IM 864: And Artemis Too - Journalism In The Age Of AI

Summary

Intelligent Machines (Audio) – Episode 864: And Artemis Too – Journalism In The Age Of AI

Date: April 2, 2026
Host: TWiT (Leo Laporte, Paris Martineau, Jeff Jarvis)
Special Guest: Kate Lee (Editor-in-Chief, Every.to)

---

Episode Overview

This episode focuses on the intersection of artificial intelligence and journalism, featuring a deep-dive conversation with Kate Lee, Editor-in-Chief at Every.to—a pioneering newsletter and AI-driven platform for tech and business writing. The discussion explores how AI technologies are reshaping newsrooms, editorial practices, and opportunities for writers, while also touching on broader themes of AI innovation, productivity tools, and anxieties around automation in creative fields. The show also covers the Artemis 2 moon launch, the leak of Claude Code, and the human dimension of adapting to new technological realities in media.

---

Key Discussion Points & Insights

1. Kate Lee’s Career Path & the Birth of Every.to

Timestamps: [02:45]–[07:06]

• Kate Lee recaps her transition from literary agent (notably discovering Jeff Jarvis) to pivotal editorial roles at places like Medium, WeWork, and Stripe, culminating in her current leadership at Every.to.
• Every.to began during the “boom” of newsletters in 2020, focusing on tech/business analysis rather than traditional beat reporting.
• The company pivoted from conventional newsletters to integrating AI tools and products—not just as content, but as core drivers of the business.

Quote:

“It wasn’t, at least initially, a big business strategy… it just stemmed from a personal obsession in wanting to understand this transformative technology.” —Kate Lee [08:18]

---

2. Every.to’s AI Product Ecosystem

Timestamps: [07:25]–[15:36]

• Products like Lex (AI word processor), Spiral (writing assistant), Sparkle (file organizer), Monologue (voice dictation), Proof (document quality checker), and the flagship “Plus One” (AI agent for productivity) all emerged organically from team needs.
• Every.to positions itself not for hardcore coders but for “builders” of all backgrounds, including non-engineers leveraging AI for real-world tasks.
• The platform combines content (newsletters, podcasts), tools, consulting, and interactive events/workshops.

Quote:

“Our head of growth is all in on Claude Code... he’s not an engineer, but has automated dashboards and agents who help him do his job.” —Kate Lee [13:47]

---

3. Editorial Practices: AI as a Writing Partner

Timestamps: [19:08]–[27:54]

• Every.to does not require writers to use AI but encourages experimentation; adoption varies widely among contributors.
• AI tools are integrated in processes like editing, drafting, and style enforcement, raising the “floor” for quality and freeing up editors from repetitive tasks.
• The platform boasts a 400-rule style guide, which has been adapted specifically for consumption by large language models (LLMs).
• Emphasis remains on human discernment: AI drafts serve as input, but the writer’s judgment is paramount.

Quote:

“Your job as an editor or writer using AI is... not to blindly accept the recommendation, but to wrestle with it.” —Kate Lee [21:21]

---

4. Journalism, Authenticity, and Controversy

Timestamps: [27:57]–[36:00]

• The panel reflects on the tension between AI-generated content and authentic journalism.
• Paris Martineau notes that while AI can speed up rote tasks (e.g., standard emails), “edge-sanding” is a fundamental flaw—AI outputs tend toward the average and risk erasing uniqueness in writing.
• Kate Lee addresses concerns about job displacement, reiterating AI’s potential to unlock new roles and augment, not replace, human expertise.
• The group agrees that AI is best seen as a tool to automate drudgery, not to supplant original reporting or creativity.

Quote:

“If I never have to copy-paste into different forms again, I will be so happy. But it’s really the AI and the human expertise together.” —Kate Lee [39:19]

---

5. Community & Interactive Learning

Timestamps: [41:24]–[43:48]

• Every.to organizes regular “camps” and workshops on AI tooling, in partnership with companies like Notion, aiming to foster a builder community.
• Subscribers get access to products, events, and ongoing support.

Quote:

“I’d say it’s about community. You’re building an Every.to community of people interested in using these tools... it’s a very interactive space.” —Leo Laporte [41:50]

---

6. Broader AI News & Industry Trends

Timestamps: [54:45]–[77:58]

• Claude Code’s Source Leak:
  The leak and subsequent takedown attempts highlight both the open-source ethos and the risks when powerful AI tools escape into the wild.
• Rising Costs & Exclusivity:
  Conversation turns to whether the most advanced AI agents might soon be priced out of reach for all but the wealthy.
• Americans’ AI Skepticism:
  Surveys reveal most Americans (except high earners) believe AI will do more harm than good—paralleling skepticism seen during the rise of social media.

---

7. Journalism in the Age of AI: Soft Skills & Human Value

Timestamps: [108:14]–[115:15]

• All agree on the enduring value of human judgment and the importance of learning foundational skills before automating them.
• Newsrooms that rely on AI-generated aggregation risk eroding entry-level training grounds for journalists.
• Commodity writing may be replaced by LLMs, but the true “work” (original reporting, synthesis, creativity) must remain human-led.

Quote:

“What you’re learning is the pattern recognition, the rhythm of it... and offloading that to a tool just doesn’t do you a service.” —Paris Martineau [108:57]

---

Notable Quotes & Memorable Moments

• On Collaboration:

  “We really are targeting builders who are on the bleeding edge of AI…not just coders, but people crossing that membrane.” —Kate Lee [13:04]

• On AI as Editor:

  “The style guide, in a way, could be a competent editor—at least on the technical side.” —Leo Laporte [31:35]

• On Job Security:

  “The prompt engineer was not a job a few years ago. Now, it’s more a skill that a lot of people have…” —Kate Lee [38:31]

• On Writing Authenticity:

  “If you’re talking about using AI to produce the core part of your output… you’re going to get something that is sanded down at the edges.” —Paris Martineau [48:42]

• On AI’s Creative Limits:

  “Creativity is breaking the patterns. AI is never going to do that.” —Jeff Jarvis [52:54]

• On Automation’s Impact:

  “AI layoffs create entrepreneurial opportunities... but the reality is most people don’t enjoy their jobs.” —David Sacks (as referenced) [100:00]

---

Important Episode Timestamps

• [02:45] — Kate Lee’s early career and her connection to Jeff Jarvis
• [07:25] — The pivot to AI products at Every.to
• [13:04] — Target audience: “builders,” not just coders
• [19:08] — How Every.to’s writers and editors use AI
• [21:21] — Wresting with AI recommendations, not blindly accepting
• [27:57] — Authenticity in journalism; concerns over AI smoothing of style
• [39:19] — AI, job displacement, and future-proofing through adaptation
• [41:50] — Building community and offering interactive AI learning events
• [54:45] — Claude Code leak and security implications
• [61:49] — Public skepticism and wealth bias in AI benefit perception
• [108:14] — Human learning and pattern recognition in journalism

---

Additional Highlights

• Artemis 2 Moon Launch: The hosts share communal excitement at the first crewed moon mission in decades, connecting the dawn of space exploration with today’s AI tools ([65:01–90:43]).
• Industry News:
  • OpenAI’s new funding round and valuation ([72:22])
  • SpaceX IPO discussion ([74:02])
  • Security breaches in common AI libraries (Light LLM, Axios) ([62:55])
• Ethics & Gatekeeping: The International Conference on Machine Learning “honey-potting” its reviewers with hidden instructions to detect LLM use ([101:54–104:58]).
• Journalism Entry-Level Concerns: Analysis of hollowing-out career paths due to aggregation/junior roles being replaced by machines ([108:14–112:41]).
• AI in Daily Life: Polls about work preferences (15% prefer an AI boss), and how primarily white-collar high-income workers foresee AI as beneficial ([97:15–99:42]).

---

Episode Tone & Style

• Conversational & Candid: The hosts and guest blend humor with sharp critical insight, never shying from addressing skepticism or ethical gray areas.
• Optimistic but Clear-Eyed: There’s celebration of AI’s ability to free up creative energy (and clerical time), but persistent reminders of what’s lost if automation replaces basic training or original thought.
• Community-Oriented: Strong emphasis on building thoughtful, interactive AI user communities—whether for journalists, “builders,” or everyday users.
• Self-Reflective: The episode closes on personal anecdotes (tattoos, vintage sounds), underscoring the core theme of technology’s role in supporting rather than overwhelming human purpose and culture.

---

Listen to This Episode If You...

• Want a ground-level look at how newsrooms are integrating (and wrestling with) AI
• Are curious about practical, non-coding uses for new AI tools
• Enjoy thoughtful debate on authenticity, creativity, and the human value-add in digital media
• Like a blend of hard news, product reviews, and personal banter—plus a historic moon launch for good measure

---

Further Resources

• Every.to — https://every.to
• For a taste of AI-assisted productivity: try, waitlist, or read more about their suite of products (Plus One, Spiral, Lex, etc.)
• Semaphore’s feature on NYC “runners” in field journalism for more on the tradition of ground reporting

---

Notable soundbites, quirky tangents (April Fools pranks, obsolete sounds, tattoos), and context omitted here for brevity but richly present in the full recording!

Transcript

[A]

It's time for Intelligent Machines. Jeff Jarvis is here. Paris Martin. Oh, two. Our guest, Katie Lee, is the editor in chief of a very interesting new newsletter, slash AI site called Every to. She says AI and writing go together like cheese and crackers. We'll also talk about new rules from the governor of California, 15% of Americans who say, yeah, I'd work for an AI boss. And the big Claude code leak. All that 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 864, recorded Wednesday, April 1, 2026, and Artemis 2. It's time for Intelligent Machines. Hello, everybody. This is the show where we talk about AI robotics, all those smart doodads all around us. Let me introduce Paris Martineau, investigative journalist at Consumer Reports. And you're. Why are you wearing a retro baseball hat? Is it because opening day happened or is just.

[B]

You know, I thought about. I thought you were going to ask me about this. When I put on the hat, I just haven't brushed my hair. And so I was like, I need to wear some sort of hat. But literally, as I was logging on, I was like, is it strange to wear a hat while recording a podcast? And I think, no. But I don't know.

[A]

You'll be the guest P for podcast.

[C]

What's the P?

[A]

Maybe that's it. You're the podcast.

[B]

The Pittsburgh podcast people often ask me. They're like, oh, are you a Pittsburgh fan? I'm like, no. Someone just bought me this hat because it has a P on it. My name is Paris.

[A]

Oh, it's the Paris hat, of course. There you go. Anyway, great to see you as you're trouncing me one more time and cross that cross play thing.

[B]

Hey, you thought I was going to win this one. You should have considered yourself lucky when I wasn't responding. That was your time to strategize.

[A]

Oh, man, you just clobbered me with your last play or two plays ago. Also here, Mr. Jeff Jarvis. He is the author of the Gutenberg Parenthesis now in Paperback magazine, and of course, the new one, Hot Type, which is on pre order at jeffjarvis.com, the story of the line of type. And it is a drama. He is, of course, the emeritus professor of journalistic innovation at the Craig Newmark Graduate School of Journalism at the city of the University of New York, Newark. Ah, but you know what? We have a very special guest today because she discovered Jeff back when he was an unknown Unsung blogger Kate Lee is here. Hi, Kate. It's great to see you.

[D]

Hi. Thanks for having me.

[A]

At the very young age of 24, you became a literary agent famed for bringing bloggers. And this is what this is, early days of blogging.

[D]

Right in the very early days

[A]

to the literary world. Now editor in chief of a very interesting site called Every at Every to, which is for every two, I guess.

[D]

Yes, yes.

[A]

Great to have you. And I really should let your. Discover your protege, Mr. Jarvis, handle this interview because.

[C]

No, no, no, no, no, no, no, no, no.

[D]

I'm the protege.

[C]

Oh, no, no.

[A]

He's your. You're his.

[B]

He's.

[A]

Well, I. No, Jeff was known beforehand as a TV critic.

[C]

Well, well, it's actually. It's because of Kate that this show that I came to you, Leo.

[A]

No, really, how did that happen?

[B]

Really?

[C]

Well, because Kate was my agent, so. So I talked to Kate about. About books, about this technology thing. And originally the original idea was when I had interviewed or I had heard Mark Zuckerberg say that you should bring elegant organization, people's lives. Kate thought, well, maybe there's a book in that. And I tried to start writing, I thought. And then as we're talking, I said of, you know, what would Google do? And that became the book which Kate sold to Harper Harper. And then it was because of that book that Julio laporte called me and said, why don't we do a podcast about.

[D]

Because we were going to do a

[A]

podcast called this Week in Google. And well, clearly the guy who wrote what Would Google Do? Should know a little bit about it.

[C]

And then Kate went on to other. I'm fascinated by Kate's career because she went on to other things that tied culture and technology. She was the editorial hire by Ev Williams at Medium.

[A]

The very first.

[C]

The very first. And. And dealt with a whiplash of EVs changing ideas every morning. And then content at WeWork, which was an effort upon them, and then at Stripe as the publisher of its book outlet, and now at every. And so, Kate Rashid. I haven't talked about this, but I think it's really interesting to see how you're trying to bring culture to technology or technology to culture, one way or the other. How well do they mix?

[D]

It's an interesting question, especially now, where I feel like technology is culture in so many ways, and technology occupies a place in popular culture more now, perhaps, than it ever did. I always thought of what I was trying to do was. Was like, I loved media. I love the media world. I love the Tech world. How do I bridge those in a way that, that worked for me, that I found, you know, interesting and stimulating and satisfying? I think what is, you know, again, where I really see it now is. And it's not just that it's, you know, in 2026, but that, that, you know, technology is culture. And you see technology companies knowing that they have a place and they have an audience and that they have things that they want to say and feel like they can say, and they want to say it to an audience and it's a really interesting time.

[A]

Let me ask about Every, because Every didn't it just started out as newsletters, right.

[D]

Every was founded in 2020 in the boomlet of newsletters, starting with Substack and with the Pandemic. And it started really business and technology writing, high quality business and technology writing. And the aspiration was to become really an institution for really great writing on these topics. It was never sort of scoop driven. It was never about beat reporting. It was about analysis, commentary and insights. First person insights from practitioners and the builders who are in tech.

[A]

But at some point there was a little bit of a pivot which must have been pretty controversial. The focus became AI and your content is AI generated. Well, human plus AI. Right. Facilitated, facilitated, powered by OpenClaw. Oh, wait a minute. This is the new product or something?

[D]

Yeah, that's the latest product.

[A]

Yeah.

[D]

Oh, are you on the waitlist? Did you get on the waitlist?

[A]

I did not get on the waitlist, but I am a subscriber to every Point two. So there are lots of articles chiefly focusing I think on kind of AI use in enterprise. In fact, you have a consulting business doing that. You know, articles like build your own Bloomberg terminal with AI.

[D]

Which is, which is the Bloomberg terminal. Out of my dead hands. Yeah.

[A]

So. And. But the fact that you not only have AI Assist, but embrace it and it's part of your kind of code of ethics and everything is, I think for some kind of shocking. That's not journalism. How do your writers feel? Well, they must love it, otherwise it wouldn't be there, right?

