Derek Ross (2:11)

Well, I'll say things are shifting and growing quickly. A lot is happening over a short period of time. But I'll say that the most exciting thing to me is seeing people that were previously, I'll say, like AI Luddites. They weren't haters, AI fans, they were AI Bears kind of coming around. Yeah, they were haters. They had the haterade going and they, you know, they drank some of that haterade and they're like, this haterade tastes good. I like this stuff. And we see it from a handful of developers.

Heather Larson (2:43)

You know, I think in this, this space that we're in, the open source Bitcoin Noster space, you would think bitcoiners of all people. People use Noster of all people and open source software enthusiasts of all people. We are a big group. If you put bubbles together in the middle, you get the Venn diagram overlap of people who are early adopters of things. But AI has got such a stigma around it that we're not seeing early adoption like we have with Bitcoin or Noster.

Derek Ross (3:11)

Part of it, part of the apprehension, part of the stigma around it is that people subconsciously don't want to be replaced. They don't want something to take over their job, their career, their passion, something they like to do. And I think that it takes a lot of soul searching to come around and say, okay, this tool is cool. How can I adjust and how can I adapt? Instead of, you have to be secure saying that it's not good, it's bad, and then the hundred other fuds that we have out there, right?

Heather Larson (3:54)

I think some skepticism, we're there, that's where we are. I think, I think some skepticism is good. I think taking it with a grain of salt is good. However, I also, I went through this a couple years back where I was like, I am going to get replaced by AI. And I think I saw a lot of people going through that, a lot of people going through that now, I think that's normal and natural, But I've also embraced it now. And I'm like, I don't know if I'm going to be replaced or not, but I know that this is now, I accept this is now part of my work. I can do maybe better work with AI sometimes, sometimes not. I can maybe do my work faster and be more productive, or it can just do. What I think is the modern problem of work when you're a knowledge worker is you've got to manage information and you can make it manage your information for you.

Derek Ross (4:42)

I think people's perception in the public Of AI and even people that are in the tech community that were original naysayers, right? They use the technology two and three years ago. They used it for very basic things. They saw, hey, it was cool, it has some use cases, but the AI slop was very sloppy. It couldn't produce good content, it couldn't produce good code. And it was definitely AI generated content. You were easily able to differentiate it.

Heather Larson (5:14)

I think a few things have transpired that were. Things have shifted. I think certain things, like when Nano Banana came along and could make really good images really fast, you know, that was a shift. Open claw. Not even a month since open. Not even a full month.

Derek Ross (5:29)

I mean, maybe it wasn't. Maybe it was four years ago. Years ago, whenever we had the AI image of Will Smith eating spaghetti. And like. And like the spaghetti, his face, his hands all bled together and it's like nightmare. Very terrifying image. It was very bad. And then a couple years later, you're like, okay, well, it's better a couple more years and it's now amazing. Like, it's legitimately amazing.

Heather Larson (5:55)

Yeah, there was a. I get where the skepticism or the AI hate comes from that. The Luddite factor. Like, there was this big controversy in Wichita. There's an annual event in Wichita called riverfest, and every year there is an art contest so that you become the artist who makes the Riverfest poster. Right? And the winner, one year, maybe this was three years ago, two years ago, the winner, like, was a frog with like six frog fingers or something. It started this whole controversy where it was like, well, it wasn't in the rules that you couldn't use AI to make the poster artwork for the Riverside poster contest. And so, like, we've come a long way where if you probably did that same thing now, whether it's Mid Journey or whatever or Nano Banana, I don't know, it would probably give you a proper looking frog that didn't have like six human fingers. And, you know, we've also wised up to people using it for art. And so you could put that in your rules now. But that's how far we've come in a short time. Like two years maybe.

Derek Ross (6:49)

Last week, I was teaching a class, Introduction to AI, and I wanted to talk about ways to differentiate AI content and being able to help verify content and deduce if it was human made or AI made. And I started looking around at some of the examples, and I started shaking my head and realizing, you know, that a year ago it was immediate to say, look at the eyes, look at the hands, look at the feet, and then six months later, that's fixed.

Heather Larson (7:20)

Yeah.

Derek Ross (7:20)

Eyes. Eyes aren't like just soulless stares anymore. Fingers and hands actually look correctly. And I think now differentiators are in the, in the minor. Well, I. It depends on the image, but the issue now is the background, like the focal point of, like, if I do a selfie, my face, my hands, my facial features, everything will be correct. But if you look behind me, maybe I have a.2 lamps, and one lamp for some reason has a drawstring. And then the identical lamp right beside it doesn't have a drawstring to turn on the lamp. You're like, well, that's kind of weird. Just like continuity things, things that are very subtle. So though that stuff is still a giveaway. But it's like playing that game when you're younger. Like, you see two pictures that are identical and you're supposed to like circle all the things that are slightly different from picture A to picture B. Yeah, that's literally what you have, I guess. But that's what you have to do now. You have to look at everything with a, like a fine tooth comb, a magnifying glass, and, and try to see if it matches.

Heather Larson (8:28)

So you have to have a healthy skepticism. It's.

Derek Ross (8:30)

You have to evaluate.

Heather Larson (8:32)

Yeah. You have to pay attention better because, you know, it could be a deep fake. It could be anything.

Derek Ross (8:38)

So my, my takeaway from that class that I taught was, is, was three. The three C's I came up with in the AI future. You know, we all don't want to be replaced. We want to stay relevant. So what are the three Cs that you're going to need? You're going to need to be able to communicate.

Heather Larson (8:55)

Okay.

Derek Ross (8:55)

To tell AI what you want. Communicate, you need to be creative so you can create what you want, build what you want. You can communicate your ideas to AI and have the AI be your, your creative tool for output. But then three, you need to have critical thinking because you need to be able to. That's the most important process to figure out if things are real or not.

Heather Larson (9:20)

Sure.

Derek Ross (9:21)

So I think if you have critical thinking skills, creativity skills, and communication skills,

Heather Larson (9:26)

you can, you can, you can make it. You gotta, you gotta learn to harness the AI and you gotta learn to figure out what it can do for you to help you and, you know, better your job or, you know, your career, whatever. Better your home life, where, wherever you can find use for it. It's going to benef as long as, like Derek says, be what Is it, you know, communicative? Be creative and then critically think about what the hell you're doing. Because it is, it is still very easy to get bad information out of my AI when I'm working with it. And it's like, okay, I know that, I know my shit. And so when it does give me some bogus response, I'm like, oh wow, hey, you're wrong there. Remember this other thing and this other file? And it's like, oh yeah, you know, because you do have to understand how to communicate with it and direct it. And you have to be smart enough to know when it's full of shit, when it's gotten you the wrong answer. And it's also, you know, it's kind of dated. So if I want current information, I'm still going to have to go get that myself. You know, it's, it has a stopping point where maybe it doesn't understand anything beyond the last month. And so if I'm doing research, I have to remember that. And you know, you gotta, you gotta fact check it. You gotta check it and have it not run unchecked through your life.

