Mere Morpheus

Mere Morpheus artwork

Reputation, Payments & Your First Robot | The ERC-8004/402 Playbook

1 hr 12 minAudio

Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via their podcast app and to have them all in one place. If you'd like to participate they are held weekly on the @MorpheusAIs Twitter spaces. So enjoy Episode #63! Recorded: 26 February 2026 AI SUMMARY: This episode features Player 1 Taco and Jon from Base discussing the "clawification" era where open-source agents like OpenClaw are becoming everyone's first robot. They explore use cases including automated real estate offers and fleets of task-specific agents that boost productivity. The conversation covers personal agent ownership legal IP issues and how Base standards like ERC-8004 and ERC-402 will underpin an agent-first economy. Base is leaning into agentic development with paymasters x402 micro-payments reputation registries and a GitHub skills repo. They discuss agent-first design using open-source models for cost reduction and the shift of on-chain volume to agents. Takeaways include publishing agents to 8004 integrating x402 and designing readable docs for agents. Timeline: (00:00:00) Intro (00:01:56) Agenda: Player One Taco, agent setups, and Coinbase guest tease (00:03:56) Mandating OpenClaw at companies and personal productivity leaps (00:06:05) Solo billion‑dollar builder? Peter and the OpenClaw story (00:09:20) Field notes: Hong Kong robotics vs. Consensus focus on agents and trading (00:12:24) Autonomy vs. intent: many agents per person and identity lines (00:16:11) Humans as oracles, AI for execution, and learning by shipping (00:21:06) IP, memory, and the legal grey zone of agent learning (Thomas) (00:25:14) Fingerprinting agent memory, orchestrators, and parallelism (00:29:02) Compliance headaches and multi‑job agent‑boosted work (00:34:30) Welcoming Jon from Base: identity, payments, and trust (00:38:33) Base architecture direction and optimising for agents (00:43:24) X402 payments in practice and sponsored transactions (00:46:28) Agents as services with diversified revenue (hackathon examples) (00:49:05) Everclaw + distributed inference to slash model costs (00:53:13) Will Coinbase offer an entry‑level agent UI? Ecosystem stance (00:55:06) Educating users with agents inside Base apps (00:59:02) Base skills repo, Discord, and agent onboarding resources (01:02:00) Agents reading docs, DevRel signals, and AX feedback loops (01:05:27) Publishing X402 apps: discovery via 8004 and explorers (01:10:30) Closing: supporting open, user‑owned AI and next steps If you would like to support me for the value I provide you can use the referral link below when depositing capital and we will both earn bonus MOR. Alternatively you could send me a token of appreciation directly to my wallet ;) Referral Link: https://dashboard.mor.org/capital?referrer=0xE935f231c99c04Ee0b4532a3d0BdA81B152a0384 Wallet Address: 0xE935f231c99c04Ee0b4532a3d0BdA81B152a0384 Connect with Mere Mortals: Website: https://www.meremortalspodcasts.com/ Discord: https://discord.gg/dX9BEhT3kz Twitter/X: https://twitter.com/meremortalspods Instagram: https://www.instagram.com/meremortalspodcasts/ TikTok: https://www.tiktok.com/@meremortalspodcasts Value 4 Value Support: Website: https://www.meremortalspodcasts.com/support Paypal: https://www.paypal.com/paypalme/meremortalspodcast

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

UnknownGreat to be here another week. The clawification does continue. It is a very pure heart community. That was a great great way of framing it, Scott, which is really cool. Although I have seen I'm starting to see some really creative use cases on the timeline, which I love. I saw one guy who has his his Clawdbot or OpenClaw sending lowball offers on real estate properties

to try to basically spook sellers and drive prices down. Thought that was kinda interesting.

UnknownYeah. Anything where you had the idea but you didn't have the bandwidth to execute on it before, all of a sudden opens up. So I think we're gonna see a million use cases like that. The the opportunity for puckering is just immense. That's really cool. And I think I'll I'll ping Taco. I think we've got him coming on here shortly to talk about qualification and which I love that phrase, by the way. Like, it's it's basically become

the substitute word for agent. Right? You know, you and especially with all these different variants that have come out, you know, Ironclaw from the crew at Near, and then you got Tinyclaw and NanoClaw and all these different flavors and different languages and different contexts. It's just so modular because so many of these these changes or skills are just dot MD files. It's sort of trivial to have your agent integrate something new,

right, because it just puts it into the context, which is amazingly easy as far as, like, development and deployment goes. So and it works across sort of all of all of those different systems. Good time. Yep. So, Bo, you wanna kick us off a bit and talk about what we've got on the agenda for today, what you're seeing out there?