[D]

Well, it depends and I think it's not necessarily one size fits all. I will say we pivoted to AI. It wasn't, at least initially, it was not a big business strategy. It was just because our co founders and CEO Dan Schipper, who you saw before, you know, GPT3 came out and it was mind blowing and he was incredibly fascinated by it and he is a writer as well as a CEO and founder and you know, coder and a podcaster. And podcaster. He wears many hats and he just decided for himself that he wanted to spend three months going deep on ChatGPT. That's just what he wanted to write about. And that's also when we started the podcast, the podcast that he helms called AI and I. The initial name of the podcast actually for about, I don't know, the first six months at least was called how do you use ChatGPT? And it literally was just talking to people and it was Tyler Cowen and, and Jeffrey Litt and you know, an amazing array of guests. But it was, you know, it was just focused on ChatGPT and that. So it really stemmed from a personal, a sort of personal obsession and personal interest in like this technology is transformative. I want to know about it, I want to understand it, I want to, I want to know how it works. I want to. And I can see that it's going to have huge implications for how, for, for technology and how we all work at the same time. We had also, and this is before I joined the company, but we had basically built our, our first like AI driven product called Lex, which is an AI word processor. And it was, it, you know, Dan and our other, other co founder, Nathan, they are, you know, they are, they are not. They don't consider themselves engineers, but they, they know enough that they could build this product and it became very clear that there was a future in which you could build stuff with AI and they were just able to build it. And because every started as a media company, we have a distribution list. We have spent years cultivating a list of subscribers. So when it was time to release that product to say, hey, who wants to try this out? Get on a wait list? We pretty quickly racked up tens of thousands of subscribers on that wait list. And it became very clear that that was going to be a model going forward.

[A]

Yeah, so that's one of the things that's a little confusing because it is articles. So it's like it's a newsletter, but it's also a podcast, but it's also products and it's also consulting. So it's a whole bunch of different things. I guess this is maybe the new way of being. It's not just one thing. You have a. So were the products developed in house for your purposes in house and then released to the audience? Is that the readers? Is that how you did it or.

[D]

Yeah, for the most part, the products kind of sprung from an individual itch that, you know, that, that people internally were feeling Initially it was. Was Lex, which was the word processor. I believe the next product was Spiral, which was the. Which is our writing assistant, which Dan coded out of a. On like one of our sort of down weeks at what we call them think weeks, which is where we. We don't publish. And we came to think and plan.

[A]

Kate is in the West Village and that's the local fire department rushing over to Saul Hanks for a French dip.

[D]

Yeah, sorry about the siren in the background.

[A]

That's quite all right.

[D]

And then Sparkle was our next one, which is a file. An AI file organizer. And again, these kind of sprung out of just individual needs that we. And we could prototype them. We did end up then hiring people who, you know, we can't manage all of those products ourselves, but they each essentially have one person managing those products. We have brought people in who do that.

[A]

This is so different from the AI product culture I'm used to, which is primarily GitHub. So drive you to a GitHub page, it would have some GitHub install instructions written by Claude or open ChatGPT or somebody. And this is beautifully laid out. And it's, it's, it's like a product. You're obviously not aiming at the AI fanboy here.

[D]

Why do you say that?

[A]

I just doesn't. It doesn't look like it is a. I mean, the, the AI geek code. It's not aimed at geeks, it's aimed at real people, I guess.

[D]

Oh, for sure, for sure. I mean, I, I think the way that we think of our audience is that we really are targeting builders who are on the bleeding edge of AI. And that word builders, I know, is one that has gotten a lot of play. The Wall Street Journal just did a great article about it. And I think the way that we think about it is, you know, typically, or the connotation is you're an engineer, you're a coder. And sure, that's maybe where it started and certainly where some. These coding, these, excuse me, these chat tools, these LLMs have really found product market fit between cloud code and codex. Clearly coding is a really foundational use,

[B]

but

[D]

really what we are observing internally and really what we share internally out to our audience is that pretty much everyone here is a builder in some way. Our head of growth has basically is all in on Claude Code. He is not an engineer, he is not a coder. He calls himself dumb in that regard, but he has automated dashboards and agents who help him do his job. Basically, our operations person has done the same thing. Our Customer service has done the same thing. So that's really how we think about it. I also think when we think of ourselves as the one subscription, you need to stay on the edge of AI. And that includes the ideas, which are the content and the pieces. The podcast, it includes the products, we named a few of them. And it includes trainings, which are consulting. And then we also offer course courses and camps to our paid subscribers. So we are doing a camp with Notion on Friday, and we've done camps with Codex, we do camps on openclaw, things like that. And I think we're really trying to target those people who are builders and who want to be builders and who are sort of crossing that membrane. Because honestly, the way that every operates, I can't necessarily say in every possible way, but certainly from a AI native perspective, the way that companies are going to be created and built in the future. And so we feel like we have an opportunity to sort of show what that is, how that works, how we're doing, including, you know, clearly the struggles and all, and bring people along on that journey.

[C]

So you make. And eat the dog food. Yes, inside. But what. Also, because you're an editor, you're. You're a culture person. And Kate and I, we hadn't seen each other for far too long. We ended up on a panel before ghost writers, ghostwriters, believe it or not. There's a Ghostwriters association, fairly large.

[A]

You talked about that when you were there.

[C]

They were fairly frightened by this, by these prospects. Well, ghosts are scary talk, Kate, for a minute, because you've got people who aren't coders, but now code, as opposed to a Leo who was a coder, who now says, oh, my God, this changes my world of coding. You have people who, who, who didn't code, who can call themselves idiots at it and then do it. And then you have writers and readers.

[A]

And I am going to confess to something, and maybe it's just me, but there's a bias among coders who are now using these tools that no one else can really use these tools.

[D]

Right.

[A]

Really can get it.

[C]

See, that's the problem.

[A]

And. And so this feels.

[B]

You're saying that coders and techies are being exclusionary.

[A]

Yeah, I know. That's shocking. I know.

[C]

For white guys, it's just so amazing.

[B]

I mean. Yeah, they've never done that before.

[A]

So, for instance, I mean, the first thing I saw when I went to every to is. Is your kind of your. I guess it's a. It's a clawbot. It's Open claw.

[D]

It's a plus one.

[A]

Yeah, yeah. And I thought, well, I could just do that myself. So why. So what's the pitch here? Is it for people who couldn't. Can't do it themselves or is it. It. What. What makes this better than just vibe coding your own thing?

[D]

Right. Well, first of all, not everyone is going to vibe code their own thing. Right. Secondly, it can take some time to get them set up. I mean, I'm in a. I'm in a daily conversation with my claw in telling. Training her to do things that I want her to do. And I call. Yes, I call her.

[B]

Her.

[D]

And. And it takes time. It takes time to set that stuff up. It takes.

[A]

It does. It's a huge amount of time and unbelievable.

[D]

A lot of time. It takes also money to provision it and to.

[A]

These.

[D]

These are. These are very expensive. They eat a lot of tokens. And so we are basically doing that on behalf of our subscribers. The other thing is that we are. We are, you know, plus one doesn't exist in a vacuum. It exists among our other apps. It exists among our content. So can basically you get a plus one and you are able to be integrated with Quora, which is our email management service. Spiral with writing, Sparkle for file organization, Monologue for. For voice dictation. I just lost. I think I just lost track of the last one. But essentially it's an. It's an ecosystem. And so your plus one is sort of coming loaded in with. With the every. With the every products and with. And we ultimately want to be having like, every skills that come sort of preloaded into this plus one.

[A]

And that's, of course, the thing that scares most people about OpenClaw is that it's so insecure and they're very worried about not only the insecurity, but how much it's going to cost. It's Spiral, Chorus, Sparkle, Monologue, and Proof.

[C]

Sleepy, Dopey, Grumpy.

[A]

And then plus one is the new thing.

[D]

Proof is also. Proof is also new. But yeah, those are the two most recent.

[C]

How do your writers and editors use AI? Paris is a journalist who writes.

[D]

That's a great question.

[C]

I'm curious what the process is like and what they find useful and not.

[D]

Yes. Well, first I should just say, Leo, I think you said a little earlier that you may have said that all of our writing is AI or referred to it in some way. That is not the case. But we do use AI in our writing and our editing. I will say that it's different for everyone. One thing that's just been really interesting to see is that everyone has a really different process, even if they're using similar tools. People just go about it really differently.

[A]

Some people write something first, then feed it to Spiral. Or some people would give an idea to Spiral and have it do. I can imagine that. And that's fine. That fits, you know, your style. But it's. So it's. With a partner.

[D]

Yes. We're actually publishing a piece on Monday, I believe, by our writer, Katy Perrott, who is a writer who's basically like, figured out and built these tools herself and kind of very extensively goes through her process. She's written a lot about her process, but in particular, I think there's been a discourse recently of, is AI writing real writing? And she sort of is trying to address it. She has quite an extensive process. She starts off always with her agent interviewing her. She has a panel, Leo, you referred to with Guilfoyle and Dinesh. She has a panel of 10, I don't even know how many, but 10 different agents of, like, Sedaris is for humor, you know, Hemingway is for Bremedy. She's named all of them. And she's basically always sort of pressure testing her ideas. She's never in the use of AI. We're never just with anything, whether it's writing or anything else that we do. It's never just like, oh, it spit this thing out. And so now we take it. First of all, everything we, you know, with writing and with editing, it's trained on our stuff. So it's not generic. It's been trained. We've trained it on things that have worked. And I think then the other thing that I say and that I think Katie really embodies in what she writes is like your job as an editor or writer using AI is, as I said, not to blindly accept the recommendation, but it is to wrestle with it, because again, it's not generic. If you've spent hours training this claw or LLM or whatever it is on your work, on your style guide, on your best stuff, on the things that you're aspiring to, it's not inventing things out of thin air. It doesn't mean you have to accept everything it says, but it is a good thing to be considering.

[A]

The thing that I find very interesting, all gatekeeping with notwithstanding, is that coming at it from a different point of view, instead of a technologist or a coder coming to it from that point of view, you're coming from a humanity's point of view into the Technology and doing Vibe coding. And in some ways, that's what's exciting about Vibe coding is it. It opens the door.

[C]

It finally opens it up to.

[A]

For anybody to create tools. You have a 400 rule style guide, right? Tell me about that.

[D]

I love a style guide. You know, Paris, you may. You may have them in your world, too.

[A]

I have APs over here, but I paid too much attention for them.

[D]

I feel like every job I've had, I've always been so excited to be like, let's set up a style guide.

[A]

Tell for people who aren't journalists, what is a style guide.

[D]

First of all, a style guide is a set of essentially rules and standards of how you are going to use certain language. And it can be anything from grammar, here's how we use commas, here's how we use semicolons, or don't use semicolons to. Here's how we refer to people. We always refer to someone as their full name on first Referral, and then Mr. Someone on the second referral. Things like that. There's any number of them, words you always use, words you never use. It can really encompass an enormous amount. And so I created that, that when I started. And I mean, this was several years ago, so this was pre AI. I wrote this, but we basically have fed that into Claude or into whatever the LLM of your choice to basically be like, these are your instructions. We've also had to actually rewrite it because it was initially written as a person is going to be reading this. But. But it, you know, for an LLM to read things, the format needed to be slightly different. So we did have to. We did have to adjust it in that way.

[A]

But interestingly, still kind of prose. I mean, that's what's kind of interesting about skills for these AIs is they are English, they're not code.

[D]

Oh, for sure. This is all natural language, as they say. And I think when you, you know, from what you were saying about coming into it from the humanities perspective, I think the way that we think about it is, is we have editorial standards that we aim to meet or exceed ideally. I think we try every day to do that. We're a small team. I'm not going to be able to hire a whole raft of people. But I do have standards that I want to apply across the board. And I think one thing that became really clear with AI and of course, with its ability to sort of of do pattern recognition and automate some things that are very repetitive, is that I, as an editor Were sort of constantly coming upon the same mistakes every single time. And it's like, I don't want to have to keep correcting this. And again, I'm not saying this about deep thinking about the nature of an argument necessarily, but even just the same things in a piece again and again and again, like, I should not be spending my time on that anymore, nor should anyone on my team be spending their time on that anymore. And so that's the kind of thing that it can enable you to free you up from some of that and also then to apply those standards a little bit more evenly across every. It essentially raises the floor for me of, like, the quality level of what I was getting back from my editors.

[A]

You use the word taste, though, which is a word I would never use with AI. You can give a. And AI can have. Can have taste.

[D]

I don't know. I mean, I don't know that I. Taste is not a word that is necessarily in. In my vocabulary all the time. It's certainly a word that is in the tech vernacular at this point.

[A]

Well, spiral, it says on your site is your AI writing assistant with taste.

[D]

I see. Sorry. I appreciate.

[A]

I presume you don't mean like flavor. You mean like good taste and bad taste. Right.

[D]

Well, essentially, it's been trained on like, every's editorial standards. So it's a way of. So, you know, it's a way of saying, like, here's what we do, here's how. What we do to get the best writing, we hope, out of the people that we work with. And it is now imbued in this product.

[A]

It says spiral writes with natural rhythm, concrete details and clear language. Good taste built in from the start. Yes, I guess that's good taste. And AI can do that. You find this reliable?

[D]

I use spiral every day. Again, what? The way that I use it is the way that I use it and is probably different because in terms of how other people use it, I use it for things like if I have to be writing a marketing email or I have to be, you know, doing something that is not, you know, the highest stakes, but I just want to get it out.

[A]

So it's the writing you don't want to do that you use this for.

[D]

That's how I use it again, because

[A]

my feeling is I like to write. I don't want it to write for me. But I guess you're right. I wouldn't want to write a marketing email.

[D]

Yeah, I never want to write a marketing email.

[A]

So I understand that.

[D]

And what I think is really Useful with Spiral is that it actually returns. What it returns is essentially three drafts from three different angles. And it basically allows you to be like, I like that angle or I like those two. I want to go in that direction. And then you're sort of constantly pruning and constantly sort of talking to it. Again, that's not to say that people. We know from our data that people are actually using Spiral to write books and to write articles and things. So the use cases are individualized. This is just how I use it.

[A]

Paris, would you use something like that?

[B]

No, but, you know, I think to each their own. And these are.

[A]

You like to write though, right?

[B]

I like to write. And I also. I don't know, I think I. There's been a lot of debate over the last week and journalism Twitter about various tech journalists or other journalists kind of coming out to using AI for some or all of their work. I think kind of where I landed, at least for like reporting in journalism, where the writing is kind of the output that you are trying to get people to give you money to produce. Like, and there's a lot of cases where it's going to. It is something valuable to be able to produce authentic and interesting and unique piece of writing. Which I think one of the downsides of AI just based on the way that it works is that of course everything's going to be a bit smoothed. It's going to, it's. It's something that comes from the result of training on a large data set. So of course it's going to tend towards averages. But I think that's. That's not to say it couldn't be useful for like sending an email. Like, I think my. The one instance where I have used AI for writing, quote unquote, is like, if I have to slap out a bunch of request for comment emails and I don't have like literally do not have the time to write all of the sentences to episode, write every bad version and then check that. But I think. I'm sure there's a bunch of different kind of examples like that where people's writing output is more rote.

[A]

Do you get pushback over this, Kate? I mean, I imagine you do.

[D]

Pushback from whom?

[A]

The writers. Writers, editors.

