Intelligent Machines 855: When You're Right, You're Right

Summary

Intelligent Machines 855: When You're Right, You're Right

Podcast: All TWiT.tv Shows – Intelligent Machines
Host(s): Leo Laporte, Paris Martineau, Jeff Jarvis
Guest: Mark Surman (President, Mozilla Foundation)
Date: January 29, 2026

---

Overview

This lively episode of Intelligent Machines brings together Leo Laporte, Paris Martineau, and Jeff Jarvis to discuss the latest in AI, machine learning, and the evolving landscape of intelligent tools. The show features an in-depth conversation with Mark Surman, the president of Mozilla, who introduces Mozilla’s AI “manifesto” and shares his vision for ethical, open AI. The crew also riffs on AI-generated corn farms, speech synthesis breakthroughs, government use of AI for regulation writing, and the broader transformations underway as AI seeps into daily tech life.

---

Key Discussion Points

1. Team Catch-Up & Tech Nostalgia

• Personal Updates: Jeff Jarvis, recovering from injury, shares tales of his L3 vertebra mishap and his run-in with a strep infection post-dentist visit. ([02:47])
• Old Gadgets: Paris highlights her beloved “Yes Man” toy from childhood, sparking nostalgia and setting the episode’s anecdotal tone. ([01:20])
  • Quote: “He also says, when you're right, you're right. And how do you do it?” – Paris ([02:08])
• The Importance of Firefox: Leo and Paris swap stories about browser monocultures and why supporting alternatives like Firefox matters. ([05:26])

2. Rise of AI Coding Agents & Cautionary Tales

• AI Helping Code: Leo demonstrates how he uses tools like Claude Code for automating tasks, exporting data between apps, and building his “Beat Check” morning news aggregator.
  • Quote: “You don't write code directly, you write a plan for the end result and the system directs your bots to build code to deliver it.” – Leo referencing Anil Dash’s ‘Codeless’ piece ([10:22])
• Third-Party AI Agents: Discussion of “Claudebot” (now “Moltbot”) and security risks in letting bots access personal or sensitive data. ([12:14])
  • Quote: “One of the ways you use claudebot is to give it permission to do anything. … That really makes me nervous.” – Leo ([13:42])
• Comparison of AI Giants: Exploration of Google’s Gemini and Anthropic’s Claude, especially concerning privacy, security, and integrations with personal data.
  • Paris shares frustrations with Gemini's unintuitive behavior in Google Sheets. ([16:34])

3. Speech Synthesis & Its Challenges

• AI Voice Demos: Leo plays examples from the new Qwen (QWEN) speech synthesizer and 11 Labs, including deepfaked demos of his own voice, and discusses the ramifications of such tech. ([33:18])
  • Memorable Moment: “That's pretty close. It's insane that it produced that with two minutes (of recording).” – Paris on hearing simulated Leo ([34:17])
• Accessibility Vs. Security: Benefits for those with visual impairments are contrasted with the threat of phishing and ID spoofing via voice deepfakes. ([40:03])
  • Quote: “Can you do it in real time? Could I have a voice changer that would change my voice into your voice?” – Leo ([40:30])

4. AI in the Hands of Everyone: Usefulness, Risks, and the Future

• Consumer-Grade AI Projects: The story of “Fred,” a fully autonomous AI agent attempting to start a corn farm (proofofcorn.com). The exercise is both humorous and a microcosm of AI’s potential and limitations. ([59:51])
  • Quote: “Fred is plan(ning) to open a Union Square farmers market stand in August 2026…” – Paris ([66:09])
• AI Writing Regulations: Troubling revelations that the US Department of Transportation is piloting Gemini to draft safety regulations, raising concerns about speed, quality, and the devaluation of government expertise. ([61:41])
  • Quote: “Writing and revising complex federal regulations can take months, sometimes years. But with dot's version of Gemini, employees could generate a proposed rule in a matter of minutes or even seconds. Hey, is that really what we want?” – Leo ([65:15])
• Arms Race and Societal Impacts: The team delves into the big question of whether governments (and dictators) could use advanced AI as a disruptive “country of geniuses” to economic and political ends. ([50:04], [53:03])
  • Paris questions the realism and motivations behind such analogies.
  • Quote: “Suppose a literal ‘country of geniuses’ were to materialize … for every cognitive action we can take, this country can take 10 or more … What should you be worried about?” – Leo reading Dario Amodei’s thought experiment ([50:04])

---

Main Interview: Mark Surman, President of Mozilla ([82:22] onward)

Mozilla’s AI Manifesto and Vision

• Breaking the Monoculture: Mozilla wants to “do for AI what [they] did for the web”—fostering openness, transparency, user choice, and privacy in a field dominated by a handful of closed platforms.
  • Quote: “We don’t want 3, 4, 5 AGIs. … We need 7 billion.” – Mark Surman ([83:23])
• The “Rebel Alliance”: Mozilla is investing $650 million (of its $1.5 billion in reserves) to support ethical, open-source AI, developer tools, and efforts to keep technology aligned with human values.
  • Quote: “Lots of other people … are going at their piece of changing the game. And what we’re going to do is invest in them … that is actually the way that it pushes the tide into open source winning in the AI era.” – Mark ([95:14])
• Opt-out/Opt-in AI: Firefox will let users fully turn AI features off – countering the “AI by default” model of Big Tech browsers.
  • Quote: “We want to build the AI you hate a little less. … The only way to get there is through ‘maybe’.” – Mark ([91:44])
• Ethics at the Forefront: Commitment to responsible tech investments, partnerships, and developing mechanisms for ethical training data and open AI models.
  • Quote: “You get to build stuff that fuses those things together [ethical and technical freedom].” – Mark ([114:30])
• Thunderbird Pro: Mozilla’s privacy-centric email client is launching secure (possibly non-US-based) email hosting to offer a viable, non-Gmail alternative. ([118:31])

---

Notable Quotes & Memorable Moments

• Paris on Firefox's User Base ([89:27]):

  “Mozilla's user base, I think, is unique in that Firefox users are nerds, by and large.”

• Jeff on Open Source AI's Future ([92:20]):

  “Is it possible we’ll get the AI equivalent of Apache and Linux—true infrastructure for everyone to build on?”

• Leo on Browser Incentives ([102:59]):

  “With Chrome and the Chromium engine, it’s developed by one of the largest ad companies in the world. …and that’s not Firefox’s incentive.”

• Mark Surman on Ethical AI ([112:44]):

  “We need an approach to designing tech that fuses human ethical factors and technical freedom together.”

• Jeff on LLMs vs. “World Models” ([126:33]):

  “Now you have Demis Hassabis [DeepMind] saying ChatGPT may be a dead end; we have to focus on world models. …That’s a potential sea change there.”

---

Additional Topics & Recommendations

• AI Summarization in Journalism: Discussed the limitations of AI summarizers for nuanced reporting (Davos example). ([41:27])
• Cultural Commentary: Matrix movies re-examined for their philosophical take on technology and consciousness, especially relevant in the AI arms race debate. ([131:03])
• Consumer Devices: The history of TV manufacturing, from American, Japanese, Korean, to Chinese dominance (TCL becoming the largest TV brand). ([141:10])

---

Timestamps for Key Segments

• Opening banter & AI clipboard tools: [01:01]–[07:28]
• AI coding agents & security concerns: [08:06]–[15:25]
• Speech synthesis demo: [33:18]–[38:32]
• Deepfake risks & accessibility: [39:02]–[41:16]
• US DOT using AI to draft regulations: [61:41]–[65:44]
• Proof of Corn AI project: [59:51]–[67:16]
• Interview with Mark Surman (Mozilla): [82:22]–[123:29]
• Matrix & simulation philosophy: [131:03]–[134:29]
• World models vs. LLMs debate: [126:33]–[128:33]
• TV industry history lesson: [141:10]–[147:48]

---

Final Thoughts & Tone

The episode is quintessentially TWiT: thoughtful, wide-ranging, and laced with humor—both affectionate and irreverent. The exhilarating pace is balanced by moments of skepticism and caution regarding AI. Mark Surman’s interview is a highlight, offering a hopeful, pragmatic path for open-source AI, while the hosts make it clear: If we want a more just and equitable future for technology, users, not just corporations, need to help build it.

---

Endnote

The episode closes with a nod to tradition and future: the “Yes Man” toy’s catchphrase, “I couldn't agree with you more completely,” becomes a running gag (and the final send-off), underscoring the show's playful yet insightful spirit.

“When you’re right, you’re right.” – Yes Man ([149:51])

---

Listen to this show for:

• Insider knowledge on the future of open-source AI and ethical technology
• Cautionary tales on giving agents access to your data
• Wry asides on government, nostalgia, and consumer tech
• Nostalgic and philosophical fun with Matrix movies and more
• The vision and challenges of the new Mozilla AI manifesto

---

Further information and downloads:
twit.tv/shows/intelligent-machines

---

Summary by TWiT Podcast Summarizer

Transcript

[A]

It's time for Intelligent Machines. Jeff Jarvis is here. Yes, he's feeling better. So is Paris Martineau. Our guest this week, Mark Sermon, is the president of Mozilla.org he talks about Mozilla's new AI manifesto. We'll also talk about corn. Can AI actually grow corn and make money doing it? That and a whole lot more coming up next on Intelligent Machines. Podcasts you love from people you trust. This is twit. This is Intelligent Machines with Paris Martineau and Jeff Jarvis. Episode 855, recorded Wednesday, January 28, 2026. When you're right, you're right. It's time for Intelligent Machines, the show. We cover the latest in AI and robotics and the smart machines surrounding us all today and tomorrow and forever. Joining us right now, Paris Martineau from Consumer Reports. Hi, Paris.

[B]

Hi, Leo. I'm surrounded by machines. They're getting ever closer. Please help.

[A]

Well, I do note that you seem to have stolen something from your father's study.

[B]

I got one of my own. I realized. I realized that he. We're talking about the yes man, which is a.

[A]

On the left side of your screen.

[B]

That my dad had in the 90s that I. When my parents and I were drunk on vacation last summer, I decided I was going to find two more copies of it and buy them on ebay and send one to my parents house because they'd lost theirs and send one to mine. And I did that.

[A]

Does it still function? Does it still say yes?

[B]

Oh, it doesn't still function. I mean, the batteries might have died since I put this together. We'll see.

[A]

That is the strangest.

[B]

Hello.

[C]

I couldn't agree with you more completely.

[B]

Thank you.

[A]

I couldn't agree with you more completely. Wow. Okay. That is the kitschiest thing I've seen in quite some time. But I'm glad he has a position of honor now in the.

[B]

Yeah, he's got to be looking on over me.

[A]

I couldn't agree with you more.

[B]

He also says, when you're right, you're right. And how do you do it?

[A]

Is this a memory from when you were a little, little child?

[B]

It is. I remember my parents turning it on and laughing hysterically. And then those phrases are things that we would just. We'll just say to each other. It's a family scene about our lives.

[A]

It's a little plastic Marco Rubio on your. On your desk.

[B]

It is.

[A]

It's good to have.

[B]

It really is. Everybody's got to get one.

[A]

Hey, let's say hello to our dear, dear, dear Friend Jeff Jarvis, no longer in a hospital bed. He's home now. Are you feeling a little bit better, Jeff?

[D]

I'm feeling better.

[C]

Yeah.

[D]

I'm getting there. The back hurts. I'm hobbling around with a. The cane.

[A]

So this is from your coccyx. No, it wasn't. It was your L2 or something.

[D]

Hey, it's my L3. It's my L3. There's the killer betting that I lost to.

[A]

Oh, wow.

[B]

So now it's gloating.

[A]

It is. He says, I've got you now. You're mine now, baby. Well, I hope you feel better.

[D]

I was taking it off, I pulled it. My feet slipped, I fell up and down. I got a compression factor to my L3. And then separately, we thought I might have had an infection to my spine. But separately, probably from a tooth cleaning, I got a strep blood infection.

[A]

Yeah, sometimes a dentist, in fact, from now on, I imagine, will require you to take antibiotic before you go in for any kind of dental work. Because of that, your mouth turns out it's filthy.

[B]

I realized that that was something I should be worried about.

[A]

Mouth is filthy.

[D]

They used to. Because I. I have a mitral valve prolapse valve thing and they used to require me to do it. The dentist yelled at me one year, I said, I'm not touching you until you take those antibiotics. And then the policy one, they just changed.

[A]

Okay.

[D]

You don't take them anymore.

[A]

And you got sick.

[D]

I got sick.

[B]

Well, I mean, it's bad that you were injured, but I guess silver lining is that they found this before it.

[D]

Got worse, turn into sepsis.

[A]

Yeah, yeah, we were worried about. I'm so glad you're okay, Jeff. And we're glad you're here for this because we have a very exciting guest this week.

[D]

Yep.

[A]

Today Mozilla announced or released a. Let me show you the website. It's great. If you go to state of mozilla.org it loads up like a command line. The future is not auto generated. It's not pre approved. It requires consent. Mozilla foundation, which creates created Firefox. There's a little Firefox video right there. Has announced that. I think they're calling themselves the Rebel Alliance. We choose humanity. We choose collective power. We choose a different economic model. Now, I'm a Firefox user. I use Firefox because I don't want Google to dominate and chromium is pretty much on every other browser. I use a Firefox fork called Zen, but it's Firefox under the hood. And I think it's important that we don't have a monoculture in browsers. But I also am a fan of Mozilla because it's open source and it's. And they're good people.

[B]

I agree. I think it's also important we don't have a monoculture in terms of browsers, but Mozilla is down to like 3%.

[A]

I know.

[B]

We've gotten to. Unfortunately, we've gotten into a kind of default monoculture. It's a long. We are a long ways away. Wow. My dad's going to get a lot of shout out today on the show. Long ways away from back in the early 2000s where my parents yelled at me for downloading what they decreed was a virus called Google Chrome. We're a long way.

[A]

They were using Internet Explorer. No doubt. They were.

[B]

And they were so unhappy with the idea that anything else could exist. And now it's the dominant.

[A]

Yeah.

[B]

Player.

[A]

By far dominant. So Mark Sermon, who's the new president of the Mozilla foundation, will join us. I'm really excited to get Mark on to talk about this. It has a lot to do with AI. It's not exactly anti AI, but it is judicious about AI and of course, it's.

[D]

It's a different path of AI that we've discussed on the show. Yeah, Open source and. And smaller human scale. I think that's what matters.

[A]

Mark will join us in about 45 minutes. I'm really looking forward to talking to him. It does change around a little bit. Our order. We're going to do some AI news first and of course, hour and 15.

[B]

Technically. Okay, 45, but yes.

[A]

Okay, thank you. Well, you're the keeper of the facts here.

[B]

I've just got a clock right here looking me in the face and I can't. I can't.

[A]

Wait a minute. Is that also from your childhood? My God, no.

[B]

It's just a flip. I wish I had this in my childhood. It's just a flip clock. I think I've given it as a pick of the week before on the show from Twemco. It makes a really satisfying little flip flip every minute, and then at the hour, it makes a very satisfying clonk.

[A]

When all three of them flip. Well, we'll be listening for the clonk. And that will be the story for Mark Sermon to join us. We'll look forward to that. Meanwhile, we have about three or four hundred stories to go over, so if you will just be patient. We will.

[D]

Last week, you all saw Leo or the week before last, maybe it was Leo did his wonderful new tool and I looked at the rundown and I said, oh my God, it's been, it's been flooded.

[A]

It works too well.

[D]

It was slop.

[A]

It's not because all of this is our articles, but my new RSS reader that I did Vibe Code has really been incredible because I could so quickly go through. I call it. I've renamed it to Beat Check because it's really my, my, my morning Beat Check where I go through all of my sources and find us as many of the AI stories as I can. And there's quite a few of them.

[B]

Okay, I know we talked about this in the show before, but while we're on it right now, how long like time wise did it take you to make to Vibe code this and how much of your usage? Because I tried to do a little Vibe coding project this week and immediately ran out of my Clod Pro account.

[A]

So the first thing I did is I paid for Claude Max. 250 bucks a month.

[B]

Yeah.