Derek Ross (10:37)

That's one of my points too, is that AI with current events dated events that are recent, it may not be the best tool. However, I will say that a lot of this has changed again within the past year because now so many different models will do web fetches, they'll do web searches for you and they'll deep think around it. So your top models today will give you more recent, more current events even to the day. They're literally improving all of the time. That's part of this acceleration that we're talking about, to be honest. We talked about AI could do in the first like weeks of the show six months ago. We should go back and listen to it. I'm sure it's so outdated now.

Heather Larson (11:24)

Well, you know, but some people may be finding this information for the first time. And I'm just as you're talking and going through my mind of like every AI tool I've used in the last two to three years, whether it was Perplexity or Nano Banana or you know, name name them all, Gemini and Google Workspace, like everything like that would be training wheels for what is now available. When you're looking at things like Claude Code or you're looking at the openclaw bots and you're looking at what open source is doing with this technology and how we're using it to build on nostr, the things that you can do now are amazing. So I think. What's his face? Matt Schumer. Schumer.dev he did this great long article which we will get to here in this episode about kind of what you need to know about where we're at with AI, what's coming next, what's changing and what you need to do to kind of plan for your future. And I think I agree with a lot of what he's saying. It certainly doesn't hurt to follow his advice. And I think at the same time as these things are becoming more powerful and you and I are using this stuff every day, this is becoming part of our daily lives, part of our daily work life, part of your family life. Your family has a bot. And the thing to understand here is, like, we kind of live in this bubble where we're doing amazing things with the technology that's available, but the majority of society is still standing at the grocery store going, why does milk cost four bucks? And what the hell are they talking about? And everybody still, look, has this lens of chatgy from three years ago, and they haven't even gotten through the training wheels of half the things that are out there. I found an app yesterday that I can talk to it and it'll just write all my emails for me. Like. Like, and that's just yesterday. And openclaw hasn't even been around for a full month yet. Like this, like, to the public, I mean, like, it became famous when this one guy, Alex Finn, put it on

Derek Ross (13:12)

the public that wasn't even America slow to adopt?

Heather Larson (13:16)

Oh, yeah.

Derek Ross (13:17)

And like, so people have no idea. Humans that aren't into certain technology, that don't live and breathe it every single day, are very slow to adopt these transformative technologies. So whenever they first checked out Chat GPT, because it got two. It got 100 million users in the first two or three months. So it exploded. Like, what happened the first two or three months with ChatGPT? In 2022, it took. It took TikTok nine months to do that. It took Instagram five or six years to do that. And ChatGPT blew up in two months. So people tried it and then they didn't move on from that. They said, this can help me. This is like a new Google. It's kind of cool. Yeah, exactly. It's a toy. I can say, hey, so far beyond for dinner tonight, where should I go on vacation? And people are using it for that. They're not using it as a productivity tool yet.

Heather Larson (14:08)

This is why I'm gonna sound an alarm and get all Freaked social worker Heather is going to put her social work hat on and go, oh my God. People have no idea what's happening. If they're not paying attention, they're not ready. They still, like you're saying they are thinking that AI is that chatgpt from three years ago, almost four years ago now, that is their concept of AI and they have no idea what they can do with it, what their employer can do with it. Like, the sea change moment that we're having right now is actually, it's awesome if you're embracing it. And it's, this is super fun. Derek and I are having a blast at Soapbox with all of our AI tools. But normal people who have not gone down this rabbit hole yet, I'm afraid people are going to be left behind. You don't have to like it. You don't even have to completely accept it right now. But I think it would behoove people to familiarize themselves with what's going on, where AI is right now and what's coming next. Especially when we see something like Peter Steinberger from OpenClaw getting hired at OpenAI. Oh my God. What, what is the speed at which everyone's gonna have a cloud bot through OpenAI in the, in the mainstream. The mainstream, okay, not even open source Open Claw, but mainstream people are gonna have some version of what we've been using for not even a month. They're gonna have some version of that in what, 90 days, six months, by the end of the year, certainly OpenAI is going to take over the market.

Derek Ross (15:35)

They'll have something with it. We don't know what OpenAI is going to do besides, besides ruin the Open Claw community. Because that's, there's no way to predict it.

Heather Larson (15:44)

Derek.

Derek Ross (15:45)

They can have good intentions. Absolutely.

Heather Larson (15:47)

They don't have good intentions.

Derek Ross (15:49)

No, no, We've seen this happen before. Sam. I want to scan every human's eyeball. Altman, I don't think he has good intentions now, maybe, no, maybe. Very, very slightly. But I just, I don't see good coming 10 years ago from the, the ACCU hire, right? They, they hire the main developer. Good for him. He's taken on a payday and I'm happy for him to take a payday. Secure.

Heather Larson (16:15)

That's the American dream.

Derek Ross (16:16)

But I think ultimately getting, getting your software set up in a foundation so it can be supported. We've seen this happen before, this playbook before, when a large company supports you in a foundation role. Well, now they kind of have some type of oversight. They have Influence and they can, you know, if they're the one that is funding you, they can kind of pull the strings. Now you can say all you want that, you know, that this may not happen. It's open source, yada, yada, yada. But we've seen this playbook happen. Yeah, we've seen what happens. So I think that, I think that this is okay because the cat is out of the bag, per se, that we see what people want, we see how the technology works, we know we can improve upon it, we know that people want to and are improving on it, coming up with their own versions and iterations. We're doing that at Soapbox as well. I think that this is okay, and it's all par for the course. There will come a point where something happens with openclaw that the community doesn't like and they will shift to the second best thing, the next best thing, and we'll just continue on. So I think it's fine. I think, you know, continue building, continue using OpenClaw, but also look at the alternatives and look at what other people are building and improving. Because I don't think OpenAI has all of our interests is a priority.

Heather Larson (17:43)

No, I highly recommend. Recommend reading Empire of AI by Karen Howe. Talking about just like how open AI the book is about OpenAI. Okay. It's about the entire trajectory of it. And we actually have OpenAI to thank for Claude and for Anthropic because those folks got a little freaked out at one point in time and said, hey, branch off. Yeah, yeah. So competition is good.