[C]

The reaction to the Cleveland Plain Dealer. The reaction to, yeah, look, there's a

[A]

good example of Cleveland Plain Dealer said we're going to use AI and everybody was,

[D]

yeah, I just, I don't really can put. I don't really think of ourselves in that discourse. Or maybe I just, you know, we just worry about ourselves. And I don't really worry about what. What the others are saying. I think what. Where in getting adoption for some of these tools across the team, where if there was pushback, it was not. Again, for. For every writer I actually am not. Would never mandate to a writer. You have to use AI to write. It's like, if it helps you, if you find part of your process, great. If, you know, some of. One of our writers always writes first. You know, he writes the draft himself before he does anything with AI, you know, whereas Katie will do an interview with AI first. So it really, really depends individually on each person. What I think was really interesting, you know, about a. Maybe a year to nine months ago, was in sort of using our. Using our editing tools to essentially be like, you know, we had the style guide, we trained it, we train the LLM. We have like, a whole bunch of examples of, like, headlines and leads that really work and all sorts of things. And it really required incorporating that check as a part of our process. Not being optional took some time because you just think like, well, okay, I would get a draft and I'd be like, well, I read it and I'd be like, I think I see five things that I, in a significant way, would want to change about it. Did you run this through? Did you check it at all? And if the answer was no, I was like, do that and then send it to me. Because it's essentially catching the things that I would catch. But now I have to do it. And so that just took a little while.

[A]

So the style guide, in a way, could be a competent editor, at least on the. The technical side of it. That makes sense.

[D]

Yeah.

[C]

When you were at Medium, you were recruiting known writers to come there, and one of the services that you offered, this snob that I am. I just want. This was editing help. And. And a lot of people took you up on that at the time. Is that kind of the same motif here that. That you're offering things that can be helpful to writers if they wish? I'm hearing that from you, from your internal writers. What are you seeing for people who are using these tools for things outside of every. What are you hearing in terms of feedback from writers elsewhere?

[D]

Gosh, I don't know. I mean, I don't know. I don't know that I have a reliable handle on how others are saying it outside of. Again, I've read some of the same news reports and read the discourse that Paris was referring to. I don't.

[C]

I mean about using every's tool specifically.

[D]

Oh, I'm sorry, these tools aren't used

[C]

just for every writers. Right. They can be by anybody who's a member.

[D]

Yeah.

[C]

So who do you find is the, is the target for that?

[D]

And are you speaking specifically about the writing tool or other tools?

[C]

No, I guess all of them. As a writer, I think I start there, but yeah, yeah.

[D]

Well, I think they are people, you know, in general who, who have, who productivity is something that they are in general interested in because these are all tools that are meant to help you in some way be more productive or however you might define that. And maybe that's better writing. Maybe that's spending less time writing. Maybe that's spending less time on a first draft and instead getting that first draft out and then spending the time refining something. So we've, we've, you know, again, I think it's a, it's very individual in how people use it. I do think that there is a lot of. So for instance, when Spiral started, when Dan built Spiral, it was like, well, you know, we, we create podcasts every week. You create a podcast every week. But it's not just the podcast. It's the Tweet, it's the LinkedIn post, it's the article, it's the discord, it's the this, it's the that. And like, that all just takes time and time that we wanted to spend doing other things. And so we certainly know that in cases, again, that was the use case that Spiral came out of. But that is one way that I don't want to be spending my time, as I said, writing marketing emails and things like that. But we're seeing use cases of all across. Again, the data is showing people are using it for books, people are using it for proposals, people are using it for, you know, LinkedIn posts, which is probably no surprise, just what we all see on LinkedIn. But people are really, you know, the use cases are across the board.

[C]

What was your AI learning journey? What, what has it been like?

[D]

It's a great question. I actually was just on our podcast a couple of weeks ago talking about it because I have traditionally I, I, I love to edit it. Like I, my happy place. When I'm like, I have a million things to do, it's like, oh, maybe I can just edit this piece. That is, that is where I feel like I can just do my thing and I, and I know what I'm doing and it's a measure of control and I get a lot of satisfaction out of it. So this was this Was like a, this was a journey for me as well. And I should say I read and edit, top edit every single piece that goes out. So, so that, that is absolutely a core part of my job. But I also have to spend a lot of time, you know, I have a small team. Up until like 4 months ago I had, I had like half a person on my team. You know, I didn't even have like a full time staffer. How am I going to get all this stuff done? I had to hire, you know, I was, I was hiring a bunch of people, had, you know, was glad to be able to hire a few people, had a bunch of open roles and I was essentially, you know, this is one case sort of outside of writing. I should just say that I, that I, where I really saw the potential. But like I don't want to spend my time wrangling notion. I don't want to spend my time in settings, right. You know, setting up pages and connecting databases. Like that's just not where I shine. There are other people for whom that is second nature. It's not for me. And so if I can essentially work with, you know, whether it's cloud code or at the time it was OpenAI's Atlas to essentially just tell it to do it. That frees up an enormous amount of time for me from the writing and editing perspective. Jeff it's interesting because Dan, our co founder and CEO, he's been saying basically since GPT3 came out, that he really wanted to automate the rote parts of my job, like the copy editing. He just was like, I just don't want to see you deleting commas or whatever it is. It's just a waste of your time. He was trying, and this is before tools that allowed you to vibe code really came out, but he was really trying to build something like an internal copy editor that would do it. And every time, I mean this was several times over the past couple years, he'd be like, I built it, I want you to try it. And then I would try it and I would be like, yeah, it didn't catch all of these things. And so we're not using it. And so it, but it's sort of been like every three to six months there's been a different version of that that he would present to me. And I would be like, it would get better and better. But I would still, I mean I also still am like Google Docs is like my, you know, don't take Google Docs from me. But, but, but, but so it was. It was for me also like a. Okay, this is getting better. This is getting better. It's not great. Yeah, this is getting better. Okay, there. There is a point at which. And it really was last year sometime. And then of course, with. With. With anthropic releasing opus November 24th is when it happened.

[A]

Yes, I know.

[D]

Was the big. The big moment where. Where it was a clear tipping point of like, yeah, oh, if I don't do this, it's actually going to hold me back.

[A]

Do you worry about jobs, though? I mean, I often think I could probably at some point, not yet, automate the entire workflow for this podcast network, except for my wife, because, you know, but everybod. But those people, first of all, I really enjoy working with them. I like having them as colleagues. And yeah, maybe I could automate their work, but seems like that's kind of a grim future for a lot of people.

[D]

I don't necessarily have the answer to that. I do think that what it has enabled and what I think it ultimately enabled and what technology has done is. Is opening up new jobs. New jobs that maybe don't even exist yet. Different kinds of jobs. I mean, the prompt engineer was not a job a few years ago. And again, I think prompt engineering is more of a skill now that a lot of people have rather than job. But again, that's just an example of something that came out. I don't think it's going to be easy, but I think that if you embrace these tools in ways that are comfortable to you and are germane to how you work in your world, because I really think it fits. It's not just the AI. It's the AI and the human expertise. And that's where you're seeing so much of this, so much of these tools and so much of these gains being made. And so I do think that ultimately there will be new opportunities and new things for people, new things for people to be doing. If I never have to copy paste into different forms again, I will be so happy. I just never want to have to do it again.

[A]

Every two is the website. We're talking to the editor in chief and we're so glad to have you on, Kate. And by the way, you see, I signed up for the wait list for plus one.

[D]

Okay.

[A]

A lot of interesting tools on here. This is a very interesting site because, as I mentioned at the beginning, it's a newsletter. Yes, it's columns, yes, it's a podcast, but it's also products like voice dictation, you know, helping you write all sorts of stuff, including someday soon. Are you actually. Are some people set up with +1 now or is it not out yet?

[D]

So internally we're set up with plus ones and we're just now letting some people off the wait list. But it's just the product wasn't quite ready to let people off until now. But we're starting to too. And we've had some beta testers and of course that's just been really useful to get feedback.

[A]

And some of these tools you get with your subscription. I'm going to download and install monologue for sure, which is the voice dictation and the automatic file organization. Sounds pretty darn cool. And the email assistant. God knows I need that. So there's a lot of interesting. So it's tools. Oh, and I didn't even mention you're doing these events. In fact, there's one coming up in a couple of days and these are all virtual so you can attend them from wherever you are. This is every and Notion, a custom agents camp. But look at all the events you're going to do. Claude Code for Finance Writing camp. Claude Code for Beginners, openclaw Camp. There's a bunch of stuff coming up and it's your team that leads these. Or do you bring in people?

[D]

Yeah, no, I mean, I basically oversee the camps and I work with a colleague to run the courses as well. So definitely wear a lot of hats. The camps are free for paid subscribers and the courses are for a fee, although we have discounts for paid subscribers. But yeah, it's a lot. But it's great. It's a lot of fun.

[A]

If I were going to put it under a dome, I'd say it's about a community. It sounds like you're building an every two community of people who have an interest in this and who want to use these tools. And it sounds like it's a very interactive space to participate in that and that I love. I think that's kind of what podcasting was all about from the beginning, but we've kept it to just kind of a narrow slot. You're expanding.

[C]

Get busy, Leo. Start VOD coding.

[A]

Well, I need 20 assistance. Maybe plus one will fix that. I don't know.

[D]

You can get a few.

[B]

I do have to ask before you leave, what's up with. With the horse head behind you?

[D]

That was, you know, that was. Came up earlier when I just hopped on. That was a leftover from Halloween that I think we had a little Halloween party. And then we very recently in like the past, I don't know, two to three weeks actually made this into a real podcast studio. And so we felt like it needed to have a place of honor.

[A]

I couldn't refuse.

[B]

Delightful.

[C]

You could donate your log, Paris.

[B]

I. Do you have a log? A fake log.

[A]

It's a long story. Part of a Halloween cost. Yeah, she was a log lady. Caitley, thank you so much. Thank you for discovering Jeff. Appreciate that, actually, really, seriously, thank you for introducing us to Jeff. Kind of in an indirect way, but that's been a very fruitful relationship over many years.

[C]

Kate got rid of me. She left the field.

[D]

Yeah. But still, all these years later, here we are. Never.

[A]

Here we are. He showed up. Up like a bad penny. Kately, thank you so much. I think everybody should check out every dot too. There's a 30 day free trial, I think, so you can try it and see what you think. And I'm very interested. It sounds like a community that we've. It's a large one already. It says 85,000 subscribers.

[D]

I think we're actually up to about 125,000. We haven't updated our language, but mind boggling. We're thrilled you're there and you have a direct in if you want to send me any feedback, good or bad.

[A]

Okay, I'll wait for my invitation plus one. Thank you, Katelyn. Thanks, my friend.

[C]

Good to see you.

[D]

Bye, guys.

[A]

Take care. We will continue with Intelligent Machines and don't worry, we're a half an hour away from the Artemis 2 launch and we'll cover that when that happens as well. You're watching Intelligent Machines with Paris Martineau and Jeff Jarvis. Our show today, brought to you by Stash. Have you dabbled in investing here and there, but haven't been happy with how things are going? Stash helps turn good intentions into consistent progress. Stash isn't just another investing app. It's a registered investment advisor that combines automated investing with expert personalized guidance so you don't have to worry about gambling or figuring out on your own. Stash is simple. It's smart, it's stress free. You could choose from personalized investments or. Or let Stash's award winning smart portfolio do the work for you. Or even pick a combo of both. Stash is there to guide you every step of the way. Join over 1,000,000 Active Stash subscribers and finally, let your money work as hard as you do. Don't let your money sit around. Put it to work with stash. Go to get.stash.comim to see how you can receive $25 toward your first stock purchase and a view important disclosures that's get get.stash.com im get.stash.com im paid non client endorsement not a guarantee nor representative of all clients. Smart portfolios are discretionary managed accounts and subject to additional fees. See the Advisory Agreement and Deposit Account Agreement for details. Investment Advisory services offered by Stash Investments, LLC and SEC Registered Investment Advisor. Investor investing involves risk I just realized

[B]

you're wearing a corn shirt. I don't know how it fully did

[C]

not because it just seems so normal.

[B]

So it really does. At a certain point you wear so many bombastic shirts, they all kind of blur together.

[A]

Wait a minute. Did you call my shirts bombastic?

[C]

I like that. I like that as a clothesline. You should start bombastic.

[A]

Bombastic clothes for the weirdo in a us all. Well, anyway, yeah, you know, Paris, did you have an allergic reaction to what we were just talking about or.

[B]

Yes. I mean, frankly, my bad for not jumping in. I had a. I was a little late to the record today because I have had a long and complicated day at work. So I prepared a little less for this than I would normally. And frankly, when it comes to doing a confrontational interview or an interview style where you're heavily disagreeing with someone, I think that you need to actually come with significant, significantly more prep than usual. That opinion may not be shared by everybody on this panel.

[A]

No, I agree with you. I agree with you 100%.

[B]

So I was a little just listening and learning. I just thought it was also a very interesting interview because this has been. I mean it doesn't sound like her products are necessarily for journalists or people whose primary output is exclusively writing in the same way.

[A]

You know, who I actually think it's for is my wife. It's for Lisa, who is very interested in AI. It's not a technologist, uses tools like Grammarly for her writing. And I think she's taken a class in AI and I think she's. I think this is very common. She's a CEO. She's a CFO by training. And I think that there is a lot of interest at the, at that level of people in AI, but they don't really know where to go or what to do and what to start with. If you're a writer, you like to write and you don't want somebody to write it for you. That's, that's.

[B]

I mean this has been a whole. This has been like the topic of discussion on journalist Twitter. Oh, I know this past week because

[A]

people feel like it's not real. It's not.

[B]

I mean, there are a couple of stories that I think are on the rundown that touch on this. There was a Wired report that asked a bunch of different kind of newsletter writers and tech reporters about how they're using AI. And a lot of them were like, I think Alex Heath had said that he feeds a lot of his sourcing material and stuff to a LLM and has it write a first draft for him and then he kind of goes from there and he's like, well, you know, the hardest part about writing is always like going from 0 to 1 on your first draft. And I agree that's very difficult. But the reason why that is hard is because that's where your thinking happens. It's very difficult to. Our job is to synthesize a large amount of material collection through these interesting ways and highlight the value and report that out and write that out in a way that is illustrative to the

[A]

I believe it can be useful for you. I think for.

[B]

It can definitely be useful.

[A]

You have all those and stuff like that.

[B]

Yeah. You just have to be aware of the fact that if you're talking about using AI to produce the core part of your output, you're going to get something that is sanded down at the edges.

[A]

I agree.

[B]

It's going to be an output that. That is by definition not very special or particularly excellent because it is an average of a lot of different things. And that, I don't know, over time I think is going to just result in not the sort of work you want to be doing. I will say I. I think like, like we've talked about in this show before, the areas I think are most useful for AI application and journalism are like when I have a folder full of right now, like 20 different PDFs, maybe, actually maybe like 50 that are all various scientific studies on a crazy amount of things that I've gone and found and know are relevant. And I am now writing and reporting something and I'm like, God, where's that one that says this about that?

[A]

Totally do that. Yeah.

[B]

Asking NotebookLM to go through it and I have all of my documents in there. It tells me what thing I'm that's phenomenal. It's so helpful.

[C]

I'll give you two other examples this this week I was invited to write a blurb for a book by an author I admire tremendously, and I was blocked. I couldn't get started on it. It was weird. And so I asked AI to Write one. And it was crap. I didn't use it, but it unblocked me.

[A]

Right.

[C]

It made me see. Okay. Oh, right. That struck.

[A]

Kind of laxative. It's like a laxative. It's of kind.

[C]

Kind of.

[A]

It's kind of, you know, milk of magnesia for writers. Is that what you're saying?

[C]

More. More like.

[B]

It's like a laxative. Yeah. That's a crazy thing to say, Leo.

[A]

Could.

[C]

Could you not. Alka Seltzer? It's more like to burp it up. It'll be better.

[B]

It's like you got to throw in there just for marketing.