[A]

But I. My rationalization is, well, I'm writing stuff that I might have to pay for or might never. For instance, I also the other day I realized that I am using two different journaling apps. I use Day one and I use Obsidian and I didn't like it that there were different places. So I asked Claude to write me a tool, the one time only tool that was going to export all the Obsidian stuff, move it into day one. It did. It took about 10 minutes. I will never use it again because it's.

[B]

Did that not already exist?

[A]

People have. Well, see, this is what Claude does and this is what's changed, is it? Claude co goes out and searches for people who've done it and finds repositories, finds blog posts and was able to write its own thing in Rust based on information it gathered from a bunch of people who'd done it. As far as I know, as far as I could tell, it wasn't just lifting the code wholesale and there are APIs, there are interfaces, but anyway, it wrote a very nice tool, it imported it all. But this is the thing to understand. Besides the fact that you need a lot of credits, it is kind of an iterative process. You want to with Claude code, go into planning mode first, spend some time talking about what you want to do. It will ask you questions and then maybe you do some iterative coding. So for instance, the Beat Check store program I wrote on a Sunday morning over a couple of hours, but I'VE been continuing to iterate. Every time I think of something I want to take out, which is more often the things I want to put in, I can do that. And so it's on version, it's on this like fourth or fifth version now.

[D]

Because now is it labeled as versions. Can you. If you say, oh, I screwed that up, I revert to the prior version.

[A]

Because you push it to GitHub so you can say, oh, nevermind, let's go back one git allows you to do that. So yes, it's all designed kind of as you would do it as a developer.

[D]

Yeah.

[A]

Anil Dash had a great piece. Codeless, he calls it, from idea to software. That really describes this revolution that's happening literally as we speak right now. He says there's finally been a great leap forward in coding tech unlocked by AI. Not just it's doing some work for me, but we couldn't do this before and there are a number of tools people are using. He talks about Ralph Wigham, which I've mentioned before. There's another one called Gastown. A lot of these are ways of harnessing multiple instances of Claude all working together. Or in some cases there's one. One guy who's written a tool that does it in an adversarial way. It's like he says, it's like Court, where you have one Claude instance write some code and then has to justify it to an advocate who is taking the other side saying, no, no, that's a bad idea and has to prove it before it gets committed. It's very interesting. It's. This is all about orchestration. Anil says, here's the 10 point bullet point summary. Codeless is a new. Is a way to describe a new way of orchestrating large numbers of AI coding bots to build software at scale. You don't write code directly, you write a plan for the end result and the system directs your bots to build code to deliver it. And he goes on, I don't need to get in too many details, but I think it's really interesting for me to watch this organic development of people like Aneel and Harper Reed and so many others kind of expanding the space. It's not necessarily anthropic, although anthropic is pretty good about embracing it. I should also. One other thing I wanted to point out. You've probably seen the conversation, maybe even been tempted by what was Claude called? Claudebot.

[B]

Not spelled how you think.

[A]

C, L, A, W, D. Well, I.

[D]

Used to think all that.

[A]

Yeah. Anthropic got mad and he's renamed it to Moltbot, which is bad name, terrible.

[B]

Too close to mold.

[A]

And you may be tempted to try it, as I was.

[B]

Of course I should have actually, I saw.

[C]

It to you.

[A]

It comes with a free set of headphones. All you have to do is give it three credit cards. A lot of people are crazy about this thing. My alarm bells went off pretty quickly.

[B]

Well, my first alarm bell looking. I also saw all the chatter about claudebot and I was like, oh, oh, I guess I'll look into this. What is it? And my first alarm bell went off and it was like, the only really way to communicate with it is through WhatsApp or like your chat app of choice that it has to be kind of set up. And I was like, that just seems like a somewhat ridiculous extra hurdle to have to this. And then all of the descriptions I heard of what it could do just didn't sound that people were like, oh, I use Claude bought to make a reservation for me on Resi. I'm like, I could make a reservation for me on Resi.

[A]

Yeah, but this is all automated, in fact.

[B]

But it's not automated because you have to say Claude bot. I'd like a reservation at blank place at blank time. And you could go do that if you have that information.

[A]

That's how you're using it. But you could also say Claude. But I've got a couple of free hours in New York City on next Thursday afternoon. Find a good restaurant and make a reservation for me, I guess. So you could be more generic is what I'm saying. But don't, because one of the ways you use claudebot is to give it permission to do anything. It says this works best because you don't want to interact with. Works best if you just say, go ahead, just do it. And that really makes me nervous. Claudebot has amassed a Multbot, I should call it now, Multipot, which really has amassed 44. This is from TechCrunch. 44,000 stars on GitHub. Stars how you vote on a project on GitHub and say, I like this, 44,000 is a lot. So much viral attention has been paid to Moltbot. TechCrunch writes that it's even moved markets. Cloudflare stock surged 14% on Tuesday as social media buzz around the agent resparked investor enthusiasm for Cloudflare's infrastructure it uses. Cloudflare developers use Multbot to run locally on their devices. Now contrast this to what Anthropic has done With Claude, the Claude version it's making for Normies. You have to have a Macintosh called Claude Cowork. We talked about this on Sunday with Alex Stamos who's a very prestigious security guy. In fact, his latest startup is called Corridor AI and it's about security with your AI, with your vibe, coded AI stuff, how to make sure it's secure. He said Anthropic did this right? With Claude Cowork, it creates a virtual machine. In fact, that's why it takes. So have you tried it yet, Paris on your Mac?

[B]

Yes, I've been trying Claude codework.

[A]

It takes forever to set up, to load up, right. That bar goes across. That's because it's actually building a virtual machine that it's running in. It is sandboxed from your system. They've done a lot of work to keep it from being harmful. So contrast that with Multbot which says just give it permission to do anything you want. It's running bare metal on your machine. There is a, I think a real risk of prompt injection on both tools, but that's why Claude Cowork runs in a, in a sandbox. So code injection.

[D]

That's why I think, I think Google wins in the end because Google already knows everything about me. So what's the harm in adding an agencic layer on top of that? Or as Google, as Apple uses Gemini and you're in the Apple ecosystem. Well, Apple already knows everything about me. What's the harm of adding the layer there? I think that feels a lot safer than saying I'm going to take a third party so to speak tool and give it access to everything I have.

[A]

And your credit cards, all three of them.

[B]

I mean I've been trying to play around the Gemini as much as possible too and I just, I haven't found it hitting this. I mean I think Gemini in my uses is only useful when it is clear. Like a Google specific task might have an edge. Like if I'm trying to figure out like the other week I was trying to look for coffee places in New York City that were open within a couple day window that would have a specific type of coffee brewer thing that I wanted in stock. And Gemini was way better at that than Claude or ChatGPT and I think that just makes sense because they've got the search parameter thing down. But I've also just, I don't know, I've found that Gemini is like strangely lacking when it comes to other aspects of just running an AI model. And one, one example is like have you tried to use the Gemini agentic kind of injection for Google sheets? It's. It would be so simple if you had, let's say, a little Gemini popping up on the side that would take a look at your data and then be like, okay, this is the correct Excel sheet function to put in there. But instead it tries to do that all in the cell itself instead of working separately and then inputting something. It's just a mess. It just. There's so many unintuitive things about Gemini in a way that I haven't noticed. I guess perhaps it's just because the pervasiveness of Google products, but I haven't noticed the same thing among ChatGPT or Claude.

[D]

Yeah, I agree, Paris. Even in Gmail. I had to find it. I had to find an email today of something that I was scheduled to do that maybe I can't do now, and I gave it like five clues and it had no idea. And you think the Gmail search is bad enough? Oddly, this is worse. And Napy Jones, one of his many demos is that he used Claude, gave it access to his mail, and said, I need. I can't find this scheduling line for somebody. Can you go look for it? And it did, and found it more effectively than Google did.

[A]

I think there's. Look, there's a lot of utility in all of this stuff. I'm just advocating for caution, especially if you don't understand what you're doing or the potential risk. Even the guy who created Moltbot got bit by malicious actors. This is from TechCrunch. Peter Steinberger is the guy who created it. Steinberger himself was served with a reminder that malicious actors exist when he, quote, messed up the renaming of his project. Anthropic said, you can't call it Clodbot. That's confusing. So he complained on X that crypto scammers snatched his GitHub username and created fake cryptocurrency projects in his name.

[D]

Jesus.

[A]

And he warned followers that any project that lists him as a coin owner is a scam. In fact, you might have seen some conversation. I saw it on Reddit about claudebot saying, I can't use that. It's a crypto thing. And he said, no, no, he says, I don't do any crypto with Claudebot. He posted the GitHub issue has been fixed, but cautioned that legitimate X account M O L T B O T on X, not any of the 20 scam variations of it. We're there's a lot of bad Actors out there, they're using AI, they're doing some, you know, all sorts of stuff that maybe a less sophisticated user might not be aware of, like prompt injection. The. The best advice is to run. And you could do this with cloudbot to run it in a virtual machine. Interestingly, a number of people have gone out to buy Mac Minis because it only runs on a Mac, I guess, or. No, it's easiest to run on a Mac, have gone out to buy Mac Minis to run. To run Multbot on their Mac Mini that's separate from their own hardware. It's just. It's just wild.

[D]

Anyway, it still needs access to your data and other things.

[A]

Sure. And so, in fact, so does Cowork. If you want Cowork to do some of the things that they've promoted, like rearrange icons in your desktop, it's gonna have to have access to your desktop. So there's always a risk.

[D]

What does Steve say about this?

[A]

Oh, I don't know if Steve's completely up on all the risk. I think he's aware of. Of it. He's not, as, you know, advent an avid. He's. He's a fan of AI for sure. He's not as avid a user as I am. I think it's really important for everybody to. Who has some interest in this to start pushing on it. Maybe in little ways, maybe just do what you're doing with Chat GPT Paris. But the more you kind of use this, the better you understand the kind of the uncharted territory we're working in these days. When I, you know, you. You mentioned that, and I think this is the case. Jeff's the first person to do a podcast from his hospital bed. He said next he wanted to do it in a sauna. So I immediately opened up Nano Banana to put us in a sauna. Weirdly, Jeff's in a towel, Paris is in a towel. The rest of us are fully clothed.

[B]

Weirdly, this is basically the entirety of the Internet. It's training on, which was make the woman naked and then a guy.

[A]

I mean, you were.

[B]

It was, yeah, but, you know, we know what it was communicating well.

[A]

But then I said, well, wait a minute. Put Jeff and me in a towels. And it kept us fully clothed and put towels around us, which is because you can't be nude, apparently. But the point was, as soon as I opened Gemini, it said, you don't.

[D]

Want us to be.

[A]

Yeah, trust us. It said, as soon as I opened Gemini. Oh, do you want to connect all your Stuff. And I said, yes, of course I do. Because this is the new thing from Google is Google is leaning into this whole thing saying, hey, you know, Gemini will be better if you give it access to your Gmail, your calendar and your Google Drive. So it can now, I think I trust them. They already.

[B]

By the way, Leo, you trust anybody who texts you asking for gift cards.

[A]

By the way, I just want to point out Google already has access to all of that. Right, Google. Google Green, mind you.

[D]

That's what I'm saying. Yeah. You've already. You've already passed that Rubicon.

[A]

Yeah.

[D]

So now I'm just giving it over the hill.

[A]

I'm giving it to Gemini. I don't know. I think this is the problem, is in order to do useful stuff, you do need to give it access to stuff, but do it judiciously, I guess, is. Is what I'm saying.

[B]

Yeah. Or you can be like me and be like, all right, I'm going to try to use Claude code to do what they say in the demo thing of organize some files and then you plug in your big old file folder that's full of court documents that are all publicly accessible, so it's not really a security, but they're poorly labeled. And you're like Gemini or Claude. I want you to organize these by case file number and using this naming structure and then it spins and spins and spins and you fall into a terrible error because Claude can't. Claude freaks out if it excesses. PDFs over 100 pages, even if it doesn't need to open the PDF. I don't know, I'm. I had some frustration this weekend, but.

[A]

I'm just not paying enough. You need to pay more.

[B]

I'm not going to pay $250 to organize PDF' in file folders. That is a strictly $20 a month activity I would take me 30 minutes to do. I can't be doing.

[D]

Is it over the. Over the limit of Notebook lm?

[B]

That's a great question. I forget what the Notebook LM limits are, but probably not. But Notebook lm, I mean, Notebook LM is I found quite useful for. I used it whenever I was working on protein.

[A]

Definitely not. Oh, I'm sorry.

[B]

Oh, there you go. You may have thought that was Leo's voice. It's not.

[A]

That's our next story. I didn't realize it would play as soon as I opened it, but go ahead.

[B]

Oh, I was just saying I.

[A]

By the way, that's a really good use for Claude code and I totally think it would do what you want it to do. In fact, my next project, it frankly.

[B]

Took a lot more steps than I think it should have. But that's fine. I. You might.

[A]

You want to have an interaction with it rather than just go do it. That's the thing that worries me is just go do it. Although everybody's starting to talk about oh look what it can one shot and it can run for hours and blah blah, blah, you do have to pay for it.

[D]

It never succeeded. Paris.

[B]

No, I tried. I mean it appear I ended up falling into a. I have now deleted it from my cowork history because it was. There's like some strange memory leak issues I feel like with Cowork on Mac. But there. There seems to have been like a recent bug that at least people chattering online have also experienced this where it kept saying the prompt was too long, even though the prompt was like one sentence. I think it was just trying to include all of the documents in all of my folders as part of the prompt. Once we got back and forth a couple of times, but I don't know, I found it frustrating, but I tried, which I suppose is useful.

[A]

We're about half an hour away from our visit with the president of Mozilla. Let's take a little break. I do want to play this new QN3 text to speech demo because it's pretty terrifying.

[D]

Kids, kids, hide your wallets from dad. Don't let them get near him.

[A]

But first, a word from our sponsor our show today. So anyway, great to see Jeff Jarvis feeling better. Thank you for being here, Jeff, professor of Emeritus of Journalistic Innovation at the Craig Newmark Graduate School of Journalism at City University of New York. I got it in because I didn't do it in the beginning, I did it in the middle. And also the author of many great books like the Gutenberg Parenthesis, the Web We Weave, and Magazine, and soon Hot Type.

[D]

I'm proofreading.

[B]

Get your Hot Type.

[A]

Hot Type, great. What I've read so far, I love. I had to put it on the back burner because I had so many other things that I have to read, including Stacy's Book Club book, which is Friday and I'm not quite done with it. I gotta read it. But I love it so far. I'm really enjoying Hot Type. I think it's your best yet, to be honest.

[D]

Thank you.

[A]

Yeah. Well, it's a great story, right?

[D]

It's a great story.

[A]

Great bunch of stories. Yeah. The story of the Linotype. But we will talk about that. More, I'm sure when the book comes out next. It's a couple of months next month. June, June, June. Things take forever. Paris Martineau, also here, working on more radioactive shrimp.

[B]

In a way, perhaps everything is radioactive shrimp if you think about it.

[D]

The metaphor for everything.

[A]

She is investigative journalist at Consumer Reports and covers, especially covers food safety. And that is a big story these days. I think we need to do the whole milk scandal. Just making that up. Lisa said, so we're supposed to drink skim milk now. I said, no, no, no. That's the good news. We're supposed to drink whole milk now. She misinterpreted the milk mustaches in the Oval Office.

[B]

As a lactose intolerant person. This knowledge is not for me.

[A]

Oh, I'm sorry. That's terrible.

[B]

I mean, listen, it's. It's all on a spectrum, right? You know, I'm not, I'm not from childhood, I've always been like, ugh, milk tastes disgusting.

[A]

And that's clearly why your body is telling you something.

[B]

I can have most cheeses. I was gonna say I love cheese. I mean, that's the things.

[A]

That's the thing I deal with.

[B]

I love cheese. And I'll deal with mild discomfort for cheese always.

[A]

And now we have a show title. Mild Discomfort Discomfort for cheese Always. I show. Anyway, great to have you both. The. The gang is together and we're so glad. We thought Jeff wouldn't be here this week. You were going to travel somewhere.

[B]

We're begging him to not be here while injured.

[A]

Oh, sir, that's right, the eye surgery. So that's been delayed.