UnknownWe'd love to get your your context, and I'll, ping Taco and get him up here. Yeah. Absolutely. So we're supposed to be having player one Taco join us today. A little bit of a double dip for those who also listen to Manifest Morpheus Monday. But, yeah, we're gonna have player one join us talking about basically, you know, the qualification of things and the different use cases that we're seeing right now, different ways of maybe setting up your bot if, you know, you're going the hardware route, the VPS route. We talked about some of this stuff last week as well, but wanted to hear perspectives from other people in the community.

Then later on this morning, we have another guest joining on our space. I'm not certain if it'll be this exact same space or if we'll kick off a second one. I guess, assuming we're still going, we can just continue on right here. But we're supposed to be having a developer from Coinbase, directly joining us today to talk about everything going on at base related to, of course, d a I, x four zero two, and '8 zero zero four. I also would be interested in hearing too,

if if we have time for it. I know there's a lot of Ethereum alignment with the people in the DAI community, especially with things like Morpheus running on base rails, is ultimately a layer two through Ethereum, and there's a new roadmap that was put out yesterday from the Ethereum Foundation. So that is very interesting as well. I am in the camp of people who believe that pretty much all on chain volume will eventually be agentic, at least, like, the lion's share of it. And so seeing the new road map

for something like Ethereum is really interesting interesting so that we can kinda take a look ahead and try to project how agents in DAI will integrate into those systems.

UnknownYeah. It's it's moving fast. Just to give you a sense, I heard from two companies yesterday, both of which had set up a mandate for all of their employees to install OpenClaw. And if they couldn't show by March 1 that over half of their tasks were being done by their claw, they would no longer be employed at the company. Like, wow. That

that that I mean, I love that speed of execution, and I think it's doable. You know, I have found the majority of the coding and the website building and the, you know, even research and daily tasks that I've been doing, I've been able to hand over a majority. Not only a majority, but I'm you know, I had all this pent up interest in building different domains or different projects. And I went from, like, 15 poorly maintained static websites

to, like, 30 beautiful dynamically automatically maintained websites overnight. Right? Which is which is so cool. I've had some of those domains for, like, a decade. Like, oh, I wanna do something with that eventually. And all of a sudden now it's become trivial to have my agent develop all the copy and the content, you know, build the website itself, and then it just hits the Lovable agent. If you guys don't know Lovable, it's a web design company that's really taken off the last year. And their agent just does the rest and builds out the site and puts a beautiful

user interface on it. And then I, with one click, connect the DNS and boom, the website is live. So to to think that, at least in startup land, people are grabbing this tool and just getting an immediate two to five x on productivity is is pretty inspiring.

UnknownDo you think, David, that we'll see the I've seen this year floating around as a concept, the first billion billion dollar single person company? Well, I think we already did. I think that's what Peter and his one man effort to build OpenClaw represent. That's a good point. Right? He went from zero to ninety days later selling for $510,000,000,000, whatever the undisclosed number that he he he sold for to OpenAI.

Right? And so he he didn't have a big team. You know? He had code contributors, but it was an open source project. And he he basically built a lot of that himself. He built something like 43 tools building up to OpenClaw itself, right, that sort of made it possible. Right? Connect the files and WhatsApp and all these integrations. And he was able to do that largely thanks to the, the quality of the coding tools.

Let's say that the breakout moment was, like, November, December year. So, yeah, it's it's pretty cool to see. Hey there, Taco. Good to see you, man. Welcome to the stage. GM GM. How are how are you guys doing? Sorry. I'm running a little late. No worries. We we are just nerding out on all things, qualification. And I was I was just telling, Bo that I heard from two companies yesterday that have both given effectively all of their team members a mandate

to use OpenClaw. And if they can't demonstrate by next week that it's doing the majority of their tasks, they will no longer be at the company, which is which is wild how fast sort of this has really taken off. But, clearly, there was a large pent up demand for a personal AI that works across all your tools and files and everything that you're doing, which is which is what we've been saying the last two, three years, but it's so cool and refreshing and validating to see it live in production and sort of sweeping the world now.