[C]

But on the other hand, I had a copy editor on my new book. I don't want to speak too much out of school, but what the hell, I will. Who spent a lot of time adding P period, and pp period, period to the footnotes and worried about italic commas.

[A]

That should be an AI.

[C]

Turn my four dot ellipses into three dots, which were all wrong. I didn't want the P&PP. I didn't. I wanted my four dots. I wanted this. But he didn't challenge me. There was only one challenge in the whole thing, and I want a copy editor to challenge me.

[A]

Right.

[C]

I wanted one of your. Your characters to do that. And I put it through and I asked it some simple questions, like, what are my ticks? And it found writing ticks that I had done too much.

[A]

So I could go, that's good.

[C]

That was useful.

[A]

That's useful.

[C]

But reaction, you know, unblocking, reaction. I think those things can be useful. Now, as I enter, as I work on the next book, I do have these folders filled with PDFs, and I can imagine doing things like there's in the invention of the amplifier. It's a very complicated time frame. I might ask it to do a timeline for me just so I can keep that straight in my own head. Not making it wrong, but at least it becomes a working structure. But I would never. I agree with you, Paris. I would never go at it. Let me just one more thing. So this week, in Sacrifice for the show, I went and watched in the theater, the AI Doc.

[B]

Which one?

[C]

It's the one that's out in theaters now.

[B]

Isn't there two? Isn't that the whole thing? Thing?

[C]

Oh, I didn't know that. Okay. This is the one that's. It's Tristan Harris's.

[A]

Yeah, we talked about it.

[C]

Nuts.

[A]

Yeah.

[C]

But, you know, it's also because I hadn't had popcorn in theater in six

[A]

years, so you should have gone to see Project Hail Mary, the pop theater.

[C]

Did you go to show the reading? Whatever. Manville Theater. It's in the suburbs, wouldn't it? There were three people, including me in the theater, but really. Yeah, yeah. This thing is subsidized. It's like Melania. It's subsidized to get it up out there. So the doomer subsidize it. But at one point he's trying. The guy is playing the D. I don't know what AI is. What's AI and then screen for screen, it comes back, it's patterns, patterns, patterns, patterns, patterns. And it strikes me that yes, that's what AI is good at. To your point, Paris, it's going to find and replicate those patterns. To me, I came to realize creativity is breaking the patterns. Creativity is seeing the different path to go and AI is never going to do that.

[A]

That.

[C]

And so I think it's useful in some ways to get things done. But in terms of what we do, in terms of trying to think that we, we work creatively and find unique value and, and, and human interest in things, that's still our job, I think.

[B]

Yeah. And I mean. Well, I think you're right, Jeff. In, in what you're noting that it can be useful in finding these patterns and maybe like having moments where like you're blocked and you just want to try and get something on a page. But I also think it's important to emphasize that there is great usefulness in that struggle. You know, trying to find those patterns yourself and really struggling with it and struggling to.

[C]

Oh, I agree. I think I want to resist the patterns.

[B]

Like trying to even, like, even if you are, I don't know, doing all of this work yourself and then realizing that, hey, there's something about this that is blocking me and I'm not able to succinctly analyze this or identify the patterns and synthesize it onto a page is useful data in and of itself. I think that obviously there are situations where using a tool like this to expedite that process could be useful, but you have to understand that you're. What you're. The trade off is. And the trade off, you're not thinking as much. No, you aren't getting as deep of a understanding of the material. The outputs could be wrong. And I think you have to be okay with, with making that trade. And I don't know, the thing that worries me is that we are setting ourselves up in a system where even just taking writing and journalism, for example, you have these Institutional pressures to produce for a lot of journalists, more and more, faster and faster for less and less money. And that leads to people feeling like they have to turn to these tools in order to make it work somehow. But the output is not going to be half as good ultimately.

[A]

Okay, we're about 15 minutes away from the launch, so I want to just do a few news stories so we can cover what's going on. Because there's a lot going on right now in the AI space. Some big stuff. The biggest one I think of the week is the leak of Claude Code's source code. Did you go inadvertently Leak? Apparently it's not clear. Anthropic said it was a human error. It could easily been a vibe coded AI error. But the leak was immediately pulled down but not fast enough and it has now there are tens of thousands of copies. Anthropic's been putting DMCA pull down requests out like crazy, trying to get them off the Internet.

[C]

It's not illegal to have it though. They couldn't really sue you if you just have it, right? But journalist we would say that's that

[A]

you can own it like you can have it on your hard drive. But people are putting out, for instance, somebody's ported it over to Python and you can download that and there's some question about it's. It's originally typescript code. There's some question about if you download Claude code in Python and run it. There's 32,000 stars on the GitHub right now. I'd be a little nervous about it, but I don't know, the lawyers will have to weigh in on that. There are people who have taken it and done interesting things with it. For instance, Ben Davis has actually taken and putting Doom inside Claude Code.

[B]

Of course it's an important part of every.

[E]

You have to answer this question, what can it play Doom.

[A]

Yes. And it can easily quite well in fact. I don't know why that's such an accomplishment. But anyway, we've seen a lot of interesting analysis of how Claude Code works. I think I could summarize it by saying it's more than just the model that Claude Code almost half a million lines of code. It's. The harness is very important to the effectiveness of Claude Code. It's not just Opus 4.6, a lot of really interesting stuff. There is a website that I think is, if you're interested, is worth looking at. It's Claude Code unpacked and you can actually see the process of how Claude Code works with what you Type in, which I think is kind of cool. And it shows the architecture and stuff. It's less. It's less of a geeky deep dive into the code and more of just kind of a higher level how it works and what it's doing. There were some hidden features that showed up. There was one that I think was intended for April Fools. It was only supposed to work April 1st through 7th, called Buddy. It was a virtual pet, a Tamagotchi, inside of Claude code that you'd have to keep alive. There is some interesting agentic stuff, like something called Kairos. There's some very expensive things. There's an auto dream mode, which I think would be very interesting when you're not using it. When you're idle, it goes through what happened and organizes it into learnings, which is potentially very, very powerful. So this is a nice little site if you want to just kind of understand what.

[C]

Anything that surprised you in this?

[A]

No, I don't know if we really learned all that much. Well, okay.

[B]

Wasn't there something that tracked the amount of times you cursed at Claude code?

[A]

Oh, yeah. You want to see that? I could show you that. So this. This comes from Wes Boss. He was digging into the code just to see a variety of things. For instance, when Claude code works, it has a verb that goes over and over again. And there are 187 of them. He found the verbs. These are what they call the spinner verbs. These are the words they filter out 20. I won't say them on the air. They're kind of George Carlin's 25.

[C]

Even worse.

[A]

Swears. It also keeps track of when you swear it is it and sends it back to the home office. It logs your prompt as negative in their internal analytics. West Boss says. Now, I think that makes sense because that's a sign that whatever Claude did wasn't right. And so maybe I wish looking at

[C]

that voicemail jail would do that.

[A]

Yeah.

[C]

Get rid of the agent, the athlete agent.

[A]

So there was some valuable stuff. There was also a mention of capybara. I'm sorry, I get spammed and I should be in do not disturb when I do these shows. I don't know why I'm not. I apologize.

[B]

You live life with your. Every time your phone calls, I'm yolo. It makes a sound.

[A]

Oh, yeah, I'm yolo. Not only does it make a sound, because I have 18 Macs in this room. All of them make a sound.

[B]

What a nightmare.

[E]

I know.

[B]

It's like my clock audibly rang in like Eight minutes.

[A]

Oh, that's nice.

[C]

Really? No commercials, no phone calls. Boy, they live a different world.

[A]

You writers, you know how to live. The other thing I think is important, and this actually is kind of an interesting topic. There is mention in it of capybara. We've heard about this Mythos model that Anthropic supposedly is getting ready to ship. I saw one tweet, accurate or not, that said April 16th soon. That is so good that they have held it back and have been warning governments and security experts, this is going to be dangerous when we release it. There's also a rumor it will be extraordinarily expensive. OpenAI is about to release Spud. Both of these models, they say, are a step change above what they're already doing, which is already very impressive. I'm not saying this is not an inspiring brand. Mythos sounds good, right? Happy bara, Spud. It just Spud. Here comes Spud, your potato. But it brought up an interesting point that I kind of think is intriguing. What if these companies do release really good agents that are incredibly capable, but price it so that only the wealthy can afford it?

[B]

I was gonna say, is that not where all of this is going?

[E]

Isn't that maybe works all the time? Isn't how I. Yeah, but I mean,

[A]

I don't mean like somebody who can afford a three thousand dollar computer. I mean a millionaire. I mean somebody who has so much money they can spend $10,000 a month

[C]

on their volume based business. Well, if business. Or is it a chart?

[B]

Here's the people what it cost. That's the point for these services. Yeah. That only the extraordinarily wealthy would be able to do it. I mean, this is probably one of the reasons why there's a poll that I included in the rundown that one of the takeaways was that a majority of Americans across all ages and political affiliations think AI will do more harm than good to their everyday lives. The only exception is people who make over $200,000 a year. They were the only group that was like, yeah, I think AI will do more good.

[C]

At the same time, people are using AI more than ever. It's just like social media. It's the media trope being played back in the polls.

[A]

It's the movie you just saw scaring people. And actually, honestly, I think it's also people who encounter AI as chatbots have a very different impression of what AI is and is capable of than people who are encountering it in. In tools like Claude Code. It's just a very different experience.

[B]

How is cl. Claude code is still a chatbot you're interacting with. I mean, it's just.

[A]

All I can say is the quality of what you get back from it, the capabilities it has.

[C]

How about this? You're as opposed to.

[A]

I don't use it like a chat bot.

[C]

Like you're getting actions from it. You're getting.

[A]

I'm getting actions from it. From it. I'm getting. It's writing code. I'm not getting. I'm not saying. Yeah, I'm not saying, hey, how much is a stamp?

[B]

Yeah, but most. The average person, no, they're not going to get any use from coding.

[A]

I understand.

[B]

So why would they use that?

[A]

No, I'm not saying they should. I'm just saying their experience is very different because chatbots are kind of dopey. And if you're making silly pictures or if you're asking it, you know, find the best running shoes for me, that's just a very different experience than if you're using it to code and the people are using it to code, are seeing a different expression of what AI can do that is much more powerful and much more intriguing. That's all I'm saying. I'm not saying you should be doing that. I'm just saying it's a different view view of it. The other thing that's a big story. Then we're gonna have to take a break is the big time compromises, supply chain compromises. First on PYPI with the Light LLM and now this one happened yesterday, Axios, which is by the way, a library used by Claude code. The problem is when these are compromised, people are downloading them and running them without their knowledge even. They're. They're libraries that are invoked by other tools you may not know. And they. This is becoming a real serious problem. We talked about Light LLM last week, which is widely used. We now know 47,000 people downloaded it in the 46 minutes it was available. These tools are, these libraries are downloaded. You know, a Light LLM is downloaded I think 92 million times a month. 3 million more than 3 million times a day. Axios also extremely popular. So these attacks, attacks can really threaten in a big way. We're going to take a break. When we come back, I do want to watch us get us to watch the Artemis 2 launch. It looks like there go. It's a very nice day.

[C]

Cape canaveral T minus 12 minutes until.

[A]

Yeah, we're actually at 10 now, so.

[C]

Oh, we gotta get fresh on page. Yeah, 10 minutes. So it's a hold. It's a hold of 10 minutes?

[A]

Yeah, maybe. Yeah. Yeah. It does look like they're a hold. That's a normal hold, I think.

[B]

Yeah.

[A]

I just want to play Walter Cronkite, that's all. So we'll sit here at the launching pad and watch as Artemis 2 makes its way to the moon. First time we've gone back to the moon in how many years has it been? 30 some years.

[B]

Hope it's still there.

[A]

It's full tonight.

[B]

There's only one way to find out.

[A]

Yeah, let's go see. They don't get to get out. That's all. They have to drive around it. It's like when you're driving down the highway and you saw the drive in movie and you got a glimpse of the movie, but you didn't get to watch that.

[C]

Paris, have you ever heard of drive in movies?

[B]

A what in what?

[A]

You surely have been to a drive in.

[B]

Yeah, I have. Well, I haven't been to one, but I know of the concept.

[C]

They still exist.

[A]

There's one or two, I would say.

[B]

I have not been to one because I don't know where they are.

[A]

It'd be a good field trip for the Brooklyn gig gang.

[B]

I would love that.

[A]

To your zip car.

[E]

During COVID they. They. They brought them back in the Bay Area. There was a bunch of dragons that popped up.

[A]

Yeah, you get. And. And. And you have to park and the. And there's a hump where you park. It tilts your car up a little bit. And then there's a speaker on a post, a really crappy metal speaker that you hang in the window. It's a Bluetooth.

[E]

Now, by the way, I'm gonna be honest.

[B]

My entire knowledge of drive in movie theaters, other than just conceptually from reading, is from an episode of the TV show Psych that where a pivotal scene with serial killers took place at a drive in movie theater.

[A]

They're scary.

[B]

They're scary.

[A]

Yeah, because it's dark and you get out of the car to go to the popcorn stand and never anything could happen.

[C]

Anything can happen, and then you cannot find your car getting back, and you're lost.

[A]

Never happen to me, I have to say. But it was fun for dates because it was just the two of you in your car. We're gonna take a break. Come back with more memories from Gramps. You're watching Intelligent Machines with a couple of old guys and a young lady named Paris Martineau and Mr. Jeff Jarvis. Our show today, brought to you by Monarch. Oh yeah. I'm so glad I have Monarch. I tell you, tax season, right? Really gets you to look at your finances. Monarch is perfect for that because they help you make progress with your money all year long. So you don't get to April 15, look back and go, what happened? I made a lot. Where'd it go? I love Monarch. Simplify your finances with Monarch. It's the all in one personal finance tool designed to make your life easier. It brings your entire financial life, budgeting accounts, investments, your net worth and future planning all together in one dashboard on your phone. It's on your laptop. You will feel aware and in control of your finances this tax season. And you'll get 50% off your monarch subscription with the code I am when you go to monarch.com monarch isn't your average personal finance app. I've tried them all, by the way. Unlike most other personal finance apps, Monarch is built to make you proactive, not just reactive. You can use the AI tools. This is new and I love this built in Monarch Intelligence. It's specifically trained by certified financial planners and advisors. So it's smart about money. You'll get Access to the AI Assistant 24. 7. Access to a financial coach accessible from anywhere in Monarch. From questions about trends in your spending to how you pay off debt. It'll help you do that that the AI assistant has the answers. Smart answers from people who know money. AI insights too. Monarch will comb through the data because it's got all the data right to surface. Insights personalized to you, hidden patterns, identifying. You know I'm spending more. Is it. Is it inflation or is it lifestyle creep? Changes in savings rates and more and help get you on track because knowledge is power. You also get an AI weekly recap. You can let Monarch look out for your money with a personalized weekly summary that alerts you to big spending spikes, shifts in net worth, upcoming expenses. Oh, another thing they do. This is new. You can split bills with Monarch. You don't need another app. You just scan and upload a receipt. Monarch will automatically parse the items and the prices. This is the AI awesome network. Then you share a link or QR code with your group. Everybody can claim their items, settle the bill effortlessly. How about that? Achieve your financial goals for good with Monarch. The all in one tool that makes money management simple. Use the code imonarch.com for half off your first year. That's 50% [e-mail] don't forget the code. Very important. Im monarch.com we thank him so much for Supporting intelligent machines. Let's see. I'm gonna go. I guess we're still on a hold here for NASA. Let me go full screen on this. There's happy people watching.