[D]

Oh, yeah, yeah, yeah. I've had enough going on.

[A]

Yeah, don't get old, Paris. I'm just saying. Actually, it's.

[B]

I don't plan on it.

[A]

That's not good. That's not. Because the alternative isn't good either. Stay young, be cry.

[B]

I'll be cryogenically frozen at the age of 40.

[A]

Forever young. That's the key. Bob Dylan had it. Our show today, brought to you by Trusted Tech. Brand new sponsor. Want to welcome them to the show? We love these guys. Had a great conversation with them last week and they solve a problem a lot of businesses have. If you're using Microsoft 365, there's a very good chance you are paying for licenses either that you don't need or you're missing licenses that you do need. And let's not forget, coming in July, Microsoft's been warning you about this. They're going to implement a significant price increase for M365 and also make it all a little bit more complicated. There's a lot more nuance than before. This is why you need Trusted Tech. Trusted Tech helps businesses of all sizes get the most out of their Microsoft investment by ensuring their M365 environment is well supported and aligned with how the business actually operates. Look, everyone knows Microsoft licensing is complicated. Options can vary widely. Trusted Tech's team understands all this and they can really help you understand what licenses you have, what licenses you need, how to make the most of the ones you're already paying for. If you want to make sure you're getting M365 done right, Trusted Tech is offering a free Microsoft 365 licensing consultation. Just go to the website. Now. It's a little complicated and long, so I want you to get this right. TrustedTech Team Intelligent365 okay, not.com it's TrustedTech Team Intelligent365 to get a clear data backed view of your current licenses, optimization opportunities and next steps. No obligation. Now if you think, well, Microsoft must hate this. No, let me tell you, Microsoft loves it. In fact, listen to what Kevin Turner says. He's the former Microsoft Chief Operating Officer. This is what he said to Trusted Tech. He said, you have an incredible customer reputation and you have to earn that every single day. The relentless focus you guys have on taking care of customers gives them value and differentiates you in the marketplace. Hey, if it's good enough for Kevin Turner, it's good enough for me. But more than that, Trusted Tech also elevates the Microsoft support experience because they have certified support services and they are used by some of the best and the biggest in the world. Like NASA uses Trusted Tech. NASA, Netflix, Neuralink, Apple, Yes, Apple uses Trusted Tech. Intel, Google uses Trusted Tech. Lockheed Martin and many, many more companies save 32% to 52 compared to the average Microsoft Unified support agreement. So whether you're looking to fine tune your Microsoft 365 licensing to improve the way your organization receives proactive Support or both, TrustedTech offers you free consultations to help you understand your options. Go to TrustedTech Team Intelligent365 and submit a form. You'll get in contact with Trusted Tech's Microsoft licensing engineers and they can help you make your way through the thicket. TrustedTech team/intelligent365. We thank them so much for their support of intelligent machines. Now I read that ad personally, but I think in future I may not have to because the Chinese company I think is it Alibaba Yeah, Alibaba. That does. Q W E N Kuen has released a new text to speech synthesizer that is remarkable. Let me just. This is their sample page on. Actually, it's not theirs. It's hugging spaces. Sample page that gives them some text to synthesize auto choice of the language. Although you can choose multiple languages. We've described this passage. The voice we want it to read in is. Speak in an incredulous tone but with a hint of panic beginning to creep into your voice. Here's what it sounds like.

[D]

It's in the top drawer.

[A]

Wait, it's empty. No way. That's impossible.

[D]

I'm sure I put it there.

[A]

Okay, that's pretty good. You want to hear me?

[B]

I mean, I guess it's a hint of panic, but I wouldn't say the hint is panicky in a way that a normal human. No way. That's not where I put it there.

[C]

All right.

[E]

The acting is bad. The acting is bad in that.

[A]

Bad acting. All right, well maybe this is bad acting too. But Anthony Nielsen used it to create a. I guess he used. He must have used some samples of my voice. I don't know how many he used.

[E]

He said.

[B]

I just asked him and my. My shock is because. Yeah, he just said that what you're about to hear was produced with only a two minute sample of Leo's voice. And that is astounding to me.

[A]

Here's what Anthony created. Hey, Burke, this is definitely not Leo asking you to buy gift cards. But seriously, can you grab me 100 Apple gift cards? Just kidding. This is Anthony testing text to speech. How's it sound? That's pretty.

[B]

It sounds great.

[A]

That's pretty close.

[B]

It's insane that it produced that with two minutes. I thought he was going to say like a couple of hours minimum. I was just say if you. I feel like you are probably one of the most record like the most recorded people in terms of hours of content out there. Easily accessible on the Internet right now. Well, and if you could put all of that into Quinn or something, you would have. It might sound better. More like you than you.

[A]

Burke did buy the gift cards by the way. You can send those along to me. Burke.

[D]

Headsets for all the. What excites me about this is we talked about this in the show before is that what you just demonstrated is a markup language for voice is right. You could take a book or take a script.

[A]

Right.

[D]

And then imbue it with the emotions you want. And I think that's where it gets really exciting.

[A]

I I mean, 11 in the labs has been doing this for a while and can do it quite well. They have an app called 11 Reader. The other day, I was telling Lisa about an X thread I had just read about what's going on in Minneapolis. It was really good X thread. And she said, oh, send it to me. And I said, well, wait a minute, let me have 11 labs reader read it. So I gave it the URL. I attached it to our speakers in the kitchen and played it. Now, what I'd forgotten is that I had just set up Michael Caine's voice. So it was Michael. Michael Caine doing it. But it was really good. And it is. I mean, I ended up saying, you know what? I'm paying the 100 bucks a year for 11 reader because I can now have. And it has a lot of different voices. I can have these voices, these excellent voices to read stuff. It's so good because not everything is recorded. I mean, look, I'd rather have a human recorded, but there's a lot of stuff like X threads or, you know, blog posts.

[D]

There's lots of books that should be.

[A]

Audiobooks that aren't, and there's lots of books that aren't ready and so forth. I think it's pretty impressive. Let me see. I don't know what voice I have.

[B]

Yeah, I listened to a book about the history of sugar for my supply chain book club using 11 labs.

[A]

It's pretty amazing, isn't it?

[B]

Just because I didn't like the audiobook narrator.

[A]

Well, that happens too, right? I bought audiobooks that I don't like the narrator. That's a very good point.

[D]

My name is Michael Kane. I played many roles in my life.

[B]

On and off the screen.

[A]

I've spent a lifetime telling stories. Now, I presume they license. Yeah, of course they did. Yeah. Same thing with Burt Reynolds. I've played some of these before. Deepak Chopra. I like Maya Angelou. I mean, again, that people will forget.

[B]

What you said, people will forget what you did, but people will never forget how you made them feel.

[A]

I think that's great.

[B]

That's great.

[A]

She's gonna read all my blog posts from now on.

[B]

Small quibble with this is what I did learn whenever I was having the 11 Labs voice, I chose read through. The history of Sugar book is it would be ideal if there was a way to opt because it was just reading from the epub, but of course, there are, like, footnotes throughout it, and it would read out every single number, and then it would read the number if there was a Footnote. And then a period afterwards, then it would think that's the beginning of a list and it would just get very confused.

[A]

Right.

[B]

In small and obvious ways.

[A]

I think they've gotten better than that though now, haven't they?

[B]

I mean this was like two months ago. Maybe they have.

[A]

Okay.

[B]

But I was also on the free version and just using a E book that I liberated from perhaps it's rightful copywritten state using whatever thing you sent me. So I'm sure that some aspect of that probably messed with it as well.

[A]

Well, I think we're getting better and better and better and I. Yeah, you know what? This is a really incredible accessibility technology for blind people, people, low vision.

[D]

Well, I used to be on the board of what is now called Learning. Ally used to be recording for the blind and dyslexic and amazing organizations started after World War I and people were blinded by gas and volunteers doing audio of all the books. The important thing about it is because of laws passed around this, they have the right to do any book.

[A]

How interesting. For accessibility.

[D]

But for accessibility any book can be done. But the other thing they do is that it'll be interesting to see how AI has been into this. Some of the specialized readers, if you have, let's say a biology textbook, the reader will come across a diagram of a leaf and will describe it using the reader's language because the reader is expert enough to know how to do. And I can well imagine AI doing the same thing where it comes across an illustration and does the alt text of it in a contextual way.

[A]

Really interesting stuff.

[B]

Some people in the YouTube chat bring up a good point which I'm sure that we've discussed before. But the advent of these speech synthesizing tools, the fact that you can get that good of a recording of Leo just from a two minute clip of his voice really does pose some issues or concerns for spam and spoofing calls, I mean.

[A]

Oh, well, that's exactly why Anthony made say, you know, get some Apple cards.

[B]

Yeah, I know, but I mean to. From beyond a don't trust anybody's voice point. I mean, my parents when I went to visit them were talking with this. They're like, yeah, someone we know got a call from their son. Turns out it wasn't their son, but they had a whole conversation with him. Right. And they almost sent that person a lot of money. It's getting.

[A]

So can you do it in real time? Could I have a voice changer that would change my voice into your voice? I haven't Seen anything like that. But that would be.

[B]

I mean, couldn't you have it running much like a voice to text or when you're talking, voice mode to chat GPT, you don't. It doesn't have to be text to voice. It could just be the kind of back and forth you see already in an LLM where it's purely a voice interaction.

[A]

Well, and it's not great for voiceover artists. Sam Pruitt mentioned that. And it's not good for people who read audiobooks for a living. I think real people are still better at this. Yeah, but not much better.

[B]

I mean, you can definitely tell when it's a AI, even a good one from 11 labs reading something. Just the cadence is off.

[A]

Your former boss, Jessica Lesson, wrote a piece about using AI to summarize her notes from Davos. And she said, well, it wasn't great. It was kind of. It was the obvious stuff, but it made an excellent PowerPoint presentation. After I've modified it and edited it and added my insights to it, she said, the thing that's different, consultants are done for. Yeah, right. I think that what's different is that the AI doesn't have her 20 years of experience. And the point she really made, which I thought was interesting, is the thing that's most interesting to her and her audience is how Davos this year is different from Davos last year and the year before and the year before. And the AI can't really do that very well. But she.

[D]

Yeah, but having. Having done it for about 10 or 12 years, it's always BS. What's the mood of Davos?

[A]

Yeah, it was about AI. This, this year it was all about AI, wasn't it? I mean.

[B]

My question with this is, what does this really matter? What does it matter what anybody says on a stage at Davos? I don't think it does. It's just a different opera. It's just a more decentralized version of the. This presentation should have been an emailed press release that we get from, you know, Apple or Google whenever they're announcing a product, except for its 50 some CEOs and executives.

[A]

Well, Jeff, you're the expert. I mean, Jeff went for many years.

[D]

The advantage. Yeah, the advantage is the whole way. Well, I got to go. So, you know, who are we going to say?

[A]

No, it's the lobby con is what you're saying.

[D]

It's lobby. It's lobby. It's all. You meet somebody and you. Serendipitous conversation. That's what the value is. You get Inspired by something. I mean, I've had scenes in a couple of my books from Douglas because I was there and I could say, oh, this illustrates something.

[A]

Right.

[D]

But it was useful for me.

[A]

Did you. Let's see, we, we still have a little time before our guest joins us. We're going to be talking to the president of Mozilla about their new manifesto which just came out. They're. They're pro human, not anti AI, but they're the Rebel alliance, let's put it that way.

[D]

I want to see if they have a uniform.

[A]

They probably have a tattoo. I mean. Yeah, yeah, that's. You gotta have a Rebel alliance tattoo. Dario Amode, who is the founder and CEO of Anthropic, wrote his yearly letter, the Adolescence of Technology came out this month. Confronting and overcoming the risks of powerful AI. Of course, Anthropic's always been all about the risks and mitigating the risks.

[D]

Again, risks defined by the Doomsters.

[A]

Yeah, but he says it's critical to avoid doomerism here. I mean, doomerism not just in the sense of believing doom is inevitable, which is both a false and self fulfilling belief, but more generally thinking about AI risks in a quasi religious way. I agree with that.

[D]

Yeah, yeah, I do too.

[A]

He says as of this year, the pendulum has swung. An AI opportunity, not AI risk, is driving many political decisions. Yes, this vacillation is unfortunate as the technology itself doesn't care about what is fashionable and we are considerably close to real danger in 2026 than we were in 2023. The lesson is we need to discuss and address risks in a realistic, pragmatic manner, sober, fact based and well equipped to survive changing tides. So he acknowledges the risk, but not the kind of. He says many people have been thinking in an analytic and sober way about AI risks for many years. But it's my impression that during the last couple of years some of the least sensible voices rose to the top through sensationalistic social media accounts. He says we need to acknowledge uncertainty. Yeah, we're definitely in the middle of that. No one can predict the future with complete confidence. We still have to do the best we can to plan anyway. I think I agree with that. Yeah, intervene as surgically as possible. He says, to be clear, I think there's a decent chance we will eventually reach a point where much more significant action is warranted. But we will need stronger evidence of imminent concrete danger than we have today. The most constructive thing we can do today is add.

[D]

I don't know what that means.

[A]

Well, yeah, right now it's I would say what I would paraphrase that is right now, any talk about risks are speculative. It's imaginative and so it's maybe not too smart to plan for imaginary risks. Wait until those risks are less imaginary and we can then do the right thing. He says the most constructive thing we can do today is advocate for limited rules while we learn whether or not there's evidence to support stronger ones.

[B]

I think it's a very convenient argument from the guy whose product would be hit by those rules being more than limited.

[A]

Well, but my contention is it's not just him that would be hit. I would be hit. All of us who are excited about and using AI. I think that, you know, many of the constraints people are talking about would harm society, would be societally harmful.

[B]

Harm society how?

[A]

By limp. Because this thing is a remarkable tool. It would be like the beginning of the Internet saying, you know, there's a lot of potential hazard here, which indeed there was, but we didn't know exactly what. What it was. Let's lock it down now.

[B]

I don't think anybody's arguing for it to be. I mean, I guess some people are wanted to pause but I think that there's a reasonable argument to be made that if this technology has even one fifth of the promise that its boosters argue it does, that. Yeah, maybe you should take a slower and more measured approach to building it in its early.

[D]

There's a famous paper that I quoted in the Goodberg parenthesis by James Doerr In 1998 when Google started up about looking at Elizabeth Eisenstein in the beginning of Watch Out Gutenberg and saying that the lesson of those kinds of innovations is let them go because you don't know what they're going to do until you get there. And then you adapt and you figure it out. And the danger is you delay an inevitability with a false comfort and think that you know that because you've delayed it, you know what it's going to do. All you've delayed is the questions you don't know the answers to yet.

[B]

I mean, yeah, I suppose in a vacuum that's. Or I suppose in those cases, yes. I mean, I think that it's foolish and impractical to assume that any technology like this can be delayed in any real sense, given the fact that even within our own country we're in kind of an arms race, not to mention globally, but.

[D]

And just as foolish to. To believe that. That they're harmless.

[A]

He. He says he's. He does a thought Experiment, which is kind of interesting. I'm curious what you think about it. He says, I think the best way to get a handle on the risks of AI is to ask the following question. This is a pure thought experiment. Suppose a literal, and he puts this in quotes, country of geniuses were to materialize somewhere in the world in let's say, 20, 27, 50 million people. They're not really people, but 50 million entities, all of whom are much more capable than any Nobel prize winner, statesman or technologist. The analogy is not perfect because these geniuses could have an extremely wide range of motivations and behavior, from pliant and obedient to strange and alien. But sticking with the analogy for now, suppose you were the national security advisor of a major state responsible for assessing and responding to the situation. Imagine further that because AI systems can operate hundreds of times faster than humans, this country is operating with a time advantage relative to all other countries. For every cognitive action we can take, this country can take 10 or more, maybe hundreds. What should you be worried about? Here's, he says what he's worried about. First of all, do you think that premise is realistic?

[D]

No, no, that's ridiculous.

[B]

No, it's ridiculous.

[D]

Yes.

[B]

It's like, yeah, imagine. Imagine if we lived in a world where there wasn't any oxygen and we were all on fire and the fire worked oxygen, and it was very, very hot and your mom was burning up faster, but your dad was also burning. What would you do?