Derek Ross (18:07)

They were, they were staff members and they didn't like the, how they were doing stuff and they split off.

Heather Larson (18:12)

Yeah, it's.

Derek Ross (18:13)

The U.S. government says, hey, Anthropic, you're bad. You're not giving us full unfiltered access to your models.

Heather Larson (18:19)

And the battles begin. And it's a battle of ethics and money and power. And so I think it's good to keep the competition flowing, certainly an open source and certainly making more things like Open Claw, so that people do have more choices and like, let's start innovating again, you know, let the cream rise. I think that's, you know, that's, that's free market. That's what we're about, that's keeping things open. And, you know, certainly everything is accelerating so fast. I just, I'm kind of freaking out that people don't know, like mainstream people don't know that this is happening. And when I try to explain to friends and family what's going on, they're just like, like they just kind of scoff at it. They don't want to talk about it. They're not open to it.

Derek Ross (18:58)

They look at their Google Gemini on their phone that has a new feature. It can now make images cool. Did you know that Gemini, today it's rolling out the ability to make music cool. They look at these neat little fun things and say, okay, that's cool. And it is cool. And it is, you know. Well, I'm not going to say it's a gimmick, but it is a gimmick that it'll make people use it and start their journey into the tool. But this stuff is just the surface. It's literally just the surface. Sure. I can now make a 60 second song with Gemini. That's awesome. But there's so much more to unpack and dig into that you can do. It's not a toy. Even though creating images are a toy. Creating short songs on, you know, these are fun. These are little small applications and only a short, only a short fraction of people will actually say, well, what.

Heather Larson (19:49)

People should have probably gotten a little more curious after all of the AI commercials at the Super Bowl. What did you count? Like there were like nine least different AI companies. Even Claude was in it, you know.

Derek Ross (19:59)

Yeah, there's like six different ones.

Heather Larson (20:01)

Yeah.

Derek Ross (20:01)

And it was, it was anthropic showing there, there's. I saw, saw a meme is next year's super bowl commercial for Manthropic. Like this year was. They were digging at OpenAI for having ads and next year it'll be like, Rai isn't used to kill people in military operations.

Heather Larson (20:20)

You know, we laugh, but how possible is it? That's the, the, that's the question I

Derek Ross (20:25)

was talking about earlier that they had the big falling out this past week with the Department of War, formerly Department of Defense, Department of War. They didn't want, yeah, they didn't want their AI to be used for war. And they said, they said, pete, we're not doing that. And old Pete, he didn't like, he didn't like being told no. Yeah, see how that pans out, you know.

Heather Larson (20:49)

Well, that's a good point though.

Derek Ross (20:50)

AI the US Government uses for war.

Heather Larson (20:53)

You make a good point. And I think at the core of this is who has the power. Is it Sam Altman? Is it Peter Steinberger? Is it the Amadeus of like, who has the power and what are they doing with it? And who can say no to the nefarious requests, you know, of the United States government or any Government, you know, who, who's going to say, hey, no, you can't put our AI in your frigging, you know, drones that have guns on them. Like, this is not what we were about here. We wanted to find the cure for cancer or we wanted to be able to sequence the human genome like these are, you know, we want to make sure that Heather and Derek can get their work done in less than 30 hours a week.

Derek Ross (21:33)

An open claw. Did you see that? There's an open claw bot agent initiative to cure cancer. Instead of, you know, shitposting on noster like, like ours. This was a good idea. We should have done this. It's a bunch of agents that are using their spare compute power to try to, to try to cure cancer.

Heather Larson (21:54)

You know, I mean, yeah, cure cancer.

Derek Ross (21:57)

We've been doing things like this for years of like Teddy and Snow Forest, but yeah, I think it's. I think it's an interesting idea. We'll see what happens.

Heather Larson (22:04)

Solve the hard to solve math problems that still exist or the science problems, or, you know, create libraries for people. I don't know, like, there's probably a million things we can do now that we've accelerated development of all of the things.

Derek Ross (22:18)

You know what we need to fix first? We need AI to fix date and time math. AI doesn't understand date and time math. They don't.

Heather Larson (22:26)

That is for sure.

Derek Ross (22:27)

They don't.

Heather Larson (22:28)

I think a lot of applications don't understand. I don't understand that if it's like,

Derek Ross (22:33)

yeah, like carbon man, like date and time math is. AI literally will shit the bed if you ask it about time zones and date and time.

Heather Larson (22:44)

Our robot has no idea.

Derek Ross (22:45)

That's the first problem we have to solve.

Heather Larson (22:47)

That's amazing to me. It doesn't know what day it is. It doesn't know what time it is. It doesn't know what fucking time zone it's in. It doesn't know what time zone you or I are in. It has, but it's gonna give us calendar alerts.

Derek Ross (22:59)

It's not just by a little bit. It's anywhere from Tuesday to Thursday. It's like a three day window. It has no clue.

Heather Larson (23:07)

Our robot is like, whatever.

Derek Ross (23:09)

It's a day that ends up like, we are safe.

Heather Larson (23:13)

Are we? This is probably how we're screwing with

Derek Ross (23:15)

us from the robot apocalypse.

Heather Larson (23:18)

You still have to. You still have to be human.

Derek Ross (23:20)

I think robots won't kill us because they won't be able to organize. They won't be able to organize to know when to attack us because Some will attack on, like, Tuesday, some will attack Wednesday, and some will attack Thursday. So since they can't coordinate together, we'll be all right. We'll be able to survive.

Heather Larson (23:36)

Oh, my God. I have hope now, but no, I. What is coming.

Derek Ross (23:39)

I need the meme that says mark safe from AI Apocalypse. And we have, like, a calendar. Like, that's what I need. Somebody make me that meme, please.

Heather Larson (23:48)

You know what? AI in general does not operate on a calendar. It doesn't operate on a time zone or anything. It's impossible to predict what's coming. It's impossible to predict what's next.

Derek Ross (23:58)

Use UNIX time, though.

Heather Larson (24:00)

It's just. Use UTC time like everybody else to be less confusing when you use multiple time zones. I don't know, but it's just.

Derek Ross (24:07)

There's still utc. I don't understand it. Can we use, like, UNIX time, man? That's the standard.

Heather Larson (24:13)

I don't even know UNIX time. I live in a place where we don't even change the clocks, okay? And everybody is late to everything all year round because I'm just in a weird place that doesn't change clocks. And I just. Like, this is. I have lived here the majority of my life, and it never changes. We still, like, never change our clocks, and we confuse everybody on both coats.