[B]

The chat suggests to me that things are happening. You're not up to date. I don't believe.

[A]

It's not just a hold, is it. Is it something else?

[E]

You're not live.

[B]

It looks like you're not live, Leo.

[A]

Oh, well, I want to be live. How do I get live?

[B]

You gotta not be doing. I don't know, but it's not that.

[A]

Are you sure? I think it's live.

[B]

Is he live?

[E]

Because I don't know.

[C]

The.

[E]

The timecode at the bottom says there's six hours and 23 minutes into that already.

[A]

Well, that's just, you know, that's just them. The clock is still top. No, I'm live. Yeah. 10 minute countdown. We're cleared for launch. We're just waiting. And by the way, our own intrepid space hosts, Rod Pyle and Tarek Malik are not there. I don't think they're in that. Maybe they are in that audience of dignitary. Those are dignitaries. They're not very dignified. I think they're out there at the bleachers. You know, there's a great seat across the water from the launch pad where you get to see it actually happen and feel it. Yeah. More importantly, I was really sad. We. We were at the Kennedy Space center last month and the rocket had been pulled back. They had to put off the launch. We came this close. There they are in the bleachers. That's the bleachers.

[C]

Is the gantry already pulled back or is that.

[A]

Not yet. They're about ready to restart the countdown.

[B]

What's a gantry?

[C]

The thing that holds the rocket up.

[A]

Well, there's Elmer Gantry, but that's a different thing. That's a preacher and then there's.

[B]

You're telling me there's not a preacher holding that rocket up?

[A]

Well, I think the good Lord is holding that rocket up with his. With his mighty, you know. No, it's actually that tower which will slide back, but that's pretty close to launch. Because the thing. You want the thing to fall over, right?

[B]

Can we all get corn shirts? Is that.

[A]

I think we all need corn shirts. I think we all need corn shirts.

[B]

Can. Yeah, I'll get a normal. I'll get a corn shirt like yours, Leah. But then we've got to get a custom black one for Jeff where it's like all the black Corn shirt, Grayscale, black.

[A]

All right, let's see what else there is. We got a lot of news to get through.

[C]

We got time for it. Yeah.

[A]

Google. Oh, actually we should mention this. OpenAI. We did mention it early, I think closed. Oh, this was on Windows Weekly. Another round of financing. 100 $122 billion. That gives them an $852 billion valuation.

[C]

So losing Disney's 1 billion last week was no big.

[A]

Unbelievable. And what I thought was kind of interesting, you know, SpaceX is also apparently today going to prepare for their IPO, which should be a pretty spectacular IPO, but I doubt they'll raise 122 billion.

[B]

Yeah, I mean they confidentially filed today. Right.

[A]

Yeah. So it's interesting that you can still, if you are growing at the rate that Open A is growing, not making money, obviously, but if you're growing, you could still go to the investors and say, hey, give us some money and they'll give you enough money so you don't have to go public. There are rules. We were talking about this on Windows Weekly. There are SEC rules. Google was in the same position. They didn't really want to go public, but they had to because. Because the SEC rules say, I think if you have more than a thousand investors, you have to go public.

[C]

You are de facto a public market.

[A]

Yeah. And that was because they were giving stock options to their employees. I don't know what situation Intelligent Machines is in. Countdown has resumed, so we are now nine minutes away from launch. We will cover that launch, but we'll continue until then. Yeah, SpaceX is filed confidentially. But how do we know that if they file says Bloomberg says it, so that's.

[C]

Well, Bloomberg knows.

[A]

They know.

[B]

I mean, Bloomberg. Yeah. Knows.

[C]

I don't see SpaceX as a business. You know, is it, is it a trillion dollar business?

[A]

It's, it's the same thing as Uber, as we work as OpenAI. These are companies that aren't currently profitable. But there's an upside. There's a huge upside.

[C]

Side, well that, that's. No, that's what I'm questioning. There's only so many rockets to go up. There's only so much.

[A]

They make money. Quite a bit of money on starlink.

[C]

Fine, fine.

[E]

But money from the government, doesn't it like it's all government?

[A]

Yeah, that's the thing they're selling. There's these, you know, this isn't a SpaceX launch, this is a NASA launch. But yeah, they're making a lot of money from the government. I think there's probably thinking of things like asteroid harvesting, asteroid mining.

[B]

Yeah, because that really goes so well.

[A]

Cities on the moon.

[B]

Yeah. Both Bloomberg and the Wall Street Journal have reported that SpaceX is filed.

[A]

Do they say what they expect the strike price to be or.

[B]

They're aiming to raise between 40 billion and 80 billion dollars in an offering.

[A]

See, isn't that interesting? So OpenAI without going public. Public Raised twice that.

[B]

Well, yes. And that's why we're all kind of squinting our eyes at that.

[A]

And you have to figure that SpaceX will really. Is it already part of Tesla and XAI or is there. Have they not merged all through. There's the gantry moving just.

[B]

No, it did SpaceX combined with in February to create a 1.25 trillion juggernaut in the biggest corporate tie up by value in US history. Which is the most Wall Street Journal clause of all time.

[A]

A juggernaut.

[C]

A.

[A]

Well, yeah. So the AI. There's the AI investment there too. Maybe that's part of it.

[B]

I mean, what will be interesting which they notice that once the filing is made public, it's going to have a bunch of interesting and never before seen information about the combined company's operations because they obviously have to go into detail about all of this.

[A]

This is why Google didn't want to go public. So a lot of companies would prefer not to go public because suddenly you are.

[B]

Yeah, you got to open the books, tell all the people the reasons why you're a bad investment.

[A]

Tesla's public. Right. I have to say companies have gotten very adept at non GAAP accounting and hiding. And for instance, Apple. Apple no longer says how many iPhones they've sold. You'd think that'd be material, of material interest to investors, but somehow they get away with it.

[C]

I guess the argument is that if, if you really want to know that and you don't want to invest, then you don't invest. It works for them.

[A]

But the whole point of.

[C]

I know.

[A]

Disclosure. By the way the Trump administration is thinking. Or the SEC is thinking of eliminating quarterly reports.

[C]

Right.

[A]

Saying maybe we just do this twice a year.

[B]

What? Why?

[A]

It's a lot of work. Every three months. You got to get all the financials together.

[B]

So. Sorry. That you have to tell the public and your investors what you're up to every three months, multiple times.

[A]

My point is that it's slanting away from investors towards these companies.

[B]

I mean.

[A]

Yeah, that was the whole idea.

[B]

It is like the bare minimum.

[A]

Yeah, well, I mean they're the next people.

[E]

Right? Because first they screw over the customers and then they screw over their investors.

[A]

That's called inside ification.

[C]

So you wait another three months to find out more?

[A]

Yeah, yeah. Google is rolling out AI inbox beta for AI ultra subscribers.

[B]

That is a statement that sounds, felt like just fake. That sounds. That sounds like you generated that statement algorithmically.

[A]

Okay.

[B]

What is an AI Ultra subscriber?

[C]

I don't even know.

[A]

It's a personalized briefing that surfaces the actual information you need instead of a message that you have to open and read.

[B]

I, I mean there's no company that I would trust to do this less than Google, who anytime I try to use the search function in Gmail, it returns awful. Anything that I did not search for it is like astounding how bad it is.

[A]

And this is returning the things I'm famous for. Search.

[B]

Yes, it's baffling.

[E]

They need you to increase the search queries. They need you to ask again and again.

[A]

Well, they said that they have a new engine, quote engineered privacy environment where your information is processed in a separate space and doesn't leave that space so that Google doesn't have access.

[C]

So I signed up for Google workspace studio that's supposed to send me. Notify me of urgent emails.

[A]

Yeah, this is an example.

[C]

Jeff is urgently asked to donate 25 to Mary's U. S. Senate campaign tonight.

[A]

It worked.

[C]

First FEC deadline.

[A]

That was urgent.

[C]

It's urgent.

[B]

It was so urgent.

[C]

Jarvis needs to coordinate a preliminary meeting with the panelists and submit administrative materials. Urgently urgent.

[A]

I think people are just going to game it right and put the word.

[C]

Chicago's membership team is issuing a final, final urgent appeal for donation.

[A]

I wish I could have a button that eliminates all solicitations for money.

[B]

It's also just very. I don't know any Google rolling out any features like this has my hackles raised because Gmail and G suite are I think, notorious for not allowing you to turn off any of the extra AI features or else it completely wrecks the basic functionality of Gmail. If you try to turn off the awful little, little purple line that says how you should rephrase every sentence of your email. Not for grammatic ish. It's not because the grammar is wrong. It's not because the spelling's wrong. It's just like a better phrase. If you want to turn that, that off. You have to turn off all spell check. You have to turn off really any sort of ability to sort your inbox in into like little categories. You have to turn off basically every feature that makes Gmail. Gmail because it ties in those features with like Paris, you're just.

[C]

The P stands for picky.

[B]

You're just supposed to stand for picky. Yeah.

[A]

Now I have to point out Gmail launched on this day Today, April Fool's Day, 22 years ago. 2004. Can you believe this? And in 2004 I thought it was a pretty good idea to create a Gmail account. Laporte not a good idea.

[B]

Two minutes to launch. Two minutes to launch.

[A]

Should we go to the video?

[B]

I'm sorry to interrupt.

[A]

Yeah, it's exciting. I have the audio. I think I do. Let me make sure it's.

[C]

Oh, that's neat.

[B]

Two minutes is a long, very a very sensual shot we've got going on here.

[A]

That's the rockets behind.

[C]

But what happened with your. Your laporte Gmail?

[A]

Well, Google has announced that you can now change your Gmail name.

[C]

Was every possible name you want has taken long years since.

[A]

Yeah, I mean if I change it it's not going to be LaPorte. It's going to be LaPorte 5879999 or something. Right. Something that knows spammer will. Will know. Unfortunately you keep your existing Gmail. Amen. Because of course they're not going to turn off that. You'd lose all your email. So I'll still get all the spam I get at laporte.

[C]

Yeah. And you could revert to your name when you. When you regret change.

[A]

I mean this is why you just

[B]

got to create a second one and then you have two separate. It's just another name boxes that you can look at.

[C]

I guess it's recyclable from this vehicle.

[A]

Nothing unfortunately.

[C]

Nothing.

[A]

It should be be. It was. It was intended to be but these are reusable. The pods on the side were reasonable.

[B]

Do they ever get those astronauts who were stuck up at the space station?

[A]

They did, yes.

[C]

Okay. And they both retired the astronaut.

[A]

No, no, that's a different one. We're talking about the ones who were stuck there because the. Because of the leak in the Boeing return vehicle. They both retired from the astronaut corps.

[C]

One minute.

[A]

Yes, the guy, the astronaut was returned a little while ago it turned out. Yeah, he went somehow lost his ability to speak and he's fine speaking again. Yeah, he's fine now. He just wanted to come back. Here we go.

[B]

I'm not sure that I've ever watched a mission to the moon happen live.

[A]

Oh, you missed it back in.

[B]

Actually no, I'm certain that I haven't because it's the first time in 50 years.

[A]

And he'd say we're about.

[C]

What year was the last time we went to the moon?

[B]

50 years ago.

[C]

No, the last time.

[B]

No, the last time. Astronauts. It says on NASA's website we're sending astronauts around the moon for the first time in 50 years.

[A]

Wow.

[C]

She's. You and I are old deal

[A]

25 engines. Holy cow. Yeah. These are very powerful engines. Oh, my gosh. Very fast.

[F]

Whoa.

[A]

Now, bound for the moon, humanity's next great voyage begins. That's. That's steam, by the way. They water cool the launch. So that vapor is steam.

[C]

Florida's not humid enough.

[A]

It's beautiful.

[C]

Yep.

[B]

That's so cool.

[A]

I'm always inspired by watching these

[C]

and praying a little bit.

[A]

Yeah. You and I are old enough to remember watching.

[C]

I was in the newsroom when it happened.

[F]

Integrity passes the alternate vehicle target milestone. Mission Control, Houston.

[B]

See?

[F]

Good performance.

[A]

Paris wasn't even born in 1984.

[B]

No, I was very much.

[C]

Orwell has nothing on her.

[A]

It's. I like doing this because it's fun to watch this together. Yeah, It's a good communal.

[C]

We need Jimmerby.

[A]

Yeah, that's the cheers coming from the bleachers there. It's apparently something to see. That ground really rumbles. There's a huge.

[B]

Oh. To slip the surly bonds of Earth.

[A]

Actually, that's what Reagan said when the Challenger astronauts died. So we don't say that now.

[B]

Sorry.

[A]

That's not the way to. That's not really what they're trying to do right now.

[F]

Integrity is 14 miles in altitude.

[A]

Eight miles. So five days away from the moon.

[C]

How many orbits around the Earth before it swings?

[A]

I wasn't sure I asked that question. I don't know if anybody knew that. Probably not a long time now.

[C]

If you were really Walter Cronot, you'd have a model. You'd be explaining all this.

[A]

He would have these models that they would take apart. And it goes like this and it rotates like this. And you'd have an accident or somebody sitting next to you.

[C]

Surprised they haven't given us the camera inside.

[A]

Well, what's amazing about these now is there's the people at the bleachers. Look for Rod and Tarica. They're out there. I'm sure there's of course, a bunch of tourists.

[F]

Separation.

[A]

Hey, there's your mom and dad. Whoa.

[B]

How cool. Hey. To the moon as this.

[A]

There's a shot. Look at that stage. RCS ready. Watch. We're going to.

[C]

One shot. Inside was mainly of the feet of two of the astronauts.

[A]

I know, it was interesting that you could see that they got a foot cam. Yeah, they have these big boots.

[C]

Cuz they're, they're double deckered in there, so.

[F]

2 minutes 45 seconds of mission elapsed time into the Artemis 2 mission thrusters on Integrity and upper stage confirmed in a ready state ahead of service module fairing separation.

[B]

Do you think that person's chose like, do you think there's like a little audition for who has the best voice to kind of do the public announcement?

[A]

Well, it might be now. It used to be Capcom and these guys and now it's.

[C]

Now it's a.

[A]

Now some P guy. Yeah. Oh, there goes the fairing. They don't need that anymore. They're out of the atmosphere. Gosh, you guys, I wish Richard Camel were here because he knows exactly what's going on. He was an amazing tour guide.

[C]

Houston, Integrity. Good last jettison.

[A]

Great view.

[C]

Wow.

[B]

Would hate to be a fish right now.

[C]

Team Eco 8/02

[A]

we see the same on board.

[C]

Stan.

[A]

Amazing. And Houston has you loud.

[C]

Can I go to the bathroom now, dad? Says the kid going through the

[A]

outstanding.

[C]

Stan, we have you the same.

[F]

3 minutes 50 seconds into the flight of Artemis 2 wise.

[C]

I wonder where that camera is.

[F]

Hanson, cross the boundary to space with good com checks. GPS signals acquired after last jettison. Now working on internal checks to verify accuracy. Light dynamics officer analyzed the time of main engine Cutoff confirmed at 8 minutes 2 seconds. Time of MECO.

[A]

MECO main engine cutoff.

[B]

Oh, Darren points out a. Makes a good point. It's all fake because we know the Earth's flat.

[A]

Right?

[C]

Right, Right.

[A]

Just now in space. As Brand points out, the Karman line is the order between atmosphere and space. Pretty amazing. I love the pictures. These pictures from the. That's the best part spacecraft and we're getting such good pictures nowadays thing.