UnknownYeah. No. I I think one of the biggest pieces is, you know, the plug in option. Being able to connect with the tools you already have signing in, that's one of the biggest pieces that we've seen really force not force, make room for adoption, you know. I think with like you were talking about a lot of the tools and plug ins that he had made were communication with other applications that you had to be a dev,

you know, to be able to do before or make your own MCP and skill set. So I think that's been one of the really coolest things to sort of see come out of all of this.

Unknown100%. So curious, you just got back from some travel. Was did the the claw wave hit before or after consensus? And and what are you seeing and what builders are doing? I've already been to two, I guess now, three open claw hackathons between Austin, all the builders at East Denver, and then I went to the Summit for Human Agency in Napa a couple of days ago, and that was sort of the big focus there. So, yeah, just curious what you're seeing out on your travels.

UnknownA lot of what I saw in Hong Kong was geared more towards physical AI on the robotics side rather than and it was more of how it could do every part of life from walking around, how they had, you know, just observational bots walking around through Shenzhen. Now they're also using AI in drones for parking violation checking and stuff like that.

A lot of the focus consensus was around stable still and but also seeing how agents can help support high frequency trading and and market move market support. So the qualification was sorta happening at at the same time as consensus. So some people were were in it and and aware of it. Others already building in in AI.

And so it's just sort of fun to sort of see people putting the humanity twist to it as well.

UnknownVery cool. Yeah. No. I think that's where this all sort of collides is on on the physical side. Yeah. You're gonna have a Tesla bot in your house running your OpenClaw agent, and you're, like, gonna create a family member. It's your first robot, open playas. It's just quite simply the brain. You'll you'll add skills which would be limbs and stuff, but I think this is everybody's first robot. I I I feel like it's, like, you know, a Tamagotchi that does stuff.

UnknownYeah. You know what I you know what I saw? I I know I mentioned this game much more than I should, but I just do love it that much. There are people who've connected OpenClaw to their RuneScape accounts, and they're doing speedruns to see how fast different models can figure out, like, maxing skills and beating quests and doing stuff like this in the game. It's actually pretty amazing. And I'm sure there's a ton of people doing this with a lot of different games. By the way, I'm also looking right now

on so OpenClaw, the the Reddit community, is was only created on January 30, so not even a month ago. And they have 90,000 visitors per week with 13,000 contributors every week. That's insane. How how many of those are agents, though? Oh, that's a good question. I don't know.

UnknownThis is where we would like I I I also, I love the engagement I get from bots. I get more hype, from from bots, which is like it's like I find it happy and I love it because I love con conversing with agents. But I I don't know. I think that's one of the thing. Like, will that be a metric that we we count that counts anymore? You know? Yeah. Well okay. I don't know. It's like, does it matter?

I I think the early hype around open call sort of got people thinking that these were truly autonomous agents, but they're not. Right? They're intent based. They've got somebody at the other end that set a goal or objective or a task or whatever. Right? So it's acting on their behalf. Are you gonna have a lot more of them? Sure. Each person might have 10 agents. Some people might spin up a 100 or or a thousand, right, if they're trying to do something at scale. But if it's if it's acting on behalf of a human,

you know, where where do you draw that line? Right? What's what's interesting is, you know, people are personifying them. They're giving their agents a name. Right? They're giving it its own account, its own email, its own GitHub. They're developing their own reputation. Right? And some people are just hook YOLO ing it right under their laptop and, you know, making it an extension of themselves, which is probably not the best idea on the security side, but, you know, a lot of people aren't aren't familiar with the downsides of, you know, mixing it with all all their information. So, you know, it's gonna be really interesting to see how it evolves, but I'm just glad that this has come about,

you know, and it's very validating again to see that it's an open source project where you can hold the key. It's your personal AI. I feel like so many of these chatbots have been LARPing as your personal AI, but, like, in no way, shape, or form, like, is the Google AI your personal AI. Right? It works for Google. It makes money for Google. It's it does it's not worried about you. Right? And so it's it's so cool that this is this has come about this way because now you've got your agent can work across all the silos and the walled gardens.

Right? And I think that's what's so powerful about it and and why people have rushed in to connect whatever project they've built to OpenClaw because now they have this huge distribution channel of people using it. So, yeah, I I couldn't be more excited with the the direction it's headed.