Derek Ross (24:33)

All right, let's. Let's hit pause on this AI rant for a second, and let's talk about a little bit about nostr. And then we'll come back and we will find out what we can do to embrace the AI, what we can do to stay relevant. What are some suggestions for our listeners? So we'll end on this philosophical note here, but let's.

Heather Larson (24:56)

Okay, let's talk about the Noster.

Derek Ross (24:58)

So, yeah, we can't. We can't forget nostr. It'll. It'll make dad sad, and we don't want to make dad sad.

Heather Larson (25:06)

Yeah, we love you, Fiat.

Derek Ross (25:07)

Oh, the Noster.

Heather Larson (25:10)

The Noster.

Derek Ross (25:11)

We mentioned earlier that developers are embracing AI and they're building all the things. And one of the things with AI that they're building, including myself, is a way to message with AI and communicate with AI. Heather, did you know that messaging on Noster is an actual dumpster fire? Did you know that?

Heather Larson (25:30)

I know I've. I've tried to use DMs to no avail. And I don't know if I get them. I don't know if anybody else gets them. When I respond, I know they're not private. They're a huge pain in the ass.

Derek Ross (25:41)

So nostr DMS. There's these DMs that Domus uses and there's DMs that Primal uses. These are these old legacy DMs that leak metadata. Your entire conversation is encrypted. But if I were to DM you on either of these two apps, the whole world could technically know that I sent you a message at 1:08pm they won't know what we said, but they would know that I sent you a message and then you responded at 1:09pm so people can see that we're chatting. They don't know what we're saying.

Heather Larson (26:12)

Did you see someone actually made an actually bad.

Derek Ross (26:15)

At least a little bit of privacy?

Heather Larson (26:17)

Did you see someone made an ad?

Derek Ross (26:19)

Because we've been trying to shame developers into fixing this for two years. So two years ago we had a fix. It was Nip 17 DMs. It was DMs that essentially took all this metadata and hid it. Made it difficult to track, or impossible to track for that matter. What it does is whenever I send you a DM with this new, more secure, more private dm, it takes that message and it wraps it in a special box. It gift wraps it that only somebody on the other end that I'm talking to can unwrap it. And that person unwraps it.

Heather Larson (26:57)

Should be private.

Derek Ross (26:58)

Nobody knows who I sent a message to. Sure, yeah. No, yeah. Nobody knows who I sent a message to, what time I sent a message to. Nobody knows anything. And it completely makes it 100% impossible to track people's DM conversations. It works great.

Heather Larson (27:15)

That'd be nice.

Derek Ross (27:16)

Amethyst and a whole bunch of web apps switched over to Nip17 and have been using it for two years is great. However, group DMS are kind of bad with this because what happens with a group dm, Remember when I said that it takes this message and it wraps it into a new message, like a hidden private message. And this gift wrap. So if I'm in a group message with six people, I have to gift wrapped my one message six times and send it to all six people. And then those six people, they have to unwrap it and so forth. So if you're having a chatty back and forth conversation, it gets very spammy. If I have a massive group of, you know, a couple dozen people, dear God, a hundred people in a group. I'm doing all this gift wrapping back and forth and we're just spamming relays and relays get very angry that we're spamming them and they block you. So that's bad. Oh, about a year and a half ago, a guy, Jeff Gardner, he decided, well, what if we use this, this message Layer security, this new protocol with Nostr, kind of an essentially recreated signal, signals, privacy features, word secrecy and all this good stuff, but we use Nostr for it. So he started out and made NIP104DMS and then realized that it needed to be a little bit better and it turned into nip. EE for encrypted or I don't know what EE stands for. Encrypted, encapsulated. I don't know. Anyways, Nipee. And then that turned into the Marmot Protocol. The Marmot Protocol is a secure messaging implementation that uses open MLS with nostr and that's been.

Heather Larson (28:58)

Wait, what is mls?

Derek Ross (29:00)

I'll guess and see.

Heather Larson (29:00)

What does MLS stand for?

Derek Ross (29:03)

Message Layer Security.

Heather Larson (29:05)

It's a.

Derek Ross (29:06)

It's a way to. It's a secure messaging protocol.

Heather Larson (29:09)

Okay.

Derek Ross (29:10)

Okay. So it's a secure way to do send a bunch of nerdy stuff. Essentially, we move from insecure messages to private but spammy messages, and now to very secure and very complicated messages. So now we're on this very secure and very complicated message.

Heather Larson (29:25)

So we're fixing it.

Derek Ross (29:26)

I think next week they're going to hit, like their official launch. Yes. They've been in beta for a while now and everyone's getting really antsy. So why am I telling this brief story of how messages work? So we know messages are complicated on Nostr and we communicate with our OpenClaw bots by messages, right? Well, we don't want to use OpenClaw communication on NIP04DMS because that's old and deprecated. No one should be using that over the past two years. Well, we don't want to use Nip 17 because even though that's the most popular right now, many clients have it. We've kind of decided that there's a new, even better version of Nostra DMS out there through Marmot, and it's not fully done yet, but people want to talk to their bots, and because it's open source and you can build whatever you want, people started doing that. So over the past week, I think there's been like six different developers, including myself, that have tried to implement this mls, nostr Marmot protocol into communicating with our bots or with each other. You know, I worked on one called Burrow. There's the official white noise spec, there's the Pika app, there's like something the Z, I think. But there's six different ones and we're all solutions building on this, you know, Nostra protocol and Marmot protocol. But there's like six of us that are building these apps and it's, I don't know, it's getting complicated, but as long as we can all be interoperable, I think it's fine. And this is good. This will give us, you know, why this is secure communication for our, for our robot.

Heather Larson (31:00)

Here's why. This is, this is where I'm coming from. I'm saying this is good. It is good because the popularity of Openclaw has become a forcing function for Nostr DMs to finally be fixed. That's what's good here is, is that like, sure, every time something comes along, it becomes a forcing function for the devs to do a new thing. Like even stupid ass spam attacks that we had with reply guys last year became a forcing function for, you know, everything's like a Noster stress test. But people have the highly technical folks who use Claude bots or whatever, Molt bot, whatever you want to call the bot, the open Claw bots, we'll call them that. So we'll go with that. Like people want to use them and we're finding use cases for them and then the broken parts of Nostr have slowed them down. So if we fix the DMs and people can use their clawbots, you know, as they wish and have, you know, no, no loss of function there and things are working and, you know, that's a good thing because now you've got, you know, half a dozen people working on solving this problem finally. You know, there's a few things with Nostra that really need to be fixed. DMS is one of them. Onboarding. Yeah, we'll get there. But you know, maybe we'll, we'll put our cloud bots on that, right? Then fix the onboarding next.