[C]

Yeah.

[E]

Move your cursor, Leo.

[A]

Outstanding.

[C]

Stand we.

[A]

Sorry about that. There it is. There is no cursor in space.

[C]

That was God moving it in the game.

[A]

Move your cursor.

[F]

Five minutes into the flight, Commander Reed Wisman confirms he can has visuals of his destination Paris back in the day.

[C]

This is an incredibly wobbly picture where the would go all around the screen because there was no holding steady.

[A]

Yeah they have some guy on with a very long telephoto lens.

[B]

Someone doing sketches and kind of mailing them to your address.

[A]

Amazing shots. These are incredible.

[E]

Now there's like a GoPro mounted to the side of the rocket. What is. Where is this camera?

[F]

At this point, three good main engines are all that's needed to carry integrity to a nominal main engine cutoff target.

[A]

I can't wait to hear Rod and Tarek talk about their experience.

[F]

Joel? Houston. They're watching 75 miles altitude 330 miles down range approaching 10,000 miles per hour.

[B]

Very.

[C]

That's what's amaz.

[A]

Thousand miles an hour.

[C]

You're Boeing jet.

[B]

We put our mind to it.

[A]

Yeah.

[C]

Imagine what else we could do.

[B]

Someone in the chat. Oh, I guess make a AI video. Video of one of the corns on Leo's shirts taking off.

[E]

You can't make AI videos anymore.

[B]

I was just about to say I almost interrupted myself.

[F]

Google has it up still at the seven minute mark.

[C]

I was. The animations were always great too.

[F]

Miles in altitude, 460 miles downrange integrity.

[A]

Expect nominal shutdown down. It's good. Everything's gone great. Very encouraging.

[B]

Patrick asked a great question. What's that bright spot in the middle and what's the stuff coming off of it?

[F]

Elapsed time. Shutdown plan is as expected.

[A]

I think it's refraction again.

[F]

The time of meeting is confirmed at eight minutes. Two seconds into what?

[C]

That. That shot of the. It's the engine.

[B]

Yeah, yeah. What's that?

[C]

That's the engine.

[B]

Oh, that's like. Like from below.

[C]

It's an extreme.

[A]

I don't know what that is.

[C]

Telephoto, I think.

[F]

Seven minutes of mission lapse time.

[A]

All right, well, we should probably continue

[B]

on instead of just the really compelling

[A]

audio of us just silently watching in awe.

[B]

Watching.

[A]

Yeah. Very cool though. Very cool. Congratulations. Godspeed, as they say.

[B]

They should do a podcast from up there. They should do a podcast on the moon.

[A]

That would be cool. I volunteer for that. Let's see, what else? Are you gonna change your name? Did you change your name on Gmail? We were talking about that before the launch.

[C]

I've got a bad one.

[B]

What would I change it to?

[C]

That's the thing.

[B]

I have all the Gmail addresses I'd want and I have them, I guess separated because I. This wasn't a thing before, but now I just have all of my logins. I'll be logged in at the same time. We'll just toggle through them like the different user.

[A]

So yeah, I don't really use Gmail anymore, so I guess I don't really need to worry about it. Have you tried the ATI app on Blue Sky Blue Skyers?

[C]

You can't. It's not available for us.

[A]

No, it uses AI to give you control over your social feed. And is it right now the most blocked account after JD Fans on Blue Sky

[B]

AT Tracks?

[A]

Yeah. It allows users to design their own social media algorithms and create custom feeds within an AT protocol.

[C]

So you're making your own algorithm, which is a good thing, right?

[A]

Yeah. This is what, 8. This is what Proto was all about, was control. 125,000 users have blocked Addie's Blue sky account.

[C]

Why?

[A]

I don't know. Oh, it's AI. Yeah. That's worse than the White House and ice.

[C]

I think it makes a good sense. I saw somebody on. On the Socials on Blue sky who said that if the algorithm is supposedly addictive, whoever it was, I wish I could remember said, well, I don't use Facebook and it's got an algorithm. I use Bluesky. There's no algorithm, and I'm addicted to it. But now you can make your own algorithm.

[A]

Yeah. Well, we'll see. You know, it's early days yet. Once we get access to it, we can see how that is. The Electronic Frontier Foundation. I just did a nice interview with their director, Cindy Cohn, a couple of weeks ago. That's available on our Twit feed. Is suing for answers about Medicare's AI experiment. Now, Jeff, you and I are on Medicare because we're seeing your shinig. They. Medicare is proposing to use an AI algorithm to determine what treatment we get.

[C]

The death panel turns out to be a machine.

[A]

It's worse than a death pen. I'd rather humans made these decisions.

[C]

Amen.

[A]

This was announced by Dr. Oz last year.

[C]

That gives me confidence.

[A]

It is. The program is known as wiser, which is an acronym that stands for Wasteful and Inappropriate Service Reduction. Did they really think this through? This name is not good. It uses AI to assess prior authorization requests. Ah, interesting. Prior authorization requests for Medicare beneficiaries. Oh, I see. Prior authorization requires medical providers to obtain advanced grants approval from the insurer before delivering treatment. So it is a request. Can I deliver. Can I fix Jeff's bum ticker before I could get.

[C]

Which already.

[B]

That is a. That is a system that exists in order to put more hurdles towards people accessing care. It's already a speed bump. And you're adding robotic speed bump on top of that. Sorry.

[A]

Rolled out in January in six states. 6.4 million beneficiaries by design. Weiser incentivizes contracted companies to deny prior approval against the best interests of patients because vendors are compensated in part on the volume of healthcare services they Deny.

[C]

Well, this is worse. I'm not on Medicare. I'm on Advantage. This is the only choice I was given. So Aetna is paid a lump sum and its job is to then be as efficient with me as possible.

[A]

Give you as little care as possible. So they already minimize their. That's exactly right. So EFF Is has asked for injunctive relief for violation of the Freedom of Information Act.

[C]

Go eff.

[A]

They want to know more about them. Yeah. Thank you, eff. We appreciate it. Governor Newsom has signed an executive order requiring safety and privacy guardrails from AI companies that do business with the state. I'm not sure what those guardrails find.

[C]

Guardrails, Right.

[A]

Yeah. What is.

[C]

What is those?

[A]

They'll have to explain their safety and privacy policies around AI. The state will examine policies carefully on how the companies prevent exploitation of individuals, including the spread of csam. The government will also consider whether AI models are used to monitor individuals or block certain speech. Companies will have to explain how they're avoiding bias in their systems. Not sure that's a bad thing, but. And they also. He called on state officials to begin watermarking AI generated or manipulated videos they create. Certainly the state officials who are creating AI videos should definitely watermark them. I would agree on that. I don't know if you can compel private individuals to do that, but you certainly, if you're working for the government and you're doing that, you. You should label. And David Sacks, we thought might be leaving the White House, but no, he has a new role. David Sacks, the podcaster and AI and cyber coin czar at the Trump administration, is going to continue to advise, but I'm not sure how to find out.

[C]

We didn't go over this last week. The. The White House 12 point AI plan was.

[A]

The attention span of the White House is so short.

[C]

Well, but it was. It was written. It was written as if by an adult. Yeah, there was booby traps in it, but it actually had some decent things in it. Nothing's going to happen, though.

[A]

Nothing's going to happen to happen. You're watching Intelligent Machines. Jeff Jarvis. Leo Laporte. Paris Martineau. Great to see you. Good to have you. Now, let's talk about these surveys that you were talking about. Paris. 15% of Americans say they'd be willing to work for an AI boss. 15%. Well, that's not.

[C]

15% of Americans say their boss is so bad. How bad is it?

[A]

Oh, bad.

[C]

I'd rather work for AI.

[A]

A Quinnipiac poll, only 50. It really should be. Only 15% say they'd be willing to work for an AI. The majority said they would would not be willing to swap their human boss for an AI people manager. Nevertheless, you may be ending up working for an AI. Companies like Workday have launched. This is from TechCrunch. AI agents that can file and approve expense reports on your behalf. Amazon has deployed new AI workflows to replace some of the responsibilities of middle management. You know, you can get rid of middle management, I think that's okay. But of course all those people are out of work.

[C]

My friends in various universities, Workday is the most hated thing.

[A]

Yeah, nobody likes Workday. Yeah, I used to have to. This is pink. AI is being used to replace layers of management in what some are calling the Great Flattening.

[E]

It's true that in a lot of tech companies there are a bunch of people who don't do anything.

[A]

There are a lot. Middle management tends not to be useful. And most people I think would be much happier with less hierarchy anyway. Quinnipiac says Americans are wary about what it means for job prospects. 70% of the respondents say they believe advances in AI will lead to a decrease in the number of jobs job opportunities. I don't know what the other 30% were thinking. More than half the U.S. this is the one you were talking about, Paris say AI is likely to be harmful.

[B]

And the only segment they split up all of these responses whether you think AI is going to do more good or do more harm into a bunch of different things. You know, they split it up by like your political beliefs, what generation you're in, income, what you know about AI. In all cases they were adamant you're going to be doing AI is going to do more harm than it's going to do good in. The one exception is people who made over $200,000 a year. Those people said yeah, I think more good to my day to day life

[A]

then it's going to help me.

[B]

Which I think is notable.

[A]

Anthropic asked the question of several actually 80,000 people, if you could wave a magic wand, what would AI do for you? The largest group cohort said 18% said people hope for professional excellence. It would help them do better.

[B]

That's such a funny category.

[A]

Then personal transformation, then life management, time, time freedom, financial independence, societal transformation, entrepreneurship, learning and growth. And only 5% we're hoping for creative

[C]

expression because they're going to do it on their own. Right Paris?

[B]

Hopefully.

[A]

That's the survey. Survey says what other. Before we move on, what other stories were you excited about Paris?

[C]

You first. I got. I got one or two.

[A]

I'll give you one real quick. Most people. This is from you, Jeff. Most people don't enjoy their jobs, says the Perplexity CEO.

[B]

Shocking. Wow. I'm so glad we have somebody on this case to figure it out.

[A]

And so that means good news. They're going to get laid off. But it's a chance to start your own AI venture. AI layoffs create entrepreneurial opportunities. This was, of course, on the all in podcast with David Sacks and Jason Calacan during gtc. Aravan Srinivas acknowledged, yeah, AI could displace jobs, but the reality is most people don't enjoy their jobs. So it's a boon.

[E]

No, the reality is most people don't like having to have a job.

[A]

Most people enjoy paying the rent. Most people.

[B]

I don't think most people. Yeah, most people do not enjoy paying the rent. They enjoy the rent having been paid

[A]

by the job they don't enjoy. All right, I gave you a minute to think about something. Did you find something you care about?

[C]

Line 80. Really kind of interesting story.

[A]

A biggest AI conference. What is the biggest AI conference?

[C]

It is the ICML. God, now I gotta remember what that stands for. All of these conferences, they presume, you know.

[A]

Right. Well, you know, it's International cooperation and Machine Learning specialists.

[C]

It is the International Conference on Machine Learning. Right?

[A]

Yeah.

[C]

So they put out a call for papers and in it they buried invisible instructions.

[A]

So they put. They put these in the papers that they sent out for peer review.

[C]

Right.

[A]

If a reviewer fed the paper into an LLM, the invisible instructions would then insert telltale phrases in the review so that they would know, hey, you used an LLM for the review of the papers. Five hundred and Six reviewers used AI. 398 had their own papers rejected as a result. So in other words, a lot of people, what the hell? They did, agree not to do that. I guess violations of the LLM review policies.

[C]

But there's a fair amount of upset that this organization would do this to their own people.

[A]

It's the International Conference of Machine Learning. You would think, think they won't use AI.

[C]

Exactly.

[A]

This is what they do.

[C]

Exactly. The decision to assume from the outset that your reviewers cannot be trusted to read a paper and think about it without a machine doing it for them tells you something far more important than the number who got caught.

[A]

It trapped him.

[C]

It tells you that the institutions responsible for advancing human knowledge no longer believe human judgment is the default. So, no, it's just a Helper just to help her. Those persons pissed that ICML did this.

[A]

Yeah, it was only 1% of all reviews. So it wasn't.

[C]

Yeah.

[A]

That widespread. Actually, I'm surprised it wasn't more widespread.

[B]

I mean, don't most of these systems catch prompt injections like that?

[C]

Well, I don't understand how, if it was a paper to review, how did it end up in the review they wrote?

[B]

I mean, I guess probably what it would be is the classic like white text.

[A]

Yeah.

[B]

In a document thing where you upload it in like you upload a PDF to Claude or Chat GPT and you, the human looking at it, just see your normal PDF with the black text you read, but hidden in there is white text or very small text that you don't easily see, but that the LLM says.

[A]

But the issue is to insert the word spaghetti into the review.

[B]

And then a lot of these systems now have kind of a built in defense for that, which is there's something in this that told me to insert the word spaghetti. I'm not going to. And it will notify you of it.

[A]

So they're not even using the best AIs. The ones that got caught were using crappy AIs. What else do you got? You have another story, Paris, that you would like.

[B]

I mean, we talked about, talked about this earlier, but there was this Wall Street Journal piece that was a. It's line 92. It's a profile of this journalist at Fortune, Nick Lichtenberg.

[A]

He's very relaxed. You know why he's very relaxed?

[B]

Because he uses everything.

[A]

Yeah. You don't write in your lap. Have you ever written in your lap with your feet up on the desk? No, no.

[B]

I mean, not in many, many years.

[A]

That does not look like a good way to write a story.

[B]

They say that on a Wednesday in February, it starts. Journalist Nick Lichtenberg produced more stories in six months than any of his colleagues at Fortune delivered in a year. One Wednesday in February, he cranked out seven. I'm a bit of a freak, Lichtenberg said. Now let me do a rare case of a Paris. Back in my day, which is back in my day when I started my journalism career a wee year, eight, nine years ago, I also regularly cranked out seven stories a day. And it wasn't because I used AI. It's because that's called aggregation, baby. You write up other people's work, you dual, you source it to them and you're basically just doing a bunch of quick blogs that really add very little to the overall e Insider ecosystem. You just are trying to scrape off some of the traffic graphic. If people wanted to read about that story. This is not, I don't know, anything novel, but it's being pitched as if it is like a revolutionary approach to.

[C]

Well, because he got there differently because how he did it.

[B]

He said AI assisted stories accounted for nearly 20 of Fortune's web traffic in the second half of 2025. Most were written by Lichtenberg. And basically how he seems to do it is he just like throws some stuff in his LLM of choice, prompts it to write an article and then pastes it in the CMS and edits it from there. And I'm just like, this is a disaster. I mean it's disaster ethically. It's basically participating in this interview and profile is advertising to your readers that you don't care enough to even do the work, to produce work for them. You're just offloading it to something else. And also, the thing that ended up making me sad about all of this is getting given my own experience. Like, yes, living in the aggregation, like the aggregation content minds wasn't great. It sucked. It was. You were not producing good journalism by any means. But it was also, I think like an essential part of the experience of becoming a journalist and getting exposed to a bunch of different. Like that when you're writing six or seven aggregation pieces a day, you are reading like six or seven big news stories every day, figuring out how those stories work and then figuring out a way to kind of make your own facsimile of that. And that is a useful skill for first job journalists or early stage interns. And it just makes me sad that I don't know, this whole section of the economy of, of the journalist ecosystem has been hollowed out and will continue to be hollowed out because like what you are, what are the earliest stage career options for this industry supposed to be?