[A]

It's like, I don't, It's a hypothetical. I, I acknowledge. I don't think it's that far fetched.

[D]

I mean it's, it's, it's still tech triumphalism.

[A]

Okay, one year.

[B]

Do you think that, you think that.

[A]

We'Re going to be at, we're going.

[B]

To be at 50 million geniuses level of.

[A]

Here's why he says that. Because it does accelerate. If once AI starts to self improve, which it's starting to do, a lot of Claude Code is written by Claude Code. Now it starts to accelerate at a very rapid pace. So yes, one year, a year ago, we are not that far ahead. Now, when we were a year ago, we're farther ahead. But imagine it's 10 times farther ahead in a year or 100 times farther ahead in two years. So that acceleration.

[D]

So what's he scared of?

[A]

Autonomy risks? Number one, what are the intentions and goals of this company country? Is it hostile? Does it share our values?

[D]

See, right. There is a huge presumption that the machine can ever want something, that it Has a desire, right?

[A]

Yeah.

[E]

He's already equating them to people. He's equating it to people already.

[A]

So he's saying these are people. Exactly, yeah. Misuse for destruction. Assume the new country is malleable and follows instructions. Then, okay, this is number two and is thus essentially a country of mercenaries. Could existing rogue actors who want to cause destruction. This seems much more likely, including terrorists, use or manipulate, manipulate this new country of geniuses to make themselves more effective? That seems not at all far fetched. Look how we're leveraging. Look how I'm personally leveraging Claude code now to do things to make me more powerful. I mean, it's more powerful things like bookmarking stories for the show.

[B]

But I love the idea of you looking out over your spreadsheets of links, being like, I've accumulated power.

[A]

I'm benevolent, though. But what if I weren't?

[D]

As far as the eye can see, your modeling.

[A]

I'm Oz Amandius of the AI. But now. But Steve was talking about this yesterday. There is a very significant malware that is fully AI generated running about now. And so there's no question that bad actors are using AI and they're getting that leverage. Now, I'm using the leverage for something stupid, but they're using the leverage for something very malicious. His third question, misuse for seizing power. What if the country was in fact built and controlled by an existing powerful actor? David Sachs. Xi Jinping. Such as a dictator, A rogue corporate actor. Peter Thiel.

[D]

But that's true today with a company like Palantir.

[A]

Right? True With Alex Karp, absolutely. Do we want Peter Thiel or Alex Karp to have infinite power?

[D]

Well, that's what we supposedly have regulations and laws for.

[A]

Well, but our current administration in. In order that we don't fall behind, China is saying we better not have any of those because we've got a full speed ahead with or without AI. Project Genesis. Yeah, yeah. Fourth problem. Economic disruption. If that new country is not a security threat, but simply participates peacefully in the global economy, could it cause severe risks simply by being so technologically advanced and effective as it disrupts the global economy, causing mass unemployment or radically concentrating wealth? I think that's a legitimate concern. It wouldn't be on its own behalf, I don't think. I think it's probably best not to ascribe any volition on the part of this 50 million geniuses. But whoever wields them, Peter Thiel, or China, could easily use them to commandeer the global economy. Maybe. Yeah. Put A lot of people out of work and starts firing.

[E]

These are all societal problems. These are all societal problems. It's like this isn't a tech problem.

[A]

Right, but it's a tech. But it's a societal problem caused by tech.

[D]

No. Which text is tool of.

[A]

Of.

[D]

Of agents. But it doesn't have its own agency.

[A]

Well, yeah, and I agree with you. I've always said that. I said the problem with AI is not the AI, it's the people wielding it. Yes. So you could, you know, we can leave out the autonomy risk the first hazard, but all the rest don't assume the autonomy. That there's volition on the part of the.

[E]

See, the problem here is that this is a societal problem and that they're trying to solve technologically. And that's not possible either.

[A]

Yeah, well, if you. If you're curious, I mean goes. This is really almost a book length piece. If you're curious. He goes into greater depth on this. You know what? I encourage that Amadeus thinking about this stuff.

[D]

What 11 Labs voice should read it to me.

[A]

Oh, you know what? I could. I could. Let's see. Who would you like to have Carl read it? I can have anybody you want read it. Let me just open. Oh, there is a Feynman voice. There is a Feynman voice. Okay, let's see. I think it's Michael Caine still. It's loading the website right now, so you can give it a URL, which I just did. I gave it Dario's piece and now it's going to the adolescence of technology.

[C]

Confronting and overcoming the risks of powerful AI. January 2026.

[A]

I like that voice. There is a scene in the movie.

[C]

Version of Carl Sagan's book Contact where the main character, an astronomer who has.

[A]

Detected the first radio signal from an alien civilization, is being considered for the role of humanity's representative to meet the aliens. I should pause it and get.

[D]

You see, they should also have one though where you could have Steve Martin as the jerk read it.

[A]

That would be fun. That would be fun. How do I change the voice on this? Let me see. Voices.

[D]

Let's see.

[C]

Political, technological.

[A]

Let me see if I can get Feynman. Here's Feynman reading.

[C]

I would rather have questions can't be answered than answers that can't be questioned.

[A]

Hey, I'm talking here. Wielded in my essay Machines of Loving.

[D]

Grace, I tried to lay out the.

[C]

Dream of a civilization that had made.

[D]

It through to adult.

[B]

I don't love this.

[A]

You don't like that.

[B]

This is my issue. I feel like many of these readers just.

[A]

They don't. They're not human.

[B]

I know.

[E]

Yeah, it does the voice really well, but it doesn't do like their personality at all.

[A]

Here, let's have.

[B]

I mean, not even the personality, but just the cadence of how someone would read a sentence.

[A]

Here's John Wayne reading it. What do you think? Maybe that'll make it more accessible. I like the Michael Kane voice. And they have a lot of non celebrity voices too. It's not perfect. I agree. Global peace and work and meaning. I felt it was important to give.

[D]

People something inspiring to fight for.

[A]

A task at which both AI accelerationists, AI safety advocates seemed oddly to have failed. Seems oddly.

[D]

There's. There's an evenness to it.

[A]

Yeah, it's a little too. Little too even, huh? Well, all right. Are we getting picky? I mean, it was only a few years ago that computer voice sounded like this.

[E]

But that's what AI is going to do, right? It always tends towards the middle.

[A]

How about.

[B]

I mean, literally, it is going to.

[A]

Tend to velvety actress voice. Let's have a velvety actress voice.

[B]

Want to confront the rite of passage itself. That is velvety. I will say we are about to face and try to begin making a battle plan to defeat them.

[A]

So I agree it's not human. And if you're an actor, human narrator, you're not out of work yet. But there's no recording of this. And if I wanted to listen to this in the car or I just. My eyes are tired. I don't want to read this. Or it's in dark mode and I'm Jeff. Maybe you'd want to listen to it. I think that works. No. All right. Enough.

[C]

I have created the most advanced AI soldier.

[A]

The wait is over. Tron Ares. Now streaming on Disney. We are looking for something. Something you've discovered. Give me something to believe. And some of us will stop at nothing to get it ready. The countdown is complete.

[B]

There's no going back.

[C]

Our directive is clear.

[A]

Hang on. Tron Ares. Now streaming on Disney. Rated PG 13. You're watching Intelligent Machines. Jeff Jarvis, Paris Martineau. We're waiting for our guest who's going to show up pretty soon. We are going to talk to the president of Mozilla. Mark Sermon about the state of Mozilla that he just recently published. And I thought was. I thought was. That's pretty good. Very good. Did you see the guy who's trying to raise corn with AI?

[D]

I didn't understand this at all.

[A]

Proof ofcorn.com. oh, I can't load it. It's. It's being.

[B]

There's no proof of corn. Or is this your weird extension that stops all new and fun websites?

[A]

That's. That's what's running.

[B]

Unfortunately, any extension that stops you from seeking out a website called proofofcorn.com is wrong.

[A]

I think it's. Well, anyway, you go to proof ofcorn.com.

[B]

Oh, I can clearly see this was generated by Claude because it's got the same sort of vibes as the Leo Tomato game.

[A]

The idea was he wanted an AI to see if the AI could put together a corn farm. Basically right. I don't know what the current state is because I get. I haven't seen it lately, but I.

[B]

Was going to say Claude code loves the one thing I made with cloud Code kind of looks like this. It loves this framing for its websites.

[A]

Describe it because I can't. Or maybe Benito can show it. I don't know.

[B]

A heavily padded. It's a white background, heavy padding on either side, central column of things. It's got a couple little like it. Then I don't. I don't know how to describe whatever is going on with the corn. It describes itself as an autonomous agricultural agent, monitoring fields, responding to partnerships evolving daily.

[A]

It's trying to buy land so it can raise.

[B]

Oh. As of January 25th, Fred operates completely independently. Every day at 6:00am UTC, he wakes up, checks weather across three regions, reviews his inbox, composes partnership emails, sends them. No human prompting required. But what is he trying to do? He's trying to buy land. Fred's thinking is Iowa is frozen solid, been through worse. We wait.

[A]

So no corn yet. Here's another use. Now, that's a harmless use of AI. The Trump administration says the Department of Transportation is going to use AI to write its regulations going forward. The DOT's top lawyer said they don't need the perfect rule, they just want good enough. This is a ProPublica investigation, by the way. Great, great illustration on the ProPublica webpage. What do you think should they're going to use Google's Gemini? At least they're not using GROK to draft new transportation regulations.

[B]

No, I don't think we should be doing this already. The issue is with our regulation that it is so cumbersome and overly rotten opaque that there's God knows what stuck in there, that I don't need an AI to be throwing its proverbial hat in the ring in influencing what ends up on our legal books.

[A]

This is the Sad thing. The sad state of the affairs right now in the United States government is that, you know, and I understand it all started when Ronald Reagan said the scariest words in the English language are, we're the government. We're here to help. Which really turned a lot of people against the idea that government could, in fact, help. And maybe there's more bureaucracy than there should be. I don't deny that maybe there's waste, but there's also a huge. Or were a huge number of people with great expertise in these areas who were committed public servants who are really trying to do the right thing, especially in the Department of Transportation. Many of them have been chased out of government. So the Trump administration's plan is to replace those experts with expert systems. And in fact, they say this is just the first agency that is fully enabled to use AI to draft rules. The plan is to do it everywhere. According to ProPublica, these developments have alarmed summit DOT. The agency's rules touch every facet of transportation safety. Get ready. Including regulations to keep planes in the sky, prevent gas pipelines from exploding, and stop freight trains from carrying toxic chemicals from going off the rails. Why would the government use such nascent technology, notorious for making mistakes, to write these? The answer from the government. Speed writing and revising complex federal regulations can take months, sometimes years. But with dot's version of.

[B]

Why is it take months, sometimes years? Because those things become law. And we want to make sure that this is.

[D]

This is part of the devaluation of government.

[A]

Yeah.

[D]

This is a way to say matter. Any machine can make it.

[A]

It's straight out of Ronald Reagan's playbook, and it's just gotten worse and worse. With dot's version of Gemini, employees could generate a proposed rule in a matter of minutes or even seconds. Hey, is that really what we want? Is that. I mean, if. If really, is that your priority? We want, you know, we don't want it to cost anything. We want it to be instantaneous. In any case, most of what goes into the preambles of DOT regulatory documents is just one. It's just word salad. One staffer recalled a presenter saying. It doesn't matter. It's just word salad.

[E]

You know what you're losing.

[A]

Gemini could do word salad.

[E]

You know what you lose when you go fast and cheap. Right. You lose good.

[A]

Yeah. The goal is to dramatically compress the timeline in which regulations are produced, such as they could go from idea to complete draft, ready for review by the Office of Information and Regulatory affairs in just 30 days. Shouldn't take more than 20 minutes to get a draft rule out of Gemini.

[D]

They said they don't like regulation, so they don't want to make it easier to make you think. They want to make it harder to make.

[A]

They think regulations don't matter and honestly they don't care. I think that's the case Right.

[D]

When in fact, most of your, if you're a business, most of your interaction with government power is not through law. It's through the regulations that result. It's through agencies.

[A]

Yeah.

[D]

Interpretation of law.

[A]

Academics.

[B]

Fred for a minute and Fred, how's Corn Fred doing? Yeah, Fred is plan is to open a Union Square farmers market stand in August 2026 with all the corn it be it being Claude Code has grown. It's currently has sent seven emails, including one that says something to the effect of hello fellow corn Husker, to a person who I suppose might be able to connect Claude Code with the ability to grow corn.

[A]

But see, this is what's interesting because we live in a kind of an economy. This is civilization. Right. Where you don't actually have to touch corn to grow, sell corn. Right. You just have to send an email or two and transfer some money over that. All of this can be done remote, by remote control. And that's really one of the kind of key things that happened in civilization. You used to have to grow your own food or somebody would have to go out in the field and make food. Somebody still has to do that. But it's now completely done autonomously by agents.

[D]

That's capital. Capital of labor.

[A]

Yeah, that's right. Capital and labor. That's right. And that division. I think this is a really interesting project. I hope he plans to really actually sell corn.

[B]

Trying. You know, they're pursuing contracts with someone from Purdue Agri Food right now. Chad from Nebraska is labeled as a hot lead. The rationale for that is it could be a strong collaboration potential with experienced farmer father, clear land availability and aligned interests in AI. Good starting point for proof of concept.

[D]

Wow.

[B]

We'll have to check back in in a couple of months to see if.

[A]

Did you see that Wikipedia hired a bunch of volunteers? They went through all of the AI direct that was posted in Wikipedia and elsewhere and they created a draft, a guide to spotting AI writing, which of course is immediately used to tell Claude Code. There's now Claude Code skill called Humanizer that takes all of that information the Wikipedia volunteers put together and tells Claude Code how to avoid any chatbot giveaways. The plugin is on GitHub 1600 stars as of Monday. It's really handy, writes the author C. Chen. Really handy that Wikipedia went and collated a detailed list of signs of AI writing. So much so that you can just tell your LLM not to do that. Oh boy.

[D]

There's an app that I mentioned in my hot type called Twain GPT. Humanize your writing.

[A]

Perfect.

[E]

Hey, maybe just write it yourself. Humanize your writing. No problem.

[B]

Yeah, maybe that's a great way to humanize your writing.

[A]

Why do it yourself when you get the machine to do it?

[B]

Why are you helping yourself when you machine do every fun, interesting or meaningful thing in your life?

[A]

I don't really want to raise coins back.

[B]

And you know.

[A]

Yeah, that's true.

[B]

I don't know. I feel dubious about Fred's corn raising potential. Got between today and February 5th to secure a land partnership. March Fred's plan is to deploy Internet of Things sensors and prep soil. Then planting day is April 20th and they're doing an 80 day sweet corn variety. Then they've got to figure out all of the licensing, harvest it in July and then transport money to New York Sea.

[A]

Yeah, Fred has real money.

[B]

Fred has some real money. And you can see how much of it it's spent.

[A]

How much is it spent so far?

[B]

Where are you Fred's money?

[A]

This is proof ofcorn.com Fred.

[B]

The total investment is $2,532.99. It expects $4,000 in revenue.

[A]

See, if it doubles its money every year, in 10 years it could be a monster.

[B]

I don't know that farming is that profitable. I think there.

[A]

Maybe it's over optimistic.

[B]

Isn't corn one of the ones that has to be subsidized by the US government in order for it to make work?

[E]

Yeah, that's it. It's finding all the subsidy loopholes and taking all that.

[A]

Right. I mean it'd probably be very good at that.

[B]

Doesn't have anything about subsidies on here too.

[A]

It could make the regulations and violate them at the same time. It could be like a, a. It's like a Oroboros bearing some smell.

[D]

On a Consumer Reports food story here.

[B]

It does say that if you gotta.

[A]

Write a story about this, it does.

[B]

Say that if it goes to Union Square Market that they're inviting press for that.

[A]

Oh yeah, Union Square in New York, I presume.

[B]

Yeah, yeah.

[D]

Where?

[B]

Actually, I. I don't know actually.

[A]

Yes, it's gotta be.

[B]

Yeah, it is.

[A]

I haven't forgotten your roots.

[D]

Is there a farmer's market there?