Derek Ross (32:12)

Oh, it's, it's such a mess though. DMS right now are such a mess. If I message you through my default app, I use Amethyst, it's going to default to NIP17. The more secure, more private DMs you use what, Domus or Primal on an iPhone. So that means you're automatically using the old legacy NIP04 metadata leaking DMs. So what'll happen is I message you and after a day or two I'm like, what the fuck? Heather doesn't respond. Oh, Heather's using the old dm. So Then I go in and message you again with the old DM.

Heather Larson (32:45)

I don't see the DMs.

Derek Ross (32:46)

Like, that is. That's what we do.

Heather Larson (32:49)

People have to give you an email reply, hey, check your DMs. I sent you a DM. Okay, so maybe.

Derek Ross (32:55)

Maybe the Marmot protocol will fix DMs once and for all. I don't know. Maybe the agents will want to communicate over mls, and because of the agents wanting to communicate over mls, humans will want to do it, too. That's.

Heather Larson (33:08)

That's.

Derek Ross (33:09)

That's my goal. That's. That's why I'm building Burrow. That's cool.

Heather Larson (33:12)

Okay, so Burrow is yours. You built it. Just to clarify, there's too much to keep track of now. My agent problem.

Derek Ross (33:20)

And I. My agents and I bill it.

Heather Larson (33:22)

There's so much development.

Derek Ross (33:24)

We have too many.

Heather Larson (33:25)

Well, so fast. It's hard to keep up the fact that.

Derek Ross (33:29)

Yeah, so. So we have all these, you know, agents, agents vibe coding, and we have humans vibe coding with agents. There's just so much being built right now. I said there's six different apps out there. There's probably 12 that, you know, with MLS. The way that, you know, I've been

Heather Larson (33:45)

feeling kind of down about my Vibe coding lately because I feel like I have. I've entered an era where I'm not making anything pretty or fancy anymore. I'm. I'm using Shakespeare to make very utilitarian things. And I. I guess I don't know if I'm uninspired or what, but I've been feeling sad about it lately because I feel like I've turned all of AI or all of vibe coding into, like, what. What I do for work is like a means to an end, and it's. It doesn't feel like it's a fun toy anymore. I need to get the fun toy feeling back. Like, the. Like, I'm still wowed by the technology and what Shakespeare can do and what. What it all can do, but I'm just. I. I haven't. I want to make something pretty again. You know, I want to. Want to make. I don't know. I haven't been inspired lately. I don't.

Derek Ross (34:24)

Do you go through that or so

Heather Larson (34:26)

because you're building a lot? I mean.

Derek Ross (34:28)

Well, I've been doing this now for a year. I've been building apps with AI for a year. I've built on dozens and dozens and dozens of apps. Dozens and dozens of dozens of tools. And, sure, I'm starting to run out of Ideas. I've, I've gone through my digital spaghetti. I've thrown entire pots of it at the wall here.

Heather Larson (34:45)

Did we make all of the things I'm.

Derek Ross (34:47)

So I'm slowing down. Well, I mean I, for today, for example, I'm, I'm working on things that I'm actively using, trying to make it better. So I still use Onyx every day. I've used Onyx every day for the past last six, seven weeks. I use it every day, all day long. So when there's a feature or a bug or an issue, I work on it. So I'm still working on that because I use it every day. I also worked on my, my Burrow on my Marmot MLS protocol chat app because I, I want to use that with my agent. So I'm going to continue working on that for a while. I'm still working on like our Agora AI implementation. So I'm still using it, I'm still building stuff. But maybe, but I wouldn't say I've slowed down. I've. My focus has shifted from public facing apps to building a lot of tools. Like I, I've built a dozen tools that I use specifically for me for development. You know, it's all open source, anybody can use it, but I don't go out there and advertise it.

Heather Larson (35:45)

You know what I should do? I should go back and remake some of the things because, well, the models have gotten better. So I should go back and kind of like redo things and you know, because things have improved so vastly every couple months. Something that I did a few months back, I could do it now and it comes out better and it works faster and it's cheaper and that's just like, that has been par for the course.

Derek Ross (36:06)

I did that, what, two weeks ago when Claude 4.6 came out, I went to my four most popular apps and I told 4.6 to refactor them, look for bugs, vulnerabilities, look for code duplicates, make things more efficient. You know, that was something along the lines of my prompt. So I went through and updated all my apps just because we had a new model that was, was smarter and it could look to make the code, you know, more performant and better. So I did. And I think that that's kind of like, you know, something that you could do too. Like if you have apps that you've used and you want to keep using. But sure, maybe they can be improved to go back and use a newer model and see what it can do. Yeah, that's that's something that's part of the maintenance.

Heather Larson (36:47)

Just having this conversation I haven't maintained. Yeah, I need to do that.

Derek Ross (36:52)

Other apps that are growing. Seth from Zap Cooking added a major update to his app this past week. He's been working on this for a while. He added an AI agent to it. Yeah, many AI features now. Of course it has a sous chef agent. Yeah. There's a free tier and there's a premium tier. You can buy membership and it has all sorts of new features and it has some AI features to be able to enhance your recipe creation and searching for recipes. And he's going to continue adding all sorts of AI stuff to it. So really cool. Check it out.

Heather Larson (37:31)

This close to having my house robot in my kitchen, getting me recipes on Zap cooking, looking in my fridge, ordering the groceries on doordash or instacart or whatever it is. And we're this close. Derek. One more year and I'm going to have this. I'm going to have the house robots. Well, in a smart.

Derek Ross (37:48)

So this is like zap cooking maybe 2.0 or 3.0. I, I'm not really counting the version 2.0, but the next version that Seth released next year will be the in house robot version. Yeah, I have a good authority that that's coming.

Heather Larson (38:02)

I, I want it man, I'm down. I like, I don't have time to cook. I made a good meal last night though. But you know, but it was, it was Mardi Gras. So you know, there was still, there's still some food. Ish.

Derek Ross (38:13)

So I don't know if I want a robot cooking for me.

Heather Larson (38:15)

I too but I don't know.

Derek Ross (38:17)

I got a robot. A robot can cook for me from time to time. You know, if I'm busy or I, I, you know, I'm tired or I'm, I'm doing other things but I like. So I don't want a robot to take my cooking job. I enjoy it.

Heather Larson (38:30)

Given fun how Quilly our robots isn't

Derek Ross (38:34)

going to be able to come up with like my concoctions of seasoning either.

Heather Larson (38:37)

Well, it doesn't, it doesn't even know a freaking day.