[A]

You can't imagine the work Woodward and Bernstein did in the Washington Post exposing the Watergate burglary being done by a machine. A human has to do it. Now I can imagine a rewrite machine rewriting their.

[C]

If you were on the Cleveland desk at the Associated Press, you were taking articles from the newspaper and you were rewriting repeat them.

[A]

Yeah, that's dumb. Yeah, but, but if, but real reporting, you got to go out and do it, right?

[E]

Yeah. You don't get to the real reporting right away though. You need to go through those initial steps of reading other people's reporting and

[A]

figuring out before you actually do it. I Just I think, and writing up the story.

[B]

It's like such an essential step to learning how journalism even works is just reading a bunch of, like, if you're fresh in a beat. The advice I always give to, like, younger new reporters, like, just read literally everything you can on the beat, like the commodity news stories, press releases, write ups of press releases, write ups of other people's scoops, anything possible. Because ingesting all that information will make you a smarter person and better reporter and better writer because you'll start to see the patterns. And offloading that sort of pattern recognition to a tool just doesn't do you a.

[A]

That's really true in anything. When you're learning it, what you're learning is kind of the pattern recognition, the rhythm of it playing.

[B]

I mean, this is why I also. This is why, despite the fact that we have calculators, you still teach children math. Because it's important to understand the basics involved in the work before you get to offload it to something.

[E]

Yeah. And then you get to the point where you're like, okay, this is what I want to offload to AI and this is what do I want to. To offload to AI but until you get there, you need to do all the stuff yourself.

[B]

Yeah. And I'm just worried that this is creating a whole culture where that's not even going to be an option for people anymore.

[C]

Little Devil's Advocate, just for a second. And that's all it is, is that once upon a time you had to take a pencil and graph paper and draw charts and graphs and pie charts and all that, and then suddenly spreadsheets did it for. For you. And it wasn't cheating. It was just easier. It was just. It was just information presentation and it did it well.

[A]

Well, how much did you learn from drawing pie charts? Whatever that.

[B]

I mean, I do think that there would be some use case in the sense of like, maybe when you're trying to learn some, like to understand that the pie chart is going to be made up of like five. One part of it will be five. Five little identical slices, the other part will be 10. And that. That's the. To understand it by doing it by hand, to understand the proportion of it could be useful for someone if that's how you learn.

[C]

But you want to look at the USA Today stories or data visualization. I'm not arguing at all with reporting. This goes back to the story on the Cleveland Plain Dealer. They argued that they can get more stories now out of the reporting. And Paris objected at the Time. But I think there's some level writing the commodity crap that, that just gets you SEO and, and links out of volume is, is the last dying gasp of mass media and scale. We gotta go to quality. But at some point, some of the stuff that exists. All right, let me ask you this way. For many years now, long before LLMs, the wire services have used computers to write financial results stories. The basic. Their EPS was this. It's just a rewrite of the report they send out. There's no reporting involved. Maybe there should be, but there wasn't. In any case, it was just a rewriting of that. Is there any problem with that? Is it's just a commodity piece of work. Did that a lot. Did that free up reporters to do more actual reporting?

[B]

I mean, I think that's a specific use case because the reason why those commodity newswares do it is because the entire value proposition of paying for a wire service or subscribing to Bloomberg Terminal is getting that information as quickly as humanly possible at speeds that are not replicable by a human.

[C]

True.

[B]

I mean, I think for like, if you're even thinking about it from a purely like business perspective, obviously zooming out from that sort of like commodity speed, I just, I'd have to hope that if you are positioning yourself as a news company that wants subscribers, or at the very least is trying to convince people they need to come to your site and see your ads in order to read your work, you have to hope that you're offering something of value that is some sort of unique capability that isn't just whatever Claude Opus 4.6 can generate from a single prompt. Like, what is going to separate you from every other website that does that?

[A]

It's really about how you learn, isn't it? So here's a good question though. Once you learn. So for me, it was really interesting to learn that numbers squared was called squared because it referred to, in fact, a square. The area of a square is this. Multiplied the sides Together, those are 3 by 3 by 3 by 3 square. The square of 3 is 9, which is the area of the square. And when you. There's a process of learning that, that you can't. Just what I just described, I don't know if you can learn that abstractly, but if you do it, if you write it on paper and you do it, you learn it. Now, do you ever need to do it again? Probably not. You don't need to keep making pie charts and keep making cutting cylinders and bisecting cylinders and things. But it is valuable to do that initially. That's how you learn. Right? But you don't have to write a thousand articles about financial results.

[B]

No, I just think that, like, we need to think carefully about what we're offloading. We're losing to these tools and what we lose. I think that's fair.

[A]

Yeah.

[C]

It's about what we might also gain because the waste of time. My favorite example of this was always the Buzzfeed 2 color dress story. And I would go to journalism conferences and I would say, how many of you had your own version of that story? And every hand would go up. An utter waste of journalistic resource.

[A]

Yeah, no kidding.

[C]

For everyone to buzzfeed story was fine. It was cute, it was fun. It was what it was. They all rewrote that story so they could get their own SEO and their own social clicks and a complete and utter waste of journalistic resource. Don't do it.

[A]

I see that now when I do my beat check. There's so many duplications. I'm actually now having the AI do some deduplication. So I. But the trick is what's the best story? Maybe is it the original story, the one you want? I guess it is sometimes.

[C]

But sometimes somebody asks sometimes, or sometimes

[B]

someone like has, yeah, an interesting angle. They have additional reporting that they're able to add to it. And I think that that's going to be the differentiating factor for all of this. But I think these are questions. The questions the people across all these industries are gonna have to start asking themselves is like, yeah, sure, you can just like automate the most bare bones, average of average version of it, but what value does that add? You need to be thinking about what is valuable about the work you're producing, not just how much work can you produce and how quickly can you do it.

[A]

You're watching intelligent machines. Jeff Jarvis. Paris Smartno, we're glad you're here. We especially thank our Club 20 members for making this show possible. Without you, there would be no intelligent machines. There would be practically no twit. You're not a member of Club Twit. Please consider Twit TV Club Twit ad free versions of all the shows. Access to the Discord special programming that is only available to Club Twit members. Twit tv. Club Twit. Thank you in advance. All right, I've got our picks of the week. Next, let's start our picks of the week. As we always do with Paris Martineau.

[B]

I'm shocked that we're here. This is the fastest show we've done in years, I think. Leo, are you okay?

[A]

I am attempting to make the show shorter.

[C]

I didn't.

[B]

I thought that was a bit.

[A]

No, it's not a bit. I even vibe coded a little clock.

[B]

Clock.

[A]

You'll see right here.

[B]

Is it a clock?

[A]

It's a clock.

[B]

Wait, don't you have. Oh, it's a clock. That just says content with an ominous photo of a frog.

[A]

The frog is my wallpaper.

[B]

I just like that the frog looks like it's judging you.

[A]

It does.

[C]

Exactly.

[B]

Imagine a frog that is just looking deep into your soul. Yeah. Make it so that the frog is kind of horrible.

[A]

That's a good idea. I will. I will get Claude up to work with her.

[B]

Well, Leo, my pick was gonna be the thing that you put as your pick. So I've got two kind of crappy picks instead.

[A]

No, no, do the pick. No, no, you go first. So you can steal my pick.

[B]

That's the rule. Then my pick is gonna be this wonderful collection of obsolete sound.

[A]

How did you find this? I'm so impressed.

[B]

I came across on Blue sky or Twitter in one of my bookmark folders.

[A]

So here's. Don't look at the screen. Screen. Let's see if Jeff will know. This here's a sound. Jeff, what is this? Hello.

[C]

Silence. That's what the sound.

[B]

You got it. Good job.

[A]

Why is it not making the sound?

[B]

This doctor's just standing here looking.

[A]

Let's play a random sound. It's not very loud, is it?

[C]

I think I just heard my stomach. Was that it?

[B]

Oh, no. It's working for me. But it is.

[A]

They're just not very loud for you. It's just not very.

[B]

You have to click on the computer icon.

[A]

It's just not very loud.

[B]

So basically it is a collection of disappearing sounds and sounds that have gone extinct.

[C]

Like payphone.

[A]

Like a mobile phone interference. Remember that?

[B]

You'd Hear Philips Type HM3210 coffee grinder. Like the sound of winding up. A pocket watch.

[A]

A cash register.

[B]

An Olivetti Dora typewriter.

[A]

Oh, I had an Olivetti. I love my Olivetti.

[E]

Is there a cigarette vending machine here?

[A]

Oh, that's the sound you don't hear much anymore.

[E]

You know, like you have to pull the T tabs.

[C]

That old dial telephone. Let's do that for Paris.

[F]

Thank you for buying a Phillips video.

[B]

This is great. Yeah, they've got an old cash register in there.

[A]

Yeah, you don't hear this much.

[C]

Quiet. That's too quiet.

[A]

Yeah, yeah, they're very Quiet. I've got it turned all the way up. I can't. Also, unfortunately, each sound in the project is recomposed and reimagined by artists.

[C]

Oh, no.

[A]

So it's kind of now stuck. Stop it. That's the old dial telephone.

[B]

Yeah. I mean, I think part of the reason why it's been reimagined is that they can play them all together, you know, so that they can actually have them and have them available for you to download.

[A]

I like this notion that you. That you're gonna have a time where you just don't hear that sound anymore.

[B]

Yeah, sounds will just, you know, disappear. I mean, they have an attendance recorder in here. They've got a lot of coffee grinders.

[A]

That. That's the drawer coming out on the cash register.

[C]

Cash. What's that?

[A]

Cash. Well, as soon as Trump signs the dollar bill, that's it for me. I'm going to coins. Oh, wait a minute. He's on the coins. Check. Oh, no, no, Petty.

[B]

I'll also shout out because it was a partially stolen pick. Semaphore had a great piece this week about the oldest job in journalism, New York Post runners, which are.

[A]

I bet Jeff knows what those are.

[B]

They are the reporters that every single day wake up, don't know where they're going to go, but they've got. They get a call and are told, hey, you got to get to the Bronx and knock on doors in this building and figure out anybody who's witnessed this crime.

[A]

Go, does that still exist?

[D]

Yes.

[B]

And they do a. That's reporters.

[A]

But I mean, but it was. It was a reporter without an assignment. It was like, you know, this is. You'll find out when they ask at

[B]

the Post, that still exists to this day. And when a lot of other newsrooms don't have this kind of specific type of assignment thing. But it is.

[C]

It's a general assignment reporter.

[B]

Yeah, it's a general assignment reporter that. No, but their specific job in the Post is to, like, physically get to a place as fast as possible and do something in person while someone else is writing something up for them.

[C]

Yeah, well, that's my job, was to be the person back at the office and rewrite.

[B]

Yeah.

[C]

And they would call in the notes. Meanwhile, I pull up the clips, I figure out what else is going on. I get the, you know, City Wire and all that, and then write a story on deadline. So I love this process.

[A]

The Semaphore piece follows one runner from the Post, Reuven Fenton, and he goes around and does a bunch of Stuff. I mean it's not just one thing in one day. Oh no, he's a busy guy.

[B]

I mean there's a. I've like met a couple of these people at various like parties or journalism events and they always end up telling some kind of wacky story about a little, A non insignificant part of their job is like when someone is murdered or dies in a very public way to track down their loved ones and bang on their door till they talk to you.

[C]

Oh, I had to do that when the plane crashed. Oh, call the first names you get are the flight crews. Oh, and so you end up calling the flight crews family and. And you know, the lie of journalism. I'm sure you want to tell people

[A]

your story about your.

[C]

Oh, Jesus, I feel ashamed of myself for that.

[E]

So are you often the person, Are you sometimes the person who has to tell them that, that, that this happened?

[C]

Wasn't supposed to be the case, but

[B]

it's certainly possible given, I mean they say in this. The nature of the runner's job is to operate inside legal and ethical gray zones. And needless to say, you can't do any of it from your desk. Quote, it's easier to hang up on someone. One of them told him. It's harder to slam a door in their face. I mean that's the kind of the purpose of this. Like, yeah, I think that this is a trap I often fall into a lot because I've grown up in this age of journalism, digital journalism, where I'm like, oh well, you know, give somebody a call, text them, you know, send them emails if you want to get the real stuff. You got to go and knock on people's doors till they tell you, no, don't, I don't want to talk to you. And that's a lot harder.

[C]

The other job was to get a photo of dead. The person to ask the grieving mom, can I, can I borrow that photo?

[A]

Are there some people who love this job or is this a paying your dues kind of thing?

[E]

The ethical gray zone. The ethical gray zone part sounds fun for some people, I'm sure.

[C]

There was one guy we had at Chicago Today, the paper that had to borrow if it was really embarrassed. He would ask anybody any question. It was like, let Mikey try it. You would send him out and he would, he would ask, ask anything of anybody the rest of us didn't want to do.

[B]

I mean there are certainly some reporters that like, that is. Their whole edge is they have no sense of shame. I would say social shame or empathy or human world. They will ask any question no matter how deeply uncomfortable it makes someone or how large of a reaction it gets from them. That could be negative. Could potentially I am always deeply concerned and anxious and wrought with guilt that I am messing someone up or re traumatizing them by not asking about something in the right language. That means I get worse stories, such

[A]

as a tweet that you fell for. That was an April Fool's tweet from the Nestle.

[B]

Okay, I'm mad about this actually. Which is I'm the biggest fan of crunch bars. It's a thing my friends make fun of me for. If they're going to bodega and you get something, I'm like, please give me a crunch bar. If I'm at the movie theater, I want to get a thing of bunch of crunch and have a little bit of bunch of crunch and some popcorn and so I see.

[A]

I like Chunkies. Do you ever have a Chunky?

[B]

We're going to get there. We've got to. I've got a one track mind. I saw this tweet. Important to note. This tweet posted March 30th, 10:13 it's not April Fool's Day day and so I got my hopes up. Says you've been pairing bunch of Crunch with popcorn for years. Now it gets its rightful spot in the concession stand. Introducing the Bunch of Crunch Concession Dispenser experience coming soon. It mixes bunch of Crunch in with your popcorn, which would be ideal.

[A]

That sounds so good.

[B]

But no, it was just a lot. It was an April Fool's lie. Not a jape, not a jest. Because it does not come on April Fool's Day. And I refuse to be considered an April April Fool.

[C]

I hate. It's my least favorite day of the year. I hate it.

[A]

I hate it. I actually, inspired by one of our club Twit members, asked Claude to come up with the base Best April Fool's Pranks for the year for this year. And there were more than I thought. Tom's Guide had an AI powered alarm clock that wakes you up by automatically brewing 8 o' clock coffee Coffee Ad Age said this year's pranks lean heavily into the blurred line between fake and real. Raising Canes enlisted McKenna Grace to unveil Cane's Sauce Coke, a fake Coca Cola variety inspired by the chain's signature sauce. ARCO Golf analytics announced smart pants and intelligent golf apparel system with advanced biometric monitor monitoring an adaptive climate control. The Royal Albert hall announced a Gen Z youth initiative called Let Them Cook. I'm not sure what that. What that means.

[C]

T Mobile had a cologne that smells like a phone.

[A]

Here is a AI sticker printing machine for Fido. It's called Pet Mode.

[B]

People kept sending me links to Indian fitness company or like some sort of wellness brand that announced a protein condom.

[A]

Oh, geez. That makes sense. Was there lead in it? Is the question.

[B]

I mean, that's. Yeah, yeah.