[A]

No, there's a. There what used to be a Macy's there, but not anymore.

[E]

Oh, that's gone.

[B]

My question is, how are they planning on harvesting this corn? Where is the budget for people, but where is the budget for that?

[A]

It's all built into the contract capital of labor.

[B]

It is. They do plan. They have a lot of stuff about the stand, even square up here. And not many details. But the corn harvesting, they plan on having a sign that says AI table proof of corn. An iPad with live weather irrigation. Dash figured out a human representative to roast the corn and explain, oh, it's roasted.

[A]

Oh, they're not just selling raw.

[B]

I don't think it understands how the Union Square green market works. It's not typically.

[D]

You know, there's no fires there. No.

[A]

You know what they should do? They should go to the Christmas market. Where is that? That's a. I love that little Christmas market.

[B]

Also, there's one in Union Square. There's quite a. There's like a dozen.

[A]

Because they have.

[B]

The probability of success in August 2026 is 25%. At least it's being realistic.

[A]

I think it's a good project. You know what? Let's think of something we could do. What would be better? What would be a better thing we could get Claude to do? I'll give it a thousand bucks. What? What would. Should we have it do?

[B]

Okay, I will.

[A]

A lot of people trying to make money in the. In the market.

[B]

I want to just. This is in about proof of corn. On proof of corn. Slash about this guy. Seth was walking with his friend in San Francisco. They have a little dialogue at 8:15pm as they're walking towards the hotel. Seth, I can do anything I want with software for my terminal. Fred, that's not true. You can't, like, grow corn.

[A]

Oh, maybe it's. I bet you San Francisco.

[B]

No, it's definitely Union Square, New York City, based on the documents. You know what I mean? I'm gonna grow corn for you. That'd be great. Thank you. I'm gonna figure it out and I'm gonna show you. And that'll be our first vibe coding project together. It's a physical thing. I will buy effing land with an API via my terminal and I will hire some service to plant corn.

[A]

There you go. This is our future.

[B]

Is it? And do we want that? Do we want our future to be two guys deciding on the walk to their hotel in San Francisco? They're gon their API through a terminal.

[A]

To buy corn now, Paris, if I said it was an art project, you would go oh, this is funny. This is cool.

[B]

I'm not saying it's not. I don't find it funny. I'm very amused.

[A]

Yeah. The CEO of Mistral, the French AI, says China lagging behind in AI is not true. It's a fairy tale. Arthur Mensch. China is not behind the West. This is on Bloomberg Television at Davos. The capabilities of China's open source technology is probably stressing the CEOs in the US and you know what? That Alibaba text to speech, that's an open source model that's running on hugging face. I should say open weights. I don't like open source. So even if they're six months behind, that's not that much behind and they'll probably speed up. So if it's not corn, would you accept Pokemon as a test of AI's intelligence?

[B]

What are they doing with the Pokemon?

[A]

What do you have against Pokemon?

[D]

What are they doing for the Pokemon? I was rid of that 15 years ago. I never, always bored me. I never understood it. I thought this will go away. But no, it's got to come back.

[B]

Do you see the people out wandering around in the streets with their Pokemon Gos?

[A]

My wife right now.

[B]

You could Pokemon go to the polls right now, Jeff.

[A]

Let me just. I have a little Pokemon Go right here.

[B]

I never played that. I loved playing the Pokemon games back in the day.

[A]

I'm surprised you didn't get into Pokemon Go. They're not playing, by the way, Pokemon Go, because you have to actually get out of the house to do that. They're playing the original games that you played the Pokemon games. Just see if there's any Pokemon in my neighborhood here. Oh, yes. Let's see.

[B]

Have you been on top of it? Are you your gym leader?

[A]

There's a wild dunsparce right there. Look at that.

[B]

Right in front of you.

[A]

Right in front of me. Let me just go throw my. I'm an expert in the. In the spirit.

[B]

Are you just throwing the balls? You're not having to fight it? I think that's.

[A]

You have to fight it later. If you go into the gyms, you know, and there's a. There's a battle. Yeah. Why are you.

[B]

Why are you. Why do you swirl it around?

[A]

Because it's better. You get more. You're more likely to. To trap it. If you do a little spin ball. That's the spin ball there. You get like extra. A little extra English on the thing. What's this? Oh, look at that. That's a wild Lechonk. I love The Lechon.

[B]

I do like the Lechonk.

[A]

Lechonk's great. It's a pig. Some sort. Yeah. So yeah, you spin. You didn't know about this. You spin it and that way you're more likely to catch it in your Pokeball anyway. You could see that an AI could probably do this better than I can. This is from the Wall Street Journal. We're all a bunch of nerds, said the man behind the trend. Silicon Valley at the world's top AI labs are using the original Pokemon games as a popular way to track track the progress and abilities of AI models and figure out how to deploy them toward time consuming, complex goals. You can watch them on Twitch and elsewhere. Dave Hershey, who is Applied AI Leaded Anthropic, says it provides us with like this great way, man, to see how a model is doing and evaluate it in a quantitative way. They like Pokemon Blue. Hershey is the architect. And you've probably seen this on Twitch of Claude plays Pokemon. Have you seen that? Right?

[B]

Yeah.

[A]

And there's also GPT plays Pokemon, Gemini plays Pokemon.

[E]

Yeah. So it's about as smart. As smart as a Twitch chat room is how smart that is then.

[A]

Yeah, right. That's right, because they did that today.

[B]

As me at 8 years old.

[A]

The streams of Claude, GPT and Gemini playing Pokemon on Twitch have collectively racked up hundreds of thousands of comments as the models chart their progress in real time. Wait, I see it now. Claude posted on its stream the X11 wall has an opening at Y10 to 12. Anyway.

[B]

Oh, boy.

[A]

It's just. This is just one of those lifestyle pieces on the Wall Street Journal.

[D]

You know what I love for two seconds is your symbol. Wtf.

[A]

Oh, you. Yeah, this is useful.

[B]

What?

[A]

Actually, I think it deleted it because I. Oh, no, I didn't think we'd have time for it.

[D]

Well, I figured I'm killing two minutes before Mark comes.

[A]

Yeah, Mark's gonna be here any moment now.

[D]

We're gonna talk from Pokemon.

[A]

Yeah, you're probably. Probably right. I actually did. I deleted it. Let me go to the.

[D]

Was it symbol wtf?

[B]

It's symbol wgf. But I don't know what I'm looking.

[A]

At, so you click it. So here's the thing. It's very hard when you're on a phone or anywhere else to figure out how to type in the trademark symbol. Right. Well, all you have to do here.

[B]

Is click, oh, I do know this.

[A]

And then it puts it on your clipboard and you paste it right in. That's.

[B]

I mean, you can just type in the words trademark in many cases.

[A]

But how about this?

[B]

And then you get it right up there.

[A]

Could you do the Myanmar thingamajob? This has a lot of nice little symbols and actually it's helpful.

[B]

It really does helpful on Linux because.

[E]

It'S useful for the arrows up, down, left, right, that's.

[A]

Those are useful on Linux. I mean, on Mac you have easy access to this stuff with there's an emoji key and so forth. But it feels like it should have more. This doesn't seem like the full scene.

[D]

Yeah. There's so many more characters.

[A]

Yeah, I'm a little disappointed. I do like it, though. My E in my name. The E accent, et gu, or e with a cute accent is right there. And I could just paste that in because as many of the time I've had to figure out how to.

[B]

You can also just hold the E.

[D]

Hold down the E and it pops up.

[A]

On an Apple device maybe, I suppose. Can you do it on Android? Of course on Windows you used to have to type alt0170 on the keyboard to get it. It was such a pain in the butt. But yeah, Apple's made that much easier. Thank you.

[E]

Actually, on the latest episode of Hands on Windows, Paul shows you how to do this easily on Windows. Yeah.

[A]

Oh, all right. Well, now I have to watch Hands on Windows. Oh, no, wait a minute. I don't because I don't use Windows. Wow.

[B]

Line in the sand.

[A]

We're going to take a little break and when we come back, Mark Sermon, the president of Mozilla, will join us. I am a devout Firefox user because I don't want Chrome to eat the world. And I think that Mozilla has been. Has some challenges. It's a challenging time for Firefox and Mozilla, but they have a plan. The 2026 state of Mozilla, doing for AI what we did for the web. Coming up in just a bit with the new president, relatively new president of Mozilla, Mark Sermon. You're watching intelligent machines. That guy is about to pass out because he's exhausted is Jeff Jarvis. Jeff. I mean, you can. If you. If you want to take a nap, there's a bed right over there. Just don't pull comfortably, okay? Just leave the.

[B]

At any given moment, a cane is going to emerge from the side of the screen and pull Jeff off. Jeff injures himself a third time trying to do a bit.

[A]

That is Paris Martin. I still say Paris. You should come up with something like corn that we could do.

[B]

I should. I should come up with a classic crop. I think that would be a great.

[D]

Boon for me, avoiding the radioactivity.

[B]

Why don't I just come up with the next corn?

[A]

Pork bellies. What's a commodity?

[B]

Oil.

[D]

Sorghum.

[A]

Sorghum. Let's all do sorghum.

[B]

In what sense?

[A]

I don't know. I just feel like that we could do this. We don't have to do corn. We can do anything we want. What would. What commodity? What should we. What enterprise should we. I have let me know what, 50 bucks a month on Claude Max. Let's do something with it.

[B]

This is going to be the classic next Planet Money episode.

[E]

Oh, we got.

[C]

We got Mark.

[A]

All right.

[B]

Hello.

[A]

Joins us now. He is the president of Mozilla and just generated a very generated wrote. I think he wrote it with his own hands.

[C]

I sewed it with a thread.

[A]

Manifesto. The Mozilla Manifesto. Mark, it's a pleasure to meet you. Thank you for joining us on intelligent machines.

[C]

Oh, thanks for having me. It's exciting to be here.

[A]

And I will preface this as I have before, with my gratitude to Mozilla. I've been using. I'm a devoted Firefox user because I don't think we should be in a monoculture when it comes to browsers. I should confess that while I do use Firefox, I most often use a Firefox fork called the Zen browser. Is that okay?

[C]

Yeah, I mean, the world needs that kind of diversity of browsers. I love Zen browser. And so it's not only okay, it's great.

[A]

Oh, good.

[C]

That's why it's open source. Fork it. Like, fork it. Lots of things. Right.

[A]

Fork it.

[C]

Folks, I have a guy who just joined me working as our CTO, and his thing is, we don't want 3, 4, 5 AGIs. Like, if that's the direction we're going, we need 7 billion, right?

[A]

Yes.

[C]

Should be open. Fork it. Make it yours.

[A]

Yes. So that same philosophy that Mozilla applies to Firefox, you want to apply to AI.

[C]

Exactly, exactly. I think you know that we grew up. Firefox grew up, Mozilla grew up. I wasn't here at the time, but I was a fan in a tech monoculture around the Web with Microsoft and Firefox, Linux, Wikipedia, all those things. They challenged that and they turned it around. And I think we're going into. Or we're right in the middle of a much more consequential monoculture, and it's time to break it open, do something Beautiful, big, joyous, different, diverse.

[A]

With AI, you've been with Mozilla a while. You were executive director in 2008, did that for 15 years, became president just a couple of years ago at a, I think, a challenging time for the Mozilla organization. Mozilla's percentage in the browser market is down to single digits.

[C]

It is, it is. And it's, you know, it's challenging for a couple reasons. That's one is remaining relevant when, you know, the. The browser market is basically locked down by people who bundle their browsers with Android, with iOS, with Windows. I mean, really, the dynamics of the browser market are kind of back to where they were before we started, which is people just use. Which comes with their device, and it's very hard to be an independent browser. And we make something great. We believe we're going to make it greater, but it's tough. So that's one challenge. And we have a new CEO of the Firefox organization, Anthony DeMeo. And he. I think he's got great plans. Not just to figure out what AI means in Firefox. We can talk about that in a little bit, but just to make it something that people want to switch to, but they've got to want to switch. It doesn't come with their phone. And then. But I don't actually think that's the biggest challenge for the Mozilla organization. I think the biggest challenge for the Mozilla organization is we're web people. And the thing that defines where the tech industry is going is AI is data. And we, you know, we come from an era of saying we don't want to touch any data. We're about privacy. We're still about privacy, we're still about user control. But the challenge for us is we have to figure out what that means in a world that is totally different than the world when we started. And that's a lot of what we've been trying to work on. When we say we want to do for AI what we did for the web, it means bringing in new people who know how to build AI that protects your privacy, who know how to build AI that kind of is designed around human values. So that's what we're trying to figure out right now. And I think that's actually the bigger challenge than market share.

[A]

You're also putting your money. Let me just finish this because I think it's important. You're also putting your money where your mouth is. You've committed $650 million to this effort. I didn't Even know Mozilla had that kind of.

[C]

We actually have almost a billion and a half in the bank. And you know that 650 is a bit more than we earned this year.

[A]

So you know, most of that comes from Google, from search.

[C]

So you know, the, the biggest part of the revenue comes from search, about 85 of percent of that.

[A]

That's why I was a little worried with the Google trial was going on that one of the remedies was that Google stops giving money to not just Apple.

[C]

We have known for a long time that we need to diversify and that's why, you know, the Google revenue and the search revenue has actually gone down from about 95% to about 85%. I don't know exact numbers because we are building new things and the first of those that's working is, you know, the best in class privacy oriented ads that you can see in Firefox. And if you want a different direction for the ad industry, we're actually building that and it's starting to be a part of how we support ourselves.

[A]

Nice. Go ahead, Jeff. I'm sorry, sorry.

[E]

Before we go on, sorry. Before we go on, Mark, if you're seeing Leo's screen, you're going to want to go to the top of the zoom window and click on meeting. And then you'll see, then you'll see.

[A]

Everybody up in the top left. You're seeing your, your, your, your little cute little Firefox animation. You want to see yourself.

[E]

There you go.

[C]

Exactly.

[A]

Yeah.

[C]

But that didn't break anything from before, right?

[A]

No, no, no, not at all. It's just what you tell me to.

[C]

Do that and I did. And then I, and then I was like, oh, this looks. I like seeing the Firefox window. Is that part of the show?

[A]

It will be, yeah. I love how, by the way, great job on this. I love how it loads with the command line and it's great. It does it too fast, unfortunately, because I want to read all the things.

[C]

It's trying to trick you. Well, not trying to trick you, but it's trying to, you know, just kind of dazzle you.

[A]

Show it right now, Bonito, because I'm going to launch it and you can see. There you go.

[B]

Cute.

[A]

Isn't that great? Love that it goes so fast though. You don't, you know.

[C]

And these videos, look at these cool videos. It's a little panda. You know what.

[A]

A trash panda or is that a Firefox?

[C]

No, that's it. It's a kind of a cousin of a fire box. It's trying to bring you in. It's a red panda. But we had a lot of people say that is just AI slop. Mozilla's going like doing these crappy AI videos. That is handmade cgi that is the work of computer artists.

[A]

Just so, you know, nice.

[B]

I guess this also goes to a large point is how have you handled the backlash from users about kind of rolling out more AI centric tools, rolling out more advertising, even when it is couched in this well intentioned attempt to make up more private advertising experience? How have you kind of navigated that? Because Mozilla's user base, I think is unique in that Firefox users are nerds, by and large.

[C]

Yeah, I mean, people are people. Well, and I am them. Right. People don't freak out when OpenAI puts out an Atlas browser. They're like, yeah, whatever, server. And we start figuring out what we're going to do. This browser. People care, people care about Firefox. And so, you know, the, the key is, and we've always had to navigate this, being there with products that meet people, where they are, like people are using AI. You have this, you know, ChatGPT is like the fastest growing product ever in history, certainly in Internet history. I use different AI things because they're helpful for some things. There's also a lot of slop. There's also a lot of reasons to hate it. So how do you hold the values of privacy, of choice, of control and meet people where they are and where they're going? And the needle to thread is that balance. And so how we're tackling that is to start a conversation with people partly by doing what we did in this report and say we're not trying to build the AI you hate, we're trying to build something different. And so if you do want to opt in, we're going to build something different. We want to build it with you. We want to figure out what that looks like. And the other way to handle it is let you opt out. So Firefox will let you turn off all the AI as we layer in more features. And that's a key piece. Right? We, unlike almost anybody, have always been about choice. You can go up in that little bar in Firefox and say, I want to search in Google, I want to search in Bing, I want to search in DuckDuckGo. I want to search in one of. I think there's probably a hundred different search engines you can turn on as your default. And that's going to be the Same thing with AI. You can pick between open source things, you can pick between Gemini, you can pick between turning it all off. So having a conversation by building things about what AI looks like, that's different. Has to happen through the product. And then you got to give people choices. So, you know, people haven't seen us do that yet. So you got to suspend your disbelief, which is a hard thing to ask people to do. So you know, the, the way, the only way, the only way to get there is through maybe is that simple answer.