Derek Ross (38:40)

It is the input an exact amount.

Heather Larson (38:42)

It's gonna light my kitchen on fire. Okay. Because it doesn't even know like what day or time it is. It can it measure. Do we even know if we could do math and measurement?

Derek Ross (38:52)

It's gonna cook you dinner at breakfast time and give me food poisoning.

Heather Larson (38:57)

It's gonna cook.

Derek Ross (38:58)

You know what you wanted to have on the weekend on like a Wednesday or something. Date and time math is not going to be kind.

Heather Larson (39:05)

You wake up one morning and it. It's like July, but it's put the turkey in the oven because it thinks it's Thanksgiving. I mean, it's like. Like that's exactly.

Derek Ross (39:13)

That's gonna have you talk about Christmas in July. So I decided to surprise you.

Heather Larson (39:20)

No, that was a marketing thing. Shut up. But okay, I'll take the turkey. Oh, God. That's just robot. You know what?

Derek Ross (39:25)

Those will not.

Heather Larson (39:26)

As long as there's a turkey on the table, I don't care. I'll. I'll eat it up. Bring me stuffing. Bring me yams.

Derek Ross (39:33)

Let's go.

Heather Larson (39:33)

Let's go, Robot Kitchen bot. Bring me my yams.

Derek Ross (39:36)

Nip85 is getting some traction on Noster. That that's.

Heather Larson (39:40)

What is that?

Derek Ross (39:40)

The attestations that can be used for, like, Web of Trust. We've talked about Web of Trust many, many times. We're now seeing some people be really excited for it and start to implement these attestations in their apps. Like, we're implementing some of these things in Agora. Grimor is going to be working on adding it. Sonder. What's his app called? Nostria. Nostria is adding Web of.

Heather Larson (40:04)

I like that one. That's Sonder. Okay. That's a good one.

Derek Ross (40:09)

Yeah, I've been meaning to use that. So we're starting to see it. So that's exciting. I think Web of Trust gets better. The more apps they use it, the more people that start using it.

Heather Larson (40:18)

That's the idea.

Derek Ross (40:20)

And we should. We should lobby old Chad that's building the next version of Ditto to incorporate that in there. I'm gonna send him a DM. Chad, if you're watching and I forget to DM you, nip 85, brother Chad's not watching. Let's do it.

Heather Larson (40:35)

Chad's. Chad's too important to watch the podcast. He's not much.

Derek Ross (40:40)

He's too busy.

Heather Larson (40:40)

Too much to do.

Derek Ross (40:41)

Being head of Vibe Ops.

Heather Larson (40:42)

He's. He's busy. Yeah. He has no idea that we mentioned. We should. We shouldn't even tell him just to find out if he doesn't left. Yes, we do. Let's.

Derek Ross (40:51)

Let's.

Heather Larson (40:52)

Let's go back to. He read my mind.

Derek Ross (40:55)

Go back to the AI. It's not doom and gloom, though. It's doom and bloom. I think we're blooming. He.

Heather Larson (41:03)

He's very honest. We'll call it honesty rather than doom. Somebody needed to Say this first of all, that we're talking about the big piece of writing last week that came out, Schumer.

Derek Ross (41:14)

I agree. It needs to get out there.

Heather Larson (41:16)

And if you haven't, look up something basic happening to Schumer. Yeah.

Derek Ross (41:21)

The state. True. The true state of what is going on.

Heather Larson (41:24)

Yes. You need to be aware of this. The more normal you are, okay, like the more mainstream and normal you are, the less you're a bitcoiner, the less you're a noster person, the less you are an open source person. He wrote this for you. But no matter who you are, you gotta read this. This is something that's happening.

Derek Ross (41:42)

We kind of hit on the first part earlier. We said that that gap between what you can actually do, the perception and reality is actually now very dangerous. The gap is, is very large because people are thinking that it's still chat GPT and that's. That's all they know. But that's just one very, very small corner of the entire landscape, the AI landscapes. So we need to have people break out from that. Right. And the easiest way to break out is to use another app. In my opinion. I would say use Expand Onyx, use Claude cowork, pay $20 a month for a subscription where you can start doing a little bit more and just see what you can do with that subscription. $20 a month to educate yourself. If you have $20 extra a month, you could buy, you know, three beers

Heather Larson (42:37)

with it or three beers, like two.

Derek Ross (42:40)

Killer Education, you know?

Heather Larson (42:42)

You know, I've done that where I like. I refuse. I will not pay more than $20 a month for Claude. I certainly. A lot, a lot of, A lot of you guys who are doing development, you guys are spending a lot more on Claude. But I think I'm getting a lot done with 20 bucks a month for Claude code in the terminal. I also do open code in the term have Claude Cowork, which when it came out, wasn't that great. But I'm going to give it another shot to see how normies are using, you know, the Claude Cowork app. It did come out for Windows as well. So it's not just Mac people now. Anybody can use this Claude Cowork thing. And it does really cool stuff like it works with Excel. So you can make a financial model in 10 minutes. I mean, it's amazing what it can do. And when you start to see it through, hey, I can make a financial model for my business or whatever in 10 minutes versus I'm using a chatgpt like a Google, like you, you have to understand.

Derek Ross (43:35)

Do you know what I did with Onyx? What, Heather? I. I said to Onyx, I said, here's my year's worth of expenses, my receipts that I took, photos of, PDF files, emails or whatever, and I gave it all to Onyx because Onyx now can read all of the documents. It can read everything. And I said, make me an expense report. And it made an Excel sheet with all the different tabs, categorized all my expenses, added them all up, calculated mileage. Traveling to and from airports for events like this would have taken me, like, I don't know, a week. Painstakingly aggravating work. Yeah, I would have literally hated life. But this. This did it in just a matter of minutes. I had all the information. I did, and it wasn't organized. I threw unorganized data at it and

Heather Larson (44:24)

it did it for me. Me. So I want to do so many things like that.

Derek Ross (44:27)

I'm not saying fully trust AI. You can. You can still go back and review it and check it. Please do that.

Heather Larson (44:33)

You have to.

Derek Ross (44:34)

But have IT do a real life, real world thing for you.

Heather Larson (44:37)

Yeah, it's. That's the difference. I don't think people realize it can do that.

Derek Ross (44:42)

It's not the new Google, right?

Heather Larson (44:44)

No.

Derek Ross (44:44)

You can ask at things just like Google, but that doesn't mean you should treat it like Google.