[A]

Put some lead in your pencil. Our very serious, legit ranking of the best live products Apple's ever made. These are all from Tom's Guide. Here's Yahoo. It's no more scrolling interface. I don't know. I don't get it. The scroll stopper. It's available in the TikTok shop. Oh. It's a thing you put on your thumb so you can't. You can't scroll anymore.

[C]

Chastity belt for your thumb.

[A]

It's the chastity belt for your thumb. No more scrolling. Here's the smart golf pants. That's not a joke. That's April Fool's launch. Big Head Mode. That's actually real. In Fortnite, they've done that before.

[B]

Yeah, a lot of games have Big Head Mode.

[A]

Big Head Mode, the big thing. Yeah,

[B]

it's a big head.

[A]

A big six speed manual shifter for your phone from Bitmo Labs. Anyway, we're glad we didn't have to do any April Fool's jokes today.

[C]

Yeah.

[A]

Jeff, your pick of the week.

[C]

Well, I just want to say that tonight is the last night.

[A]

Whatever that is. It is disgusting. I don't. What is that?

[C]

Tomorrow, Maybe I should record this for Paris. I know.

[A]

It's upsetting, isn't it?

[C]

Tomorrow, nurse Cindy is coming to take the foot and a half long catheter out of my arm.

[A]

But that's good news because you're getting better.

[C]

Yeah. Yeah. I'm not looking forward to. I'm looking forward to being gone.

[A]

So. Wait a minute. It's in. On. And it goes on.

[C]

Goes from here all the way up right over the heart because that's where the. The. The volume, the. The speed is the fastest. So the stuff will mix in. This is a pressure bulb and it cooks into my thing. And yeah, I have to spritz in saline. I have to rub it with alcohol. I've been doing this for 10 weeks. 10 weeks?

[B]

Was that 10 weeks ago?

[A]

Glad you're feeling better.

[C]

10 weeks since I got all the hospital.

[B]

10 weeks ago. You podcasted from the hospital?

[C]

Yeah.

[B]

That's crazy.

[C]

Time flies in my hospital gown. I have no shame. All Right. For a pick, college instructor turns to typewriters to curb AI written work.

[B]

Okay.

[C]

It's kind of fun.

[A]

I don't use typewriters. Yeah, great. This is a Cornell.

[C]

Wow.

[A]

Oh, but they have to do it in German, which is. That's really cruel.

[C]

It's doubly because the German typewriter keyboard is actually different. The Z is very important.

[A]

Oh, Lord, look at them. They're really struggling.

[C]

One says, oh, that's why it's called Return. Yeah.

[A]

These poor kids.

[C]

Did you ever use typewriters as a normal thing?

[A]

I think we talked about this.

[C]

Never, right?

[B]

I mean, not as a normal thing as in, like in my childhood. However, in. In college, whenever I was studying abroad in France, I would often go to the Shakespeare's Cafe, whatever it's called, and they had typewriter Shakespeare and company. They had typewriters there. And so during a time where I was at my most intolerable and I had to like, translate poems into French for some assignment, I would translate them and instead of writing them out, I would have them typewritten on there. And then my teacher would always. Professor would give them back and be like, I can't mark this up. It's too pretty. Which I always thought was kind of funny.

[A]

We did a big head mode for April Fool's Day one year in our Twitch studios. There's Father Robert and me with our big heads.

[B]

Those heads are pretty large.

[A]

This is when we had the technology. My God, look at that studio. You never saw this studio, did you? Paris?

[B]

I never saw any studio in person.

[A]

Yeah.

[B]

When are you coming to New York? And what studio are we gonna book here?

[A]

Oh, yeah, we gotta get a studio to do the show.

[B]

We need to have like three costume changes. We gotta have corn shirts. We've got to have hospital gowns. Some third thing maybe you can do.

[C]

Closes at like 1:30. When he runs out. We should be able to do it in his place, right?

[A]

Yeah, I think so. Yeah, we could do it.

[B]

But lighting and set up. Does he have Ethernet connection in there?

[C]

That's a point.

[A]

Supposedly built a studio in the. In the.

[C]

Yeah. He was going to fix sandwiches in the window, wasn't he?

[A]

Yeah, he was. And stream it all. That never really happened.

[C]

No.

[A]

You know what?

[D]

Why?

[A]

It was too much work. They have to.

[E]

Way too ambitious. That was so ambitious.

[A]

Well, no, I mean, they didn't know they were going to be selling sandwiches by the. You know, every second another sandwich. So that is about it for this show. I wish we had the big head mode now. I. We don't have a tricaster anymore. That was back in the day when we had all the kids.

[C]

That's how you do it.

[A]

It's actually pretty funny. We do this show every Wednesday right about 2pm Pacific, 5pm Eastern. That's 2100 UTC. You can watch us if you're in the club. In the club Twit Discord. Otherwise watch on Twitch, YouTube X Facebook, LinkedIn. Kick after the fact on demand versions of the show at Twitt TV IM. There's an Intelligent Machines channel on YouTube for the video. Or you can subscribe in your favorite podcast client. You'll get it automatically as soon as we are done. You will find Paris Martineau at the wonderful Consumer Reports. Are. Have you filed lately? Are you working on some massive expose?

[B]

You know, it's a great question, Leo. It's a great question with great answers.

[C]

She's not going to give them to you.

[B]

I'm not. I mean there'll be a. I'm working on a couple different stories that have weirdly competing timelines lines, but things will come out. When they come out.

[A]

Yeah. When they're okay. That's a non answer answer, as they say in the business. Yep. Yep. Anyway, thank you Paris for being here. Paris, nyc. We can work out how did your secretly British Saturday go?

[B]

It didn't happen.

[A]

I'm sorry.

[B]

It needs to happen.

[A]

No, it doesn't. No. You need to do things you enjoy.

[B]

I accompany my friend to watch him get a tattoo instead.

[A]

Oh my.

[B]

This week could be.

[A]

Oh my. Were you there just for moral support or did you join?

[B]

I was, but I might. I might get a tattoo this weekend. I don't know.

[A]

And what would you get if you got a tattoo?

[B]

I mean, that's the question. It's going to be kind of a walk in situation. We'll see. I need to.

[A]

Is this something you're going to have for the rest of your life?

[B]

I know.

[A]

Turning into a dad and you're gonna just walk in and choose it random?

[B]

No.

[A]

And he's.

[B]

Isn't that kind of fun? Does it need to be meaningful?

[A]

Yeah.

[B]

Life not meaningful.

[C]

Leo regrets his.

[A]

Yeah, but you can't. That's where the sun literally don't shine.

[B]

I already have a tattoo, so. You know, I think you have.

[A]

I've said I saw one in a picture.

[B]

Bowie.

[A]

It's of David Bowie of his head.

[E]

Ziggy Stark.

[B]

Yeah. I can't pull my arm my shirt up farther enough, but it is the Ziggy Stardust.

[A]

And why did you get David That's a very interesting thing for a person of your age to get.

[B]

He's one of my favorite artists of all time.

[A]

Really interesting.

[B]

I think just an incredible creative.

[A]

I'm a fan. I'm a fan. He used to watch our TV show. He was a fan of the Screensavers.

[B]

Oh, that's delightful.

[A]

Yeah.

[B]

Wow.

[A]

Well, that's, that's a. That's, that's a. So maybe you could get Kiss on the other bicep. Get Jean Simmons or something.

[B]

No, no, no. This arm I want to do a full, like, incorporated sleeve that I plan out, but this arm I want to do more like a little guy.

[C]

What about your acting career?

[A]

No. Yeah. What about your acting career?

[B]

All the actors have tattoos now.

[A]

They do.

[B]

It's true.

[A]

It's true. Jeff does not have is the only person.

[B]

Wow. This is, I think differentiate. This is the biggest point of conflict in our podcast history, I think is you two being genuinely disappointed. Dismayed by my tattoo.

[A]

As your parents no doubt would be as well.

[B]

I really want to. Part of the thing is I want to find someone for the little ones. I want to get more medieval wood cut style tattoos.

[A]

Oh, like a pentiment tattoo. Why don't you get Gutenberg?

[C]

You should have Gutenberg.

[B]

That's actually very interesting.

[A]

Hey, get a line of tape type. Help Jeff promote his book.

[B]

That could honestly be interesting.

[C]

Put the COVID up, Put the COVID up. It's very intricate.

[A]

So when you go to the tattoo artist, do you go in with materials or does he have a book that you.

[B]

Well, so this is the thing is, I've normally been the sort of person that would only my David Bouie tattoo. I pitched it to a tattoo artist, we worked with it. They like had some, you know, examples for what it looked like. We decided on. On it. But I've been meaning to get more tattoos for the last.

[A]

I've been meaning, like get more tattoos

[B]

years but never have because I'm like, well, like, I don't know what to do. Don't pitch it.

[A]

How old are you when you Got David Bowie?

[B]

20.

[A]

Really?

[C]

Do your parents have tattoos, if I may ask?

[A]

No, no, of course not.

[C]

What did they think about you getting a tattoo?

[A]

It's her generation.

[B]

They were dismayed. They. They probably have the exact same reaction. They often. I'll see them in the like. So do you still feel okay with your decision to get a tattoo? And I'm like, yeah, I do, actually. Thanks.

[A]

Yours is fairly innocuous. Now. Paul has a good suggestion in our club. Twit you should get your New York City tree maintenance certificate tattooed so that you'd always have it available in case of questions. A certified tree maintainer, a civic tree person.

[B]

You know, I. My editor at work had heard me bragging about it and has signed up and is going to be a licensed citizen tree pruner as well.

[A]

Wow. Wow. You're starting a movement.

[C]

You know. The tattoo gun. The tattoo device was actually a successor to an invention by Edison that was meant to create stencils to duplicate documents.

[B]

Ooh.

[A]

On your wrist,

[C]

little holes in the paper so you can make stencils to do it. And that didn't work, but instead it puts little holes in your skin.

[A]

I'm at Trees New York trying to find the. If they have a citizen pruner kind of license or anything.

[B]

I mean, I could just get like the parts department. It's a London plane tree logo leaf, which is.

[A]

That's a good. That's a good look.

[C]

That's a very New York.

[A]

Yeah, I think that'd be good. That would be. It would say you kind of want

[B]

to get a New York knife or sword sort of situation here.

[A]

It looks a little bit like a hypodermic inserting itself into a cancer cell, but, you know, other than that.

[C]

So when. When Leo does finally come to New York, we have to, I think, take him to your tattoo parlor to get a. Claude. Let's get logo.

[A]

Let's get tattoos, guys. Let's all get tattoos. Let's all get. Get brains tattooed.

[B]

Yeah, that could be fun.

[A]

I would do that.

[B]

I would do it.

[A]

I would do that. That would be pretty funny, honey.

[B]

Listen, what else is. What else are you gonna do with your skin?

[A]

Yeah, what else you gonna do? Jeff Jarvis's new book, Hot Type is available for pre order. You go to jeffjarvis.com and you could also get the Gutenberg parenthesis, which is fabulous. At Paperback magazine. His history of the magazine, the web we weave. He's written many books and. And soon he's going to be editing a series of books about AI which is pretty darn exciting.

[C]

Intelligence, AI and humanity. Bloomsbury Academic.

[A]

Nice. Thank you, Jeff. Thank you, Paris.

[B]

Shout out to everybody sending me woodcut and cool medieval tattoo recommendations that we're in New York City. Keep them going, guys.

[A]

This is hard to find good tattoo

[B]

artists on the Internet.

[A]

We are all going to get. When we go to New York City to celebrate intelligent machines, we're all going to get this tattoo of a brain on our back.

[C]

Oh, Jesus.

[B]

Oh, boy.

[A]

That would take a long Time, wouldn't it? We need something simpler. Yeah.

[B]

Corvid detective in here.

[A]

Oh, Corvids, of course, are crows.

[B]

Yeah, yeah. Really enjoy it.

[A]

Adam Savage has a ruler tattooed on him so he can measure things with his forearm.

[E]

Honestly, it goes out of scale though, right? Slowly goes out of scale.

[B]

Yeah. But, you know, don't we all?

[A]

I think it's kind of ugly. Be honest.

[B]

I mean, it's not. I've seen a lot cuter ones. There's some tattoo, like a little line, I think.

[C]

PDP 8 switches.

[A]

I'm gonna get the Bowie lightning right on my face.

[B]

If only I can. You can kind of see it if I pull up here. Ish. You know, it's a good tattoo. It's good. You should get one.

[A]

I think I actually am impressed that you would be so sophisticated as a young person.

[C]

How much did it hurt?

[B]

Not at all.

[A]

It feels like a. It's kind of a prickly vibration. It's like a bzzz.

[B]

I mean, this was just a line, a simple line tattoo, so it frankly didn't really hurt at all, is what I would describe. I think more in depth one might hurt more.

[A]

I held my.

[E]

It hurts when you feel. When you fill it in, it's. That's when it hurts. If you want to cut.

[C]

Whose hand did you hold?

[A]

I held Jason's hand. Jason Cleanthus hand. I met him because he had a twit tattoo on his inner wrist. And he was my producer for some years and he was there and I held his hand and gritted my teeth as I subjected myself.

[C]

As Lisa screamed at you?

[A]

Yeah. She was not happy. The ultimate humiliation. She's the only one who sees it, so it's. I don't see it. Thank you everybody for joining us. I'm sorry about this last bit, but I hope you will come back. We will see you next time on Intelligent Machines. Bye bye.

[C]

Bye bye.

[A]

Hi, I'm Leo laporte, host of this Week in Tech and many other shows on the Twit podcast network. Can you believe it? 2026 is around the corner. So this, my friends, is the best time to grow your. Your brand. With Twit. Nobody understands the tech audience better than we do. We love our audience and we know how to effectively message to them. We develop genuine relationships with brands, creating authentic promotions that resonate with our highly engaged community of tech enthusiasts. You know, over 90% of TWIT's audience is involved in their company's tech and IT decisions making. Can you believe that? 90%, 88% have actually made a purchase based on a Twit post read ad. No one comes close. We're the best at this. As one Twit fan said, I've bought from Twitt TV sponsors because I trust Leo and his team's knowledge of the latest in tech. If Twit supports it, I know I can trust it. You cannot buy trust like that. Well, actually you can. You can buy an ad on Twitter. All our ads are unique. They're read live by our expert hosts, Micah Sargent me. We simulcast all during the shows on our social platforms, so everybody can be watching live. You know one of our customers, Harun Meer, the founder of ThinksCanary, he's been with us since 2016. Since 2016, he said we expected Twit to work well for us because we were longtime listeners. Listeners who over the years bought many of the products and services we learned about on various Twitch shows. And we were not disappointed. The combination of the very personal ad reads and the careful selection of products that Twit largely believes in gives the ads an authentic, trusted voice that works really well for our products. 10 out of 10 we'll use again. Thank you, Harun. We love you. And it's been nine years. That's kind of. That's the proof, right? Partnerships with Twit offer valuable benefits including over delivery of impressions. You get presence on show episode pages. So there's a link right there that our audience can click on. We're in the RSS fee descriptions, a link there too. And social media promotion. Our full service team will craft compelling creative to elevate your brand and support you throughout your entire campaign. I work on the copy myself to make it authentic, to make it real. If you want to reach a passionate tech audience through a network that consistently over delivers, please contact us directly. PartnerWIT TV. That's the email address. PartnerWIT TV. Let's talk about how we can help grow your brand. Or just go to Twit TV advertise for more information. I look forward to working with you. Thanks for listening.

[B]

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