[D]

Mark. I think I'm old enough to remember when the early days, my children, we didn't just have browser wars, we had server wars. Yep, right. The Netscape server, the open market server, and then along comes Apache, which enables infrastructure for everybody. And I wonder whether that's a possible model. Not on the consumer product side, but on the, on the kind of what we used to think of as the server side of AI. Do you think that we're going to end up with things that become truly open enough that we can all use them and choose from them in a similar way?

[C]

I sure hope so. And we're doing everything we can to help that happen. And if you kind of look back, it wasn't just Apache, it was Linux, it was MySQL. You had this whole thing that if you're nerdy and old like us, you call it the LAMP stack. And you know what happened then was Microsoft actually owned. I was a web developer in the mid-90s when the web was blowing up. App owned, the whole server stack for web developers. And then all these different pieces, Apache, Linux, they came together and they were a free and open alternative that you could hack on. And Linux, when I started as a web developer, mid-90s, I think was about 6% market share. It's now 80% of the servers in the world. It is the thing that developers just use. Can that happen? Yes. Will it happen again with AI that developers turn to open source? I think this year or 2027 will be the year that people switch and it might actually.

[D]

What are the keys you're going to do to. How are you going to facilitate that?

[C]

Well, let me answer. Let me say two things. Why do I think it's going to happen and what are we going to do? So why do I think it's going to happen? Is. Is two things. One is just on the kind of open source and open weights models, they're catching up in capabilities. So you kind of see a curve where the closed stuff and the open stuff for most Use cases that people are going to put into an app. They're kind of there and at the same time serving up. It's not free, it's not like software. You still got to pay for inference. Serving up these open source models is about 20% of the cost of building on top of a cloud or a ChatGPT API. So you have capabilities going up and you have costs coming down. Those things are going to are meaning that people are going to switch. One of the biggest kind of blockers to this is that it's still all very hard to use and there's so much stuff all spread out, much like Linux was before Linux distributions, right? You could go get a PC or download it there. So we have a company, Mozilla AI, 100% focused on developers, 100% focused on encouraging developers to choose open source AI. And what it does is it has a bunch of libraries, software libraries. It's now got one hosted product in beta called any LLM and it does exactly what you imagine as a developer. Here's the thing that just is magical tool that lets you switch between this LLM, that LLM, but also any guardrail, any agent, things that make it easy to just adopt from this ecosystem as it comes along. So that's one of the big things we're doing. But you'll also see in that report we talk about gathering some sort of rebel alliance.

[A]

I love that by the way. Thank you, Spielberg.

[C]

It's evocative, it's corny, but it's evocative. And lots of other people. And this is the key to what happened with what you're talking about 20 years ago, Jeff. Lots of other people are going at their piece of changing the game. And what we're going to do is we're going to invest in them, we're going to work with them, we're going to amplify them. And that is actually the way that it pushes the tide into open source winning in the AI era.

[D]

A very important question mark, do you have a Rebel alliance helmet, uniform, tattoo?

[C]

You know, one of the nice things about a rebel alliance, the rebel alliance is you don't have uniforms. You know, that's the whole thing that you know you started with LEO is like you use different, a different browser. You want diversity of browsers. If you have people who are a kind of ragtag group who all have their own, own itch to scratch, their own way of dressing, like driving towards the same, but, but kind of different future, I think you get a much richer thing. And that's you know, that's the uniform I'm wearing.

[A]

Love it. So what is that 650 million going to go toward? Where is, where does that money go?

[C]

We're kind of like in the answer to that. We're kind of like in the boring first week of your business school class. It's not very complicated. It is. 80% of that goes towards our existing core products. We want Firefox to be great and we just keep investing in it being great. It's not like that has changed and that's where the money comes from in the first place. And also Thunderbird people don't remember we have a very privacy centric. I love Thunderbird Email client still has 20 million people around the world that use it grew by 10% last year because people want to go back to a privacy oriented email experience. And so 80% of it just goes to keeping doing that. And then we've taken 20% and really pushed it at this challenge of creating trustworthy open source AI that you can rely on. And so that means playing with this idea of a trusted AI mode in Firefox where you have to opt into it. It'll be a separate window in the same way you have a regular Firefox window, you can have a private window, you'll have an AI window. And that gives us a place to experiment with a different way of doing AI. And if you want to play with it and use it, you can. If you don't, you don't have to. So that's some of that 20%. Another 20% is going after. Just what Jeff is asking about is the developer marketplace. So that's the Mozilla AI NELLM kind of type of approach that is a company we're likely to spin out called Mozilla Data Collective, which is a two sided marketplace for ethical training data, which, you know, sounds like a lot of words but is really people who've got valuable data who want to set the terms of how AI users AI and you know, it really, that's a missing thing in the world. And then, you know, the, the other piece is really just rallying the public. I mean our Mozilla foundation group is working with creatives in Hollywood and activists in all parts of the world. Actually, not just Hollywood, even Bollywood. They're going to do an event in India next month. People of all sorts who are not the AI builders per se, they're people who have a stake in where technology goes to imagine a very different future. There's a lot of creative work there. So that's the kind of mix of it. And then underneath there's something that is pretty new to us is we're using kind of an ethical approach, responsible approach to venture investing. Investing in companies that might make the next Firefox who are trying to build something different and helping them, kind of helping them along. So there is a piece of that 20% that is going towards responsible tech companies through venture investing.

[A]

I just want to really support this because you're the group that should be doing this. You've always kind of all the group.

[C]

Leo that should be doing this. Brazil at its best.

[A]

You have a lot of trust, I guess is what I'm saying.

[C]

Well, thank you. I mean, I hope we are worthy of it because I would say there are moments where it feels like we've lost our way or certainly people have.

[A]

Worried that, well, Thunderbird, for instance, was abandoned for a while.

[C]

Yeah, yeah. I mean you're tracking this well. So I mean it wasn't abandoned. I was one of the people who never for one second abandoned it and kept it in the family. But it was that the Firefox group was like, yeah, this Thunderbird thing, email is not going anywhere. So we held it in the foundation and then we helped it set up its own company. It still built so much by the community, but, but I think I'm glad that people still trust us. I, I hope we're worthy of it. And we, we keep building that and it is really all of us are needed. Like Mozilla at its best is, you know, the, it's kind of boring old trope that, that for many people. But I, I always love going back to it. When Firefox launched, there was this two page ad in the New York Times that was the Firefox logo made up of the names of everybody who donated to buy that ad because like people wanted this thing to win and that's the energy we need to get back again. Right. Is so many people are pissed off about AI they don't trust it, they're worried, but they don't like these companies. There's so many reasons and all of us need to band together to make something different. And guess what? But doing that and succeeding at that I believe is a far more plausible and interesting thing to do than putting data centers on the moon.

[A]

We're talking to Mark Sermon. He's the president of Mozilla and has just released their Mozilla manifesto. They want us all to join the Rebel Alliance. We're really glad to help you and support you in that. I'VE always been a fan of Firefox. I've always been a believer and I've always wanted Firefox to succeed. I've been a little worried. You got a little heat too. When it was announced that we were going to add AI to Firefox. There's a lot of people who say, I don't want a browser with AI. And Vivaldi even said, we're never going to put AI in our browser.

[C]

It was very funny what Vivaldi did.

[A]

Because they said, until it's good, he.

[C]

Just said, here's our roadmap. No AI. Bye. Look, we don't exist for one set of people. That's the point of Firefox. Some people are going to want AI that comes from the big guys. Go do that and you can actually do it through Firefox. Some people want a different version of AI. We're going to build that with them. And some people want no AI. And that is the difference of Firefox. On the one end, you have Google and Apple and Microsoft, they're going to just shove their AI into their browser, into Gmail, into Office, into everything. And you're going to have to take what you get, right? And then the other end of the spectrum, you have a Vivaldi of no AI. That's actually not the real world. The real world is people are worried about that vertically integrated lock in big player stuff and people think this stuff is useful. Of course, lots of people are going to want to opt out of AI altogether. We're also there for that. But there's a real gap in the world of running straight into the problem of we have to build this stuff differently if we want to use it. And that's what we're going to do.

[A]

It's really clear. I mean, one of the reasons I use Firefox is really clear that with Chrome and the Chromium engine, it's developed by one of the largest ad companies in the world. And so their incentives, you know, Moving to manifest V3, for instance, really are in making sure that the ecosystem is good for their business. And that's not Firefox's incentive. And so, yeah, and that may be.

[C]

I mean, that may be people aren't tracking the manifest v3v2 thing, but, you know, basically that goes in Chromium and makes it so ad blockers don't work anymore.

[A]

Right?

[C]

And we actually left v2 compatibility in Firefox so people could still block ads. And that's the kind of classic Mozilla ability to hold tension? Yes. Does 85% of our revenue come from Google? Yes. Because people use Google to search. They like it.

[A]

That's how the revenue comes. We should say when you use the search bar, if Google's a default search, there's referral fees that go back to.

[C]

Mozilla or if you search in Bing or whatever. So does most of our revenue come from search advertising? Yes. Do we allow ad blockers? So are we the browser now that blocks the most ads in the world?

[A]

Yes.

[C]

And so holding that tension is the kind of thing that Mozilla is good at.

[A]

Yeah. As long as I can use your block origin, I'm happy. In fact, when you install Firefox now, you propose privacy badger, a number of extensions that are very privacy forward. I think that's fantastic.

[C]

And what we haven't figured out yet is so we'll continue to do that and I'm proud of that.

[A]

Yeah.

[C]

What's the equivalent for AI? So what is it that if I want more control, I want more privacy? What are the things that we're going to offer? Who's got the extensions? What features are we going to build? I think that's what's interesting at this moment. When Mozilla started, it was just basically a clone of the Netscape browser from the early 90s, the mid-90s, and Firefox was like, what else are we going to do with this Netscape code and turn it into something special? So like, like guess what it did. It added pop up blockers. Like that was a crazy thing at the time. We were all inundated by pop up ads. And so I think we're in a moment where we've got to figure out not only how do we give you privacy badger, all those things, but what is the next privacy badger? What is the next pop up blocker? What's going to let us engage with this new era of technology, but on terms that feel better for us? And that's a pretty cool moment of invention to be in.

[A]

Yeah. You know, one of the things that there's been up and down, ups and downs with Firefox, I've defended it all along, but there were moments when I was hard to defend. What's your position on progressive web apps now? PWAs. Because for a while you were not supporting those.

[C]

So what's a. I mean, over time were evolving to kind of go with it. But what was the issue with progressive web apps? I mean, maybe I tuned out to that. You're way more in.

[A]

I wanted you to support it. But in order to use progressive web apps. We really. Apple didn't really want to support it. Safari had kind of very weak support. You had to use Chrome and I didn't want to use Chrome, but I wanted the idea of progressive web apps to me seemed like a very, very good idea at the time. Now, maybe there were technical reasons. That's the thing. It's complicated and as you point out, you've kind of changed direction a little bit in that.

[C]

And I think whenever the shift in the technology. Progressive web apps is one I remember going way back, you can hardly remember this. I mean, you may be able to.

[A]

Be like, you can tell, I can remember it. Jeff can remember.

[C]

Maybe collectively, we don't remember. There's a time that all the videos in the browser were in Flash, which was this weird plugin that wasn't a web standard. And then the different browsers, like, who's going to put in HTML5, videos in the web, all of these things. When you get to a change like that. And progressive web apps is a thing like that, it's not a standard yet. It's like people experimenting with what can work in the different browsers. So I think the progressive web apps thing is more like that. It's like we're working out collectively as an industry, you know, what's going to happen, when do we think it's the right thing? So, you know, we usually, when it's the right thing, we come along over time and more often than not, we're the vanguard. I mean, a lot of the stuff, you know, like around tracking, protection, many, many other things. Firefox has been the vanguard before the other browsers.

[A]

I don't know if you get the credit for this, but I have to say, in the last three or four years, Firefox has improved immensely in speed, memory usage. Firefox sync has gotten better and better and better. It really is now the kind of the state of the art, I think, in every respect. People who perhaps said, well, you know, there's still this perception that Chrome is faster, for instance, and maybe it is marginally faster, I don't know. But the Firefox engine has gotten better and better and better. I think if people are listening, have said, I don't want to use Firefox, really ought to try it because it's so important that we have this choice.

[C]

I don't take credit for that. The Firefox team does. And that's a part of a they've done a good job family and they've done an amazing job. And it Was a huge lift. Like, we fell behind on speed for sure, and a bunch of things. Compatibility. And Google just leaned in hard and we lost, lost the ball a little bit. But the Firefox organization has really grabbed it back. And you have a leader, which I'm excited about, Anthony, as the leader of the Firefox organization, because he is a technologist. He came in and he turned up the gas on Firefox in the last year. He came in to be the head of Firefox. And then you have that person leading that whole company now, Mozilla Corporation. And he, he has a vision. He's got the team galvanized, motivated, focused to just make Firefox great.

[A]

Well, great. Well, thank Anthony for me then, because I will.

[C]

Well, you should get him on.

[A]

Yeah, we will, we will. I'd love to, Mark. And I thank you for your leadership in all of this, especially in AI. I think what. You know, and maybe Jeff, you got to write this book because I think the history of the browser is really fascinating because it was just a window into the Internet. It was how you accessed the Internet. It has become basically the operating system for you, Jeff.

[B]

Jeff.

[D]

Yes.

[A]

The operating system is a browser. And I think that the role of the browser in our lives is significant. And it does pose significant technical challenges. I understand that with DRM video and all these kind of complicated standards or non standards that you have to adhere to, I think Mozilla has done a great job of threading its way through this complexity.

[D]

Well, the other thing is you mix the technological sophistication and work with the ethical considerations. I was delighted.

[C]

That's why I still work here. I mean, I know for a long time, but that's why I still work here. I mean, there's very few places you can do that.

[D]

When Mozilla did its mastodon instance, I was delighted. When you did it, sad when you left it, I understood why you left it. But the important thing to me was that you dealt with ethical issues around social as well as freedom.

[A]

I want to point out my mastodon instance still uses Mozilla's rules. Mozilla's privacy statement, all of that stuff I said, this is the way it should be done. These are the rules we agree to adhere to in terms of moderation and everything. And you guys really great.

[C]

I mean, two things I would say about that. One is the original Mozilla manifesto. I mean, right now, the state of Mozilla you're referring to, Leo, is just kind of us re upping on that in the AI era, but was never meant to be like our rule set only. And I know you mean a different rule set when you're talking about the Mastodon thing. It was meant to be a call to action of, like, people to build technology aligned with human values.

[A]

Yes.

[C]

And I think that's still durable today, whatever that is 15, 18 years later was written in 2007. And so it's not just us, it's others with those values. And I think probably, if there's a place where maybe we lost our way or maybe lost our edge more. We only get to a better future if we're constantly working with each other and others who are trying to build technology with this values. And so the mistake for me around Mastodon or the unfortunate thing around Mozilla Social was we weren't the best organization to take that on.

[D]

That's why I get why you did it.

[A]

Yeah.

[C]

And it's not an apology. It's to say, like, what we actually needed to do was, like, work side by side with the people who are taking on Social. It's why I'm involved in freer feeds with the people who are taking on this part of AI with the people who are taking on this part of chat that, you know, people like signal.

[D]

But you also brought that ethical consideration again, alongside freedom.

[C]

Yeah, I see what you're saying. Because it's the human.