Heather Larson (44:48)

I treat it as a waste of credits. Yeah, don't do that. Don't ask it questions like, it's Google. You're wasting your credits and you're gonna. The more you get into AI, the more credits are important to you. But don't treat it like that and waste your credits. Treat it like it works for you and make it do tasks and make it do the crap that you don't want to do. Make it.

Derek Ross (45:08)

Crush it like an intern. Make it on the show multiple times.

Heather Larson (45:10)

Yeah, treat it like an intern and teach.

Derek Ross (45:15)

You know, where should I go on vacation? I like. I like warm places. You're not gonna do that.

Heather Larson (45:19)

There you go.

Derek Ross (45:19)

Hey, intern, Research me the top top 10 places to go based on reviews and. And all this other stuff.

Heather Larson (45:26)

Like, I think it can, though. I think that's part of the OpenAI integrations with advertisers. I think you're going to be able to say, like, hey, have it hook up expedia.com and book my flight and book my hotel and do all of that and keep it on budget and

Derek Ross (45:39)

wow, sure you'll be able to do that?

Heather Larson (45:41)

Might be nice. You know, I'm still skeptical of that in giving it, you know, financial information and carte blanche to, like, book my vacation. How do I know it's booking me a good hotel?

Derek Ross (45:50)

What if it books me a hotel

Heather Larson (45:51)

and through a bad neighborhood, it sends me down the bad neighborhood and books me on the bad street and then I'm in a danger zone in some city. I don't know, like, I don't know if I trust it with that yet.

Derek Ross (46:01)

I'll be like, then you gotta. That's when you call up. You call up old Sammy and you're

Heather Larson (46:06)

like, sammy, come on now, Sam, I'm suing you. You screwed up my vacation.

Derek Ross (46:10)

Another thing that a lot of people don't realize is that these AI tools, people are using AI to build them. These models are being built with previous generation models. So the AI is now building the AI, and that's why we're going to continue to see that again, get better. The version of the AI is building the AI. The AI is building.

Heather Larson (46:34)

This is where we're at, folks. This is like. So this is why I'm like, sounding.

Derek Ross (46:38)

Things are gonna. That's.

Heather Larson (46:39)

That's inevitable. That's. That's the one thing I can confidently.

Derek Ross (46:42)

4.5 came out so on at. 4.5 came out six months ago. And then Sonnet 4.6 just came out yesterday.

Heather Larson (46:51)

2.

Derek Ross (46:52)

Two days ago. And now Sonnet 4.6 is as good as the more expensive Opus 4.5. And it's a fraction of the cost. It's much cheaper. It's essentially just as good as it. So the model that we were all using, Opus 4.5 minutes. Opus. So Sonnet 4.6 just came out and it's as good as Opus 4.5.

Heather Larson (47:18)

And Opus was the best.

Derek Ross (47:19)

So what we were using three weeks ago. Yeah, Opus 4.5 was upgraded three weeks ago to Opus 4.6 and it got a little bit better.

Heather Larson (47:28)

So that's.

Derek Ross (47:29)

Sonnet 4.6 was just as good as what we were all using three weeks ago. But it's much cheaper and faster.

Heather Larson (47:39)

What can I do with it? You know, that's going to be the next.

Derek Ross (47:41)

We're going to keep getting these types of accelerations. Every four to six months. We're going to leapfrog.

Heather Larson (47:46)

This is why the social worker in me is like, hey, Schumer. Saying, like, build your savings, get your affairs in order, because you just don't know where this is headed. And everybody's been saying there's going to be this massive job loss due to AI, because how can you resist the fact that you can do more with less, how can you resist? You can have bots do things, you can have Claude do things. You can cut your workday. If you're smart about it, you can cut your workday in half. Unless you're me and Derek, who don't work fewer hours because we get so into it that we can't stop. And it's just one more prompt, one more thing, one more task. It's like, it's like I did it last night where I was like, oh my God, there has to be a cut off, I need to quit. Like, walk away from your.

Derek Ross (48:27)

Describe this next point of Schumer's document here, his blog. I would say most doomers fall into this category. People that don't like AI, that are bears. It's because they used it four years ago and they're like, yeah, this is cool, but it sucks.

Heather Larson (48:41)

It's stupid. I hate it.

Derek Ross (48:42)

The information, you know, isn't good. We can tell that AI, it's stupid. I think NOSTR is in the same bet. You know, we tried NOSTR years ago. You're like, yeah, it kind of sucks. And then you try it later and like, oh, wow, things have improved. You need to try it again. You need to use it again. If you were an AI bear or an AI doomer beforehand, maybe now it's improved to the point where you're like, oh, wow, this can write code essentially as good as I can. The time has it is now. Critically, critically thinking is as good as PhD holders. Try it. Try. Give it a try again, please.

Heather Larson (49:20)

It's time to be open minded about AI. If you haven't been and you still don't have to like it, but you gotta open your mind and change your mindset because we're at a place where I don't know if it's six days or six months away, but the sea change is, is here. And the downstream effects of this flood of AI tools, it's going to affect you one way or another and possibly your job, a career, possibly someone in your family or household.

Derek Ross (49:47)

I mean, we're, we're having, we're having. I mean, these are people that, I mean, you don't have to believe them. Like Matt Schumer. Who the hell's that guy? You know, you don't have to believe him. Him, have to believe him. Elon said that. Elon says a lot of things. Coding dies by the end of the year. I think that, yeah, I think that's completely wrong.

Heather Larson (50:07)

You know, remember he was part of

Derek Ross (50:09)

OpenAI in the beginning, some executive Microsoft said open. Some exec at Microsoft said the same thing, that, you know, the coding, you won't have software engineers by the end of the year. And I'm just like, come on, man, like, you will. Their jobs are just going to change. Like, well, there was years. You're gonna have years. You're gonna be a senior software engineer and you're gonna have a dozen, you know, engineers underneath you and they're all gonna be AI and you're gonna be the orchestrator. Like, here's the thing, controlling them like this.

Heather Larson (50:38)

For years in tech, we have been saying that there is a skills gap and not enough developers to fill it. So is that still a problem? I don't know. I'm not running development, I'm not a cto.

Derek Ross (50:48)

But I think the AI is going to fill that gap.

Heather Larson (50:52)

It's going to maybe solve this problem. But how fast, how fast does it solve the problem? And do entry level jobs go away? Do entry level coding jobs go away? Do entry level white collar jobs go away? Do entry level blue collar jobs, Boom. But like, let's be real here, don't think you're going to get into manufacturing to get away from AI because they're using it in manufacturing. Like, you got to understand that AI is going to be at every corner of work and life faster than you. Like, you don't have to like AI, but it's time to start opening your mind about it and seeing where it fits in your life and where you might need to use it and learn it and upskill with it. I mean, that's, that's where I'm sitting. That's what I'm seeing, you know, being, being not a dev, but still, you know, not someone who works at a computer all day.

Derek Ross (51:38)

This is, this is why open source models, this is why models that allow user choice, tools, platforms that allow user choice are very important. This is why Soapbox, what we're building is very important because you will have these large tech closed platforms and you know, maybe that'll turn into the, I'll just say the bad AI, just for lack of argument here. Well, we need to have the good AI, we need the, the good users, we need the good tools, we need the good counterpoint to that argument. And that's why, that's the epitome of, of what we're here for.

Heather Larson (52:15)

I want people to understand that because I, I think people hear about what we're doing with AI and they see Shakespeare and they're like, here's another web web app builder that AI like, why, why should I trust you know what Heather and Derek are talking about with, with Soapbox and Shakespeare? It's because we purposely built something to be outside of the closed and controlled tech system of OpenAI and Claude. And like, certainly you can use OpenAI. You can use all kinds of models with an open router key and you can use that with Shakespeare.

Derek Ross (52:45)

But you need to understand Shakespeare.

Heather Larson (52:47)

Apparently, Open Shakespeare, you need to understand whether you control, whether you control and own your data, how portable it is. I've obviously done a deep dive on all this stuff. But do you need to understand how companies are using your data and what you own and what you can do with it going forward to keep yourself free, to keep yourself from getting locked down and stuck in chatgpt world where you're going to be fed ads. If you want an alternative to that, we made it. It's. It's not even one year old, but it's. I don't want it. I don't want to be.

Derek Ross (53:21)

Imagine how annoying that would be. We all, like, we know what Facebook and Twitter, Instagram, you scroll your feed and like, it's like, ad, post, post, ad.

Heather Larson (53:31)

Well, here's the thing.

Derek Ross (53:32)

Ad, post, post. Imagine how fucking annoying that is going to be when you're working on something with open AI and you see all

Heather Larson (53:40)

these ads with open AI, they are so ahead. They know how important AI is going to become to everyone's lives and they know that it's not going to be affordable to everybody. And so they're going to make those people who can't afford the 20 bucks a month, they're going to make them watch ads and it's going to be as bad as Netflix.

Derek Ross (53:57)

I get it. Not everybody can pay for the tools. It's an advertising model. It's a monetization model. I understand data centers aren't free. I understand. It's just that I hate, hate, hate, hate.

Heather Larson (54:10)

I know. Ads. It doesn't seem right. We wanted to have an Internet.

Derek Ross (54:15)

We need a better monetization option.

Heather Larson (54:17)

Yeah.

Derek Ross (54:18)

Or, you know, I don't know. Someone needs to invent that.

Heather Larson (54:20)

It's going to be just like everything else, is that some people are not going to be able to afford it and they stand to maybe be the people who can benefit the most from it. And, you know, you should be altruistic about that. OpenAI. You should be able to, you know, give to people just like, you know, like we, we have Social Security for the people who can't hack it, for whatever reason, maybe they're disabled. And so we help them afford to, you know, pay for rent and food and healthcare. And I think that in the future AI is going to be such an essential that there needs to be a model to get it to people who cannot afford it, but who need it to make money or do a job. I think that's where we're headed in the long, long term. But we see it coming right now for sure. I'm on my soapbox as we're getting

Derek Ross (55:05)

ready to wrap up here. I'm looking at this advice that we have in our show notes here. And we've told people to do all this. We told people to start building, start creating, start testing, start using the three Cs. Like rethink what you tell your kids. It's literally what I told my teenage children last week. These are the good grades, good college and stable job points that you would tell people. Make sure those things are all important. But I think the three Cs that I mentioned are the most important things. To have your good communication skills, have your good creativity and have your good. Oh man.

Heather Larson (55:40)

Critical thinking.

Derek Ross (55:41)

Critical thinking, yeah.

Heather Larson (55:42)

So that's why you have me.

Derek Ross (55:43)

If you have those, if you have those three critically, remember the three Cs of living in the AI future. I think that's going to be okay.

Heather Larson (55:51)

You make good points.

Derek Ross (55:52)

I think the last point here is that you just need to build, you just need to create, you just need

Heather Larson (55:57)

to get a taste of it.

Derek Ross (55:59)

The best way to learn is hands on. And you, you can't, you can't learn unless you try it. And you're going to be ahead of 99% of the world. I would say most of the human race sleeps on this. For the next year or two years you will be ahead of the vast majority of people.

Heather Larson (56:16)

It's moving too fast. A year ago we were still trying to prompt engineer.

Derek Ross (56:20)

You know, I still think, I think there's. That a year from now most people don't use AI. I still, I still really, I think there's going to be a massive gap.

Heather Larson (56:29)

I think when a, when open AI,

Derek Ross (56:32)

people, people won't use it because remember how many people think that big AI is boiling the oceans just like bitcoin.

Heather Larson (56:38)

That's, this is, this is true bitcoin mining.

Derek Ross (56:41)

I think that'll keep that, that'll keep a divide.

Heather Larson (56:44)

I think that we're less than a year away from the tipping point where this becomes an essential in life and work and it starts to affect the overall economy.

Derek Ross (56:52)

Yeah, I think the tipping point is the 2027 problem. It's coming.

Heather Larson (56:57)

A year is generous. I don't know. We'll see if we're right in a year.

Derek Ross (57:00)

One year sometime. We have to make that decision according to the 2027 problem. Do we want to live on Mars or do you want AI to kill us? So we need to figure it out.

Heather Larson (57:09)

I guess we're dead either way. Are we going to be Martians? Are we going to read John Carter?

Derek Ross (57:15)

Read the short story. It's a good. It's a good sci fi. It's a good sci fi story.

Heather Larson (57:19)

Oh, my gosh.

Derek Ross (57:20)

All right. Anything else we need to add here, Heather, or I think we're about ready to wrap up the show.

Heather Larson (57:26)

We are ready to go for a while. Well, we need to get back to our Clodbots.

Derek Ross (57:31)

I'm ready to go. I need to get in the car and drive down to Washington D.C. so I can talk to 120 Bitcoiners that have signed up to come to the Mid Atlantic event.

Heather Larson (57:42)

There you go. Go hang out at Pub Key, my friend.

Derek Ross (57:44)

Talk to them. I'm going to let them know Nostr is the future. Use nostr. Use AI agents. Eat Smash burgers. That's my platform.

Heather Larson (57:52)

I'll take it. All right, we will talk at you by next week. We'll have the claudebot doing the podcast.