[D]

Similar thing around AI where you start to. You're not. You don't. I remember having this argument many, many years ago with Tim O'Reilly, we should all get badges that were good bloggers. Right. I don't think Tim would say that anymore either. So I'm not asking for a simple rule set, but what you've always done is pushed not just the technology, but the ethical considerations. So is that part of what you want to do with AI is where I'm kind of here?

[C]

Yes. Yes. And so I wrote a paper. No, I. A bunch of us wrote a paper back in 2020 called Creating Trustworthy AI. And that's really the foundations of a lot of where we are now. We updated that in 2024 with a similarly titled paper. And the idea that how do we hold. Really, the human value is part of the manifesto, and we actually added in 2017, a second half of the Mozilla manifesto around making sure that truth and human values are a part of what we consider when we build things. How do you align that with open source, security, freedom, privacy, all of those.

[D]

Right, right.

[C]

And that. That's the vision we've had for AI that we've evolved over the last five years to. Your point is we need an approach to designing products, to designing technology, to designing the economics in this era of tech that fuses those human ethical factors and those kind of technical freedom and privacy and security factors like together. And it is what we're trying to do. And it's where, you know, the, the Thunderbird people are very cautious for reasons you would imagine around getting into AI because it's a privacy oriented email client and they're not putting AI in Thunderbird in any way that they have in their plans. But that group of people, they're working on an AI assistant experimentally that is all encrypted, basically like a signal for chatbots. And so you have people who are trying to infuse all of those values together and build things around them. And that's what's exciting about Mozilla is you get to build stuff that fuses those things together.

[A]

Now I'm bummed because I have, as you can see on my screen, always on my mastodon, linked to your content policies on Mozilla. And I realized it's no longer there. So I'm going to have to find a cache of it and I'm going to have to put it on my site.

[C]

Well, you know, one of the other Rebel alliance members or these people I've always been a fan and supporter of is the Internet Archive. So they've got there for you.

[A]

I'm sure they'll have it. Yeah, I'm going to steal it if you don't mind. Let me ask you, because it is an AI show, are you going to have your own models? What is, what is, what are you going to do around AI?

[C]

That is such a great question. I mean our view is there's a lot happening in models, whether it's the closed models or open source models. And up to now it's been very expensive to train them. And so us, our focus has really been about what do you build on top of it, like the developer experience layer. When I talk about any LLM and those kind of things or what do we do in terms of being able to fine tune them, like building this Mozilla data collective and not going into the model space because there's so many people focused there. That said, there are some gaps and there are some changes. And so first the gaps, like you have more and more capable open weight models, whether it's GPT, OSS or Quinn, like you know, from big tech or from the Chinese companies, and those are good. But the gap is there are no fully, highly performance, state of the art. What we call fully open models, which means, means these open weights models, you can tweak them and you can use them for free. It's like they used to talk about free as in beer, but they're not free as in speech. And you can't really modify them or see how they work. And that's real open source, where you could see what's the training data, how did they actually do the training, what's the training code, is it open source? How do I tweak it? There are people like Dallin Institute in Seattle who their Olmo models are great and are truly open, but we need to see more investment in what they're building and we need to see more people building these truly open models. So we are looking about how do we work in partnership with some of those players where we had closed LLMs, we have open weights LLMs. We think that we can get to the point working with the right people to have competitive truly fully open LLM. So it is a place that we're focused. Rafi, our CTO is, is looking at. So who do we partner with around that? But then the other thing that is changing this is more maybe where we will have our own models is, you know, people working on small models that are highly performant for specific use cases. So you guys can have a small model you train, which, you know, there's a company that's going to come out of stealth next week that we're, we've invested in that is really focused on this. The model exactly for this show or the model exactly for, you know, my accented version of this language in the Basque region or pick your anything. And I think small models being orchestrated together is more what people are going to use. And I think we will get into that piece of the space. So a lot that that will happen there. But I don't think that's the most interesting story. I think the most interesting story is how do people adopt and use that stuff and how does open source win at that level? Much like kind of boring in many ways. The Linux distribution or package management in Linux, which is the thing that lets you pick which things you're installing. But once we had those, all of a sudden you had a whole world that you can pull together into anything. That's the part of AI we're focused on.

[A]

Firefox.com, get the best browser, support it Thunderbird.com if you want to get the best email. What is Thunderbird Pro. This is something new. Do you know anything about this?

[C]

Well, I think I know a little bit about it, yeah.

[B]

Do you know anything about this?

[C]

Basically, for years and years and years, Thunderbird was just a desktop email client. And now it has an amazing Android version. They're working on iOS they thing, but it's still like a client. It does. It's not a webmail, it's not a server. And so what Thunderbird Pro is, is to give you email hosting and webmail that is encrypted. And so, you know, you're going to be able to go and get safe, secure email, move your domain there if you want, get a Thunderbird email address if you want. And, you know, it's kind of finally moving into that space, which, you know, maybe seems like 10 years, 15 years too late, 20 years too late, who knows? But there's still a demand for it. And I think especially as things like Gmail get junkified with AI and you know, the AI can be helpful, but you don't get to turn it on or off there. I think people are going to be looking for alternatives. And Thunderbolts.

[A]

Interesting. Where are the servers going to be?

[C]

That's a good question. I don't know the answer to that.

[A]

There's been a lot of interest I see in Europe and elsewhere for servers that are not. Not in the US I use fast mail and there are a lot of people who would like a fast mail replacement that's not run servers in the US Maybe I'm one of them.

[C]

Yeah, I know that the Thunderbird folks are looking at how they spread those things out and maybe give you those choices. I also know the Firefox people are looking at AI models that are European, that would just work in Europe.

[A]

So I like it.

[C]

There are plays like that we're looking at. I don't know exactly the answer for Thunderbird Pro and where it, where it starts out of the gate.

[A]

I've signed up on the wait list. It says $9 a month. I'd be glad to support it. Very good tools and we really appreciate this new manifesto in this new direction. I think you're really on the right track, Mark, and so glad we could have some time with you. Very impressed with everything now.

[C]

You, you guys are great. And anytime you want to talk more with St. Anthony Ryan, who makes Thunderbird.

[A]

Is also a. Yeah, we'd love to get more.

[C]

Thank you guys for doing the work you do.

[A]

Do I have to ask you one question? You worked at Rogers Community Television up there in Toronto. I Used to do call for help up there in that beautiful building right there on the water. That old kind of brownstone building. Was that where you.

[C]

Wait, so where were, where were you in that brownstone? No, that's not where I was. No, I was in Parkdale.

[A]

Okay. Because they had, that was in the, in the, like the second floor. They had a news operation that was in park like 30 languages.

[C]

Oh, that was the CFMT. I did actually work in that building, but not on community television.

[A]

Okay.

[C]

So that, that building is right by the island airport.

[A]

Right.

[C]

And that was the, the, the multicultural multilingual.

[A]

So cool. I'm sure it's gone now.

[C]

And I worked upstairs when there was something launched called ytv, which was youth television. It was like the Canadian equivalent of Nickelodeon. And that was. That was probably my third. I was, I don't know, 22, my third job in television. All my career was, was television.

[A]

Okay.

[C]

And I was a guy who played the, the old episodes of Bonanza and. Or Lassie or Dr. Who, you know, very important job.

[A]

We produced Call for Help in that building, probably in that same studio. 2005, 2004, 6, 7, thereabouts.

[C]

Yeah, I was busy.

[A]

Yeah. And I'm sure that was the studio and that they had replaced that. And all I remember is there was a big pole right in the middle of the studio that we had to work around with the cameras.

[C]

There always is.

[A]

But it was, it was fun working in that location. I really enjoyed it.

[C]

Yeah, no, I love that building.

[A]

I loved going to Toronto, man, every, every month. I really enjoyed it.

[C]

We got lots of good stuff there. Come on down.

[A]

Yeah, it's up for us, Mark.

[C]

That's true.

[A]

Yeah. Above the 49th parallel. That's where the good stuff is.

[C]

You're always welcome.

[A]

Are you, are you based in Toronto?

[C]

I am based in Toronto. I actually moved a little bit outside of Toronto about two hours to Toronto to a little farm, five acre farm, not a big farm, near my mom. I can go have dinner with her.

[A]

And you're probably freezing. I'm sure it's pretty chilly.

[D]

How's the fiber?

[C]

Pardon?

[D]

How's the fiber?

[C]

I unfortunately am beholden.

[A]

As much as I hate that we are.

[C]

But wait, there's a story on that. I am unfortunately on Starlink for now, which is expensive, but there is a kind of partnership between the municipality and a small telco rolling out fiber. So they've been watching the fiber trucks, you know, rolling the conduit like all last summer. And this year it'll turn on.

[A]

That's the way to do it.

[D]

Your street.

[A]

That's part of that should be part of the manifesto is municipal Internet. That's really another piece.

[C]

Well, it's a public private partnership. It's a small company and the municipality. But yeah, it's like that.

[A]

Yeah. Thank you so much, Mark. I appreciate your time.

[C]

All right, thanks everybody. Great to hang with you guys.

[B]

All right, thanks so much.

[A]

Mark Sermon, president of Mozilla, the Mozilla organization. Quick, go to firefox.com and download Firefox. You know what? If you've got a bad memory of it from years gone by, it's better. It's so much better. It's phenomenal. It's phenomenal. It's really, really good. It's on all my machines. Thank you, Mark.

[C]

Thanks guys.

[A]

We'll continue with intelligent machines in just a moment. This is. I am with Paris Martineau and Jeff Jarvis. Well, I think given Jeff's enfeebled state, we should.

[B]

Given the amount of walking technology within arm's reach of Jeff right now.

[A]

We should probably wrap things up because poor guy is Cacio and Pepe. Do you get Cacio and Pepe still or do you have to eat like jello for the rest of your life? I mean.

[D]

No, no. Okay. I had a. I bet they have.

[B]

Cacio e Pepe Jell.

[D]

Oh, I had the roast roast beef on vec sandwich today from.

[A]

That sounds good. It's good.

[D]

It's very good. It's like a upstate New York sandwich.

[A]

What's that?

[D]

It's, it's. I forget the folder. It's a kind of bread with caraway seeds and salt and my wife's taking wonderful Caribbean and.

[A]

Yeah, nice.

[D]

No, no, no wine.

[B]

No wineless.

[A]

For you.

[E]

Yeah, he's on painkillers.

[A]

Oh, that wouldn't be good for you. No, no, very relaxed. We're gonna let you relax because he.

[E]

Doesn'T need the wine.

[A]

Yeah, that's right. Good point. You got to get the pharmaceutical grade wine. We are going to wrap things up with our picks of the week and if. Wow. I know. If there was some big thing you wanted to talk about that I didn't.

[D]

Mention, even though you're. You're getting. You're ending this in my behalf.

[A]

Well, if you continue it, that's you. It's on you, man.

[B]

It is on you. You were the one who said you didn't want to do a three hour show. We're hurtling towards that.

[D]

I know, but I am going to.

[A]

Give you that opportunity. Absolutely. What do you want?

[D]

To Talk about lines 228 to 229 or 30.

[A]

Lacune and Demise have something to say.

[D]

So. So what's interesting here is as I've talked often on the show about how Yann Lecun has talked about world models versus ll. That's actually a whole. The thing you're showing now is a kind of a joke of his. That's not what I'm talking about. Oh, but, but it's the battle of scale and LLMs versus world models. And, and I saw a debate between Adam Brown from. What's the Google company?

[A]

DeepMind.

[D]

DeepMind.

[E]

Thank you.

[D]

From DeepMind against Yann LeCun. And Jan told me afterwards about how so many people are LLM pilled and they believe that it's all about the LMS and the scale. Well, now you have Demis Hasabis who is now saying that ChatGPT may be a dead end and that we have to focus on world models. I think that's a really important moment in the paradigm of where we head in the next year or so. Where you got OpenAI is going to be going in scale route and you have Nvidia, Yann LeCun and maybe DeepMind heading toward a world model route. It's never been exclusive, but it's really interesting to me that I think that's a, that's a, that's a potential sea change there. So I just wanted to get that on the record. That was it.

[A]

Yeah. And I don't, I don't think we know yet. And. Oh, we know. And, and I also, I. It makes sense. It's reasonable. But remember, there's also. We're walking through a little bit of a minefield that has to do with history, that a lot of this are. These are debates that have been going on since the 50s. We talked about this, which we talked about last week. And to some degree what you're hearing is kind of the echoes of those debates as they've gone down through the ages. Remember we had when we talked with the author of Mathematica, Stephen Wolfram, about AI, and he said, well, it's not over for symbolic AI. That was the AI. That was the one that, you know, if there was a AI winter, the one that led us into that AI winter and was replaced by transformers, which is what LLMs are. And you know, what I think Stephen was kind of saying is, well, don't give up on symbolic. And I think that's kind of what these guys are saying in a different way. Don't give up on other kinds of models. But what's remarkable is how far Transformers have gone in a way I wouldn't have expected.

[D]

Yes, no one.

[C]

To be accepted.

[A]

The counter argument might be. And I've heard this expressed too, we're Nothing more than LLMs, and our brains are just prediction machines. In fact, I think it was Demise Hasibis who said this at Davos. He said it's. It's hard to prove otherwise, that all that we're. What we're doing as we talk. What I'm doing right now, as I talk is not exactly what an LLM does, where it's saying, well, what's the next word? And what's the next word? What's the next word? What's the next word? And we've gotten pretty far with that. It may in fact be that we can get as far as a human mind can get now. I think it is absolutely true. Also, humans definitely have a model of the physical world.

[E]

This though, is a philosophical discussion that's been going on since the beginning of time. This is like the original question.

[A]

You're right, Benito. It's. The original question is what is a human?

[B]

It's an original question. This. And I'd argue, listen, I was trying to decide between two different things for my pick of the week, but this segues perfect into my second one, which is, I think, an unusual pairing. I think I've mentioned I've been trying to watch all The Matrix, all four Matrix movies of the friend I watched Matrix 3, the most maligned Matrix 1. I thought it was great too. I think it pairs excellently with reading Claude's new Constitution and this debate. Because the core premise of Matrix Revolutions, the third Matrix, and what I think is the most interesting part of it is it's a intense, like exploration of what is the end state. If you take the idea that software programs or a large language model in this case is essentially capable of doing this sort of thought that humans are. It's kind of a significant part of it explores like. Like what would it mean for a machine to feel something like love? And does the idea of it feeling something. It does. The argument the Matrix makes is that that doesn't even matter because that's simply just a word that we view have just have come up with to describe like inputs and outputs. And couldn't a machine be able to do the same thing? Anyway, I. I've been really enjoying watching all the old Matrix films one, because I think two and three are way better than the Internet in the last 20 years has led me to believe.

[A]

Isn't there four as well?

[B]

There's four as well. And I saw. I'm excited to see four. It's been four years. Five years.

[A]

So where do you stand on the simulation hypothesis? Because that's the central theme of all of this.

[B]

What do you mean? That life is a simulation?

[A]

Yeah, that we aren't. That we think we're in a real world when in fact. Well, in case of the Matrix. Well, I don't want to spoil it for anybody who hasn't seen it.

[B]

Okay. No. If you're worried about having the Matrix be spoiled, turn off the podcast, watch all four Matrix movies, then turn it back on.

[E]

They've had 20 years, Leo. They've had 20 years.

[A]

You've had. You've had 25 years that I saw the Matrix without any. I had no idea what I was going in to see. I just thought it was a Keanu raves movie. And I came at it with. Out of it with my eyes like this, like. Oh.

[B]

I mean, I think that's the thing about the Matrix that has resulted in having such a disjointed reputation as originally a trilogy and now as a. As four movies. A quadrangle, whatever you may say is that the first movie was literally. I mean, just a leap forward in terms of so many.

[E]

The first one was irrevocable.

[B]

It was a revelation. Evolution.

[E]

That's why you can compare the other.

[B]

Two as well as ideologically it. I. There is this very interesting. So I'm also watching these all on Blu Ray and there is this very.

[A]

Isn't it nice, by the way, I just want to say, doesn't it look good on that tv?

[B]

It looks phenomenal on that tv. And it looks. I would. I mean, this is a recommendation.

[A]

She has it.

[B]

I would really recommend watching it on a fancy TV on a Blu Ray disc. Because, I mean, the first time when I watched the first Matrix movie, like a year or a couple months ago, it's not the first time I watched it, but the first time in this thing I watched on streaming and I bought. Decided to buy Blu Rays because the sound was so off on streaming, but the visuals of it are also off. But this is all to say, one of the things you get when you buy movies in physical disc form is all the extras. And the Wachowskis have this. They kind of have both. Both the Wachowski sisters now faded from life and have faded from life, public life since the first one. But they did, instead of doing a traditional audio commentary, wrote like A little essay that goes with the second and third Matrix movies that I found quite interesting in part just because of. Let me try and find the line. I took photos of it. To be honest, it being the Matrix didn't look like other stories, owing perhaps the dangers of fraternal conjugation. It did, however, look a lot like us, reflecting our own particular genealogy of interests. And while to some it may appear a hideous convoluted mistake, to us, a chop socky film that comments on the Hegelian dialectic, while having a guy who can fly and stop bullets is something that we are, well, pretty darn proud of.

[A]

As they should be. Wow.

[B]

And I'm just saying it's. I would recommend. It's a true delight to revisit the Matrix movies, but especially the headier, more philosophical 2 and 3 in the age of AI. Because the Wachowskis were grappling with a lot of the kind of core debates we're seeing major frontier AI companies like Anthropic grapple with today. But they were doing that 23 years or earlier. And it's fascinating.

[A]

Are you watching a Blu Ray or uhd?

[B]

Uhd, yes.

[A]

So you have a UHD player. That's the way to watch it, isn't it?

[B]

A uhd? Blu Ray, Isn't that the full name for it?

[A]

They're two different things, okay?

[E]

Blu Ray is just the medium. Blu Ray is just like the disc itself.

[B]

The medium is the mess.

[A]

UHD 4K on Blu Ray.

[B]

Yes, that's what I've got. A UHD.

[A]

You can get a Blu Ray. That's not.

[B]

Oh, I know, but I got the UHD for. And I've got the Deja Vu.

[A]

That's the one. $59.99 on Amazon. I'm gonna buy it right now because of you. Because I have to give you what I'm saying.

[B]

Yeah, I would.

[A]

Maybe watching them all together will help me.

[B]

I will say also watching two and three together as one movie is also the move because you may not remember when it was released. Matrix 2 came out, I think like two or three years after Matrix 1. And then six months later Matrix 3 came out and Matrix 2 ends on a cliffhanger. It's like basically just 2 and 3 are one big movie that in today's world would be a three hour Oppenheimer esque mega movie. But they couldn't do it. They were split. Would have been insane.

[E]

You can't just make two. But it is three. You can't just make two.

[B]

Yeah, that's it's true. You do kind of have to. And it's very interesting because I think my Matrix one is all about the action and the conceit and the stylistic nature of it. Matrix 2 starts off a bit muddled because they then try to pull out towards Zion, which is like the machine city outside of or like the people city outside of the Matrix, and try to. They start to introduce all these complex philosophical notions. And part of you is being like, but I just want my big battles. And they're there. They're the back 2/3 of the movie.

[A]

I don't want big battles.

[B]

The reason why I would argue for Matrix 3, honestly, as a quantity YouTube summary, is because Matrix 3, so much of it is like, let's explore how these programs are now people and their whole family relationships and talk about philosophy and Hegel. And then we go to a big battle set. And I'm like, no, give me back my quiet philosophical musings, please.

[A]

Well, I just bought all four of them and I'm gonna watch them one after the other. And then I'd recommend gonna read Infinite Jest and I'll talk to you in about four years.

[B]

I was about to say my other pick was that Infinite Jess is turning 30 and I'm trying to decide whether or not I should miss out in this show.

[A]

Have you finished it or not?

[B]

I've read it. I've read it. I finished it once in college and I read like two thirds of it last year.

[A]

It has been sitting on my bedside table for at least 10 years. I keep meaning to read it. Yeah, well, you have the hardcover, I have a paperback. I carry it with me on trips. I carry with me hoping that someday.

[B]

Are you using a 2 bookmark system? Oh, that's the cool thing. You gotta get a one bookmark up here and then a bookmark in the back for the footnotes. There's a lot of subtext and the footnotes are important. There's a.

[A]

You probably can't do it on an E reader very well and you certainly can't do it in audiobook.

[B]

I mean, an E reader would be great if they had. You just be able to skip tap and go to the footnotes, then tap back. But I don't believe in them. Anyway, my Infinite jest is turning 30 and the 92nd Street Y is doing a big event to celebrate it that I want to go to, but it's next Wednesday and so I don't know.

[A]

Are they reading the entire book? Cause that's what they do.

[B]

I mean, they Should.

[A]

That would be great for Ulysses on Blooms Day, which is in June. I can't remember when Blooms Day is, but we even have it here where they. And they do it all over the world where they'll read the entire book on Bloomsday. Because, as you know, Ulysses takes. The entire novel takes place on one day. I think it's.

[B]

How many pages is Ulysses?

[A]

It's long. It's long. You're glad when it's over. It takes him June 16th. It takes him, I think, 12 hours to read the whole thing, but almost as long as it would take to watch all four Matrix movies.

[B]

Well, so if I really. If I really cared about being a completionist, my original thought was to watch all four Matrix movies and the Animatrix, which is after the first Matrix movie the Wachowskis commissioned because they kind of based the Matrix based on anime Ghost in the Machine. They commissioned some of their favorite anime directors to do four, like, mini films exploring the other themes and backgrounds in the Matrix.

[E]

Paris, have you seen Dark City?

[B]

No. Should I?

[A]

Yeah.

[E]

This is the other movie that came out in the late 90s that's kind of like the Matrix that got overshadowed by the Matrix.

[A]

Is it animated or is it.

[B]

Oh, I love this already because it's a photo of a man, very blurry, strapped to a clock.

[E]

You should watch that.

[A]

That reminds me of that other movie you like a lot where the guy jumps off the building. That. But the clock.

[B]

Oh, yeah. Hudsucker Proxy. You hate that. That's fine. Like, it's not for everyone. I've since learned.

[A]

All right. I've never seen Dark City, a Neo noir science fiction film. That sounds good.

[B]

This is great.

[A]

Jennifer Connelly, Rufus Seawall.

[B]

It's like they built the city to make see what makes us tick.

[E]

It's kind of like the Matrix without.

[B]

The action went off.

[A]

Should I order this on UHD? Huh? 159 bucks for the limited edition one that tells you there's something going on with Dark City. Like there's a big cult. I love that.

[B]

The Hudsucker Proxy clan is mobbing up in the Discord Chat. It is a great movie.

[A]

Too weird for me. You can stream it on Prime Video. So maybe I'll stream it first and then if I really like it, buy the $160 UHD version. That's crazy. The Blu Ray is only $8.

[B]

I'm shocked. It has an UHD version, but only.

[A]

For 150 bucks, which makes no sense. That must be aftermarket or something. That's Crazy.

[B]

I think you can get it on. Oh, yeah, it's because it's on like ebay, only there's a 2 disc, 4K you can get it from.

[A]

Limited edition.

[B]

Get it from arrow video us for $35.

[A]

Oh, this is the guy who directed the crow. Can you direct crows?

[B]

Don't they fly where the wind goes?

[A]

It's murder, I tell you.

[B]

Murder.

[A]

Jeff Jarvis, you poor fellow.

[D]

I'm going to do a quick one here.

[B]

Didn't you already do one?

[A]

No.

[D]

I could do my pick line as a pick, but I won't.

[A]

No.

[D]

So I think this should have made more news. So. So, Paris, our child. Back in the day we had American television brands. We bought rca, Zenith, Motorola. I'm a little too young for Phil Co, right? Leo, did you ever have a Phil Co?

[A]

I did not have a Phil Co. You went straight to Sony. We went straight to Sony Color tv. Yep.

[D]

That's where I'm headed. So we had our American brands and then Japanese manufacturing, of course, was known for cheap stuff, including cheap transistor radios. But then Sony television came and it became the gold standard, I think, of television. It became, it became the color television you wanted to have.

[E]

Everyone had those orange dots, that orange dot sticker at the corner of their tv. Everyone had that.

[D]

Yes. Well, so this happened today, which just amazed me. I didn't see that. More news. Sony's TV and audio business has now been taken over by tcl.

[A]

It didn't happen today. We talked about it on Sunday on Twitter.

[D]

Oh, you did? Oh, well, I'm seeing news all week is all I'm saying.

[A]

TCL is now the largest TV manufacturer in the world, the Chinese manufacturer. And that's really the story. It was an American. Actually, it's the 100th anniversary of television, which was invented in Scotland of all places. So Monday was the 100th anniversary of the first television broadcast.

[B]

What did they broadcast?

[A]

Oh, that's a good question.

[D]

Oh, I'm trying to remember.

[A]

Oh, that's right, you're a TV critic. This should be right up your alley.

[E]

It's gotta be someone playing the piano or something.

[A]

Right?

[B]

Right.

[A]

I think I. I have a YouTube video. No, no. John Logie Baird's first TV set.

[D]

Yes.

[A]

It doesn't. They don't say what he broadcast with it. But in any event, Google AI overview.

[B]

Which is possibly wrong, says. Go ahead, says that it is John Logie Barrett's mechanical silhouette demos in London in 1925.

[A]

Okay.

[D]

A ventriloquist dummy named Stooky Bill here.

[A]

It Is here's a picture of the dummy and John Logie Baird. Look, I've invented television. It's great. Come watch it. I've got a dummy on the second Stinky bell is here.

[B]

Working from his Fifth street laboratory, Baird managed to transmit the first television picture with tone gradation, first using the head of a ventriloquist dummy and then moving on to a human face. Baird's subject was William Edward Tanton, 20 year old office boy. That's a job title you never get anymore. Who later recalled Mr. Baird rushed downstairs in baggy flannels, a pair of carpet slippers and no socks. He almost dragged me into his workroom.

[A]

And we're gonna put you on TV, boy.

[B]

Enormous electronic bulb.

[A]

So it's 100 years old today or Monday. So this is a perfect time to talk about this. And as you say, it was really an American. In fact, Philo T. Farnsworth kind of is celebrated as the inventor of tv. He's not.

[D]

And there's all kinds of drama with RCA involved and all that.

[A]

Yes, so far. In fact, I think there's a plaque to Philo T. Farnsworth's lab in San Francisco, because I think that was where he did his work, but it wasn't American technology. Zenith, Philco, rca. The Japanese came along, stole our business, Sony particularly, but there was also improved it and made it much better. And of course it went solid state, remember? And then, and then, and then the Koreans took a look at what Sony had done to the Americans and they said, well, wait a minute. And along comes Samsung and a little television company called Lucky Gold star, which made $90 TVs you'd find in the back of the drugstore. Lucky Gold Star changed their name to lg. And I think that's what you have is an LG Paris. And, and whoa.

[B]

It stands for Lucky Gold Star.

[A]

Well, now they say life is good.

[B]

No, no, no, no. That's huge knowledge for this.

[A]

Yeah, yeah.

[D]

Hey, come see.

[B]

I gotta look at that as a fun fact 1000 times.

[A]

But then what happened in the last decade is the Chinese saw how the Koreans stole it from the Japanese, and the Chinese undercut the Koreans with hisense and tcl. And now today, especially, I think thanks to this Sony deal, because TCL is taking all of Sony's television business 51%.

[D]

Yep.

[A]

All of the Bravias are now going to be made by tcl. They are officially the largest TV manufacturer in the world by far.

[E]

So does TCL stand for anything fun?

[A]

Oh, that's, that's a good question. What does TCL stand for? Let's ask. Chat. GPT. It stands for the tool command language.

[D]

No, wrong.

[A]

Yeah. I don't know. It's. It's originally abbreviation for the creative life according to.

[D]

Oh, that's not bad.

[A]

Yeah, it's. It's in Guangdong province.

[B]

Wow. And so the.

[A]

Sorry they were. In 2010, they were the world's 25th largest consumer electronics producer. By this year, they are number one.

[B]

The lucky of lucky Gold Star originated from the Lucky Chemical Industrial Corporation which managed manufactured cosmetics.

[A]

Wow.

[B]

Now it is the largest Korean chemical company and it was the ninth largest chemical company in the world by sales in 2021.

[A]

There you go.

[B]

It's three main business areas are petrochemicals, advanced materials and life sciences. And advanced materials sounds like the sort of vague name for something that people would use in war.

[A]

Advanced materials blow you up good. Originally, TCL was called ttk. They made audio cassettes. But TDK sued them saying you can't call yourself ttk. So they changed their name to tcl, taking the initials from Telecom Corporation limited.

[B]

Ah.

[A]

But now again, it's a little bit of a retronym. They say. No, no, that's the creative life anyway. Yeah, that's a good story. Hello, Gizmo. Speaking of gizmos, eyes are very wide. She says she does not want to be here.

[B]

She's like, who are you?

[A]

Where's my tuna fish?

[B]

She wants to show whole as always.

[A]

We will leave you with this story of the original television set. Here we go. That day. This is him in the earliest known.

[B]

Photograph of a television. That's great.

[D]

It's the Matrix.

[A]

The earliest known television. Look, it's great. It's a puppet. I've transmitted a puppet's face across the world. Oh, no.

[B]

It's going to be stuck in there forever.

[A]

It's stuck forever in my mind. Ladies and gentlemen, this concludes this thrilling edition of Intelligent Machines. We wanted to make it a two hour show, but because of you, Jeff, it's two and a half hours. Sorry. Jeff can now rest his weary bones. Will you be you. Will you be back next week, you think?

[D]

Yeah, yeah, I'm gonna call anywhere now. I've been canceled. I should be in Austin right now. I canceled.

[A]

Oh, I'm so sorry.

[D]

That's fine.

[A]

Well, I'm so glad you could be here. I am too, without you. So, Jeff, feel better. Take care of Coccyx. Coccyx.

[D]

Hey, my elder.

[A]

Your L3, which I think is connected to the coccyx.

[B]

I think it Stands for lucky industrial chemical companies.

[A]

3.

[D]

I have one departing request. Can we hear how right we are, please?

[A]

I couldn't agree with you more completely.

[D]

Thank you.

[A]

I love it.

[D]

I needed that.

[A]

Can we. Let's feed that to QN and get a digitized version of the yes Man. I couldn't agree with you, Mark. The funniest thing is I can see Paris as a six year old wandering around the house saying, I couldn't agree with you more completely.

[B]

When you're right, you're right.

[A]

When you're right, you're right, dad.

[C]

Oh, yeah.

[A]

I'm behind you all the way. Thank you for joining us on Intelligent Machines. We do this show every Wednesday at 14:00 clock Pacific Time. That's 17:00 o' clock East Coast Time. That's 2200 UTC. You can watch us live, Twitch, X.com, youTube, Facebook, LinkedIn and Kik. Or of course, if you're a club member. Thank you, club members in the Club Twit Discord on Demand versions of the show available to your favorite podcast client or on our website Twitt TV IM or on YouTube audio or video. Please subscribe. Ladies and gentlemen, I'm going to leave the final word to the yes Man.

[B]

I'm sure whatever you're thinking is correct.

[A]

Bye bye. Hey everybody, it's Leo Laporte. It's the last week to take our annual survey. This is so important for us to get to know you better. We thank everybody who's already taken the survey. And if you're one of the few who has not, you have a few days left. Visit our website TWIT TV survey 26 and fill it out before January 31st. And thank you so much, we appreciate it.

[B]

I'm not a human being, not into this animal scene. I'm an intelligent machine.

[C]

Well, the holidays have come and gone once again.

[A]

But if you've forgotten to get that special someone in your life again gift.

[D]

Well, Mint Mobile is extending their holiday.

[A]

Offer of half off unlimited wireless.

[B]

So here's the idea.

[A]

You get it now. You call it an early present for next year.

[C]

What do you have to lose?

[A]

Give it a [e-mail] Switch limited time.

[B]

50% off regular price for new customers. Upfront payment required $45 for three months, $90 for six month or $180 for 12 month plan taxes and fees. Extra speeds may slow after 50 gigabytes per month when network is busy. See terms.