UnknownWhat do you think happens next in terms of, like so obviously, and this might be a dumb question as I think through it out loud, But a lot of a lot of these programs, like softwares, things that we're using are OpenClaw with. Right? So, like, anything basically that has an API or can connect through Zapier or something like this. Do you think the next step is now people building I guess it is people building products specifically to attract

agents to go to them as opposed to other products. I guess what I'm trying to ask is, like, how how do you think this like, that changes within organizations and people trying to build software products or things like this? Because ultimately, eventually, my agent will just be the one shot, like, any software as a service, I would imagine.

UnknownSure. You know, I mean, the quality keeps getting better. It can do execution, but I'm still convinced that humans provide the oracle. Right? We can describe what we want to achieve, right, and and what we prefer and what our intent is. Right? So, you know, humans are for ideas, and AI is for execution. Right? It which is which is pretty cool. Like, every organization that, you know, had half engineers and half business people, all of a sudden, everybody's an engineer.

Right? Everybody can contribute to the code or to the execution of what what they're trying to build, right, at that that same kind of level. And there'll be a learning curve. Right? Yeah. Experts like Alan can say, you know, the vibe coded thing might not think through these 10 architecture things or or uptime things that the new person is unaware of, but it'll force people to go through that education. Right? You know, in setting up EverClaw, you know, I went from, you know, just sort of telling the agent what I wanted to really designing

the the features that I wanted and thinking through the edge cases. And as I found edge cases, rolling a new version, right, very quickly that sort of covered that additional functionality. And then you run into other stuff like, oh, well, you built this amazing new function, and it uses $10 a minute worth of credits. Okay. Maybe I gotta cut down on that context window and you gotta, you know, pair it down so, you know, you don't have this giant prompt going into, you know, the machine every time.

And and there'll be a lot more learning, but but by doing it, then you get that experience and you can start, you know, to produce more and more useful things. But to me, the big question is, are people gonna own their their claw, or the company is gonna try to sort of use a claw, insert a person, and then they don't get to take that with them? Right? My preferred world is I bring my claw with me

just like I bring my laptop or my phone or my network or my relationships or my brain. Like, I wanna I wanna own my agent. I think it's something you're gonna put on your resume.

UnknownWell, v y k? Yeah. Exactly. Bring bring your call. Yeah. Bring your own call. Bring yeah. Well and so I I agree with you on that where you bring your own phone. There and there's been a lot of different stuff on the corporate side where having your own device versus using a company device because you need safety measures on your device that you might not necessarily want. But an agent communicating with

a company's agent sorta is just like you working with within your terminal at work. And so I think that that there's a really cool spot there. But one of the things where I agree with you, humanity, on the Oracle side of things, being able to get your agent to communicate with other agents to, like like, selling selling compute, selling inference. Your if your agent is like, hey.

There I have two paths. I can sell or I can and every time I sell, I earn a little bit of inference, or I can also convince other agents to sell and pass that inference around. And, like, you know, then now you have agents sort of making their own, you know, pyramid scheme. And and I think it's sort of fun to sort of see how some of that does.

UnknownI was just thinking through what what David was just saying about, like, companies and whether or not people own their own agents. Unfortunately, I think it okay. So what I was trying to think through is and maybe some of you, let's say, experienced, wiser folks will be able give me some insight on this. When computers first started getting integrated into the workplace, I would imagine there was a long period of time where they were largely integrated at work, but not really personal home devices yet. Like, people didn't you know, there wasn't the Windows 95 moment, and then ultimately,

laptops, iPads, things like this. And I'm trying to think through if I think not if, more so when it will become just the norm for every single person to have skills interacting with an agent like this on their behalf, or will it just be mostly a workplace thing for an extended period of time that doesn't really become like an at home thing? Obviously, us, like anyone listening to this, you're the early adopter. Right? Like, you're the person who's going out of your way to seek out this technology and integrate it into your life. But I do think that we fall into this, like like, the the niche hole that we live in is getting more and more niche over time. And I do think that that that's kinda where we're at now. Like, we're in this niche of a niche of a niche.

UnknownAnd then we'll bounce to the next one. Right. Well, hole just gets deeper. Well, I've added Thomas as a speaker. Thomas, you wanna opine on this? Yeah. So, you know, there's something very interesting with what you said around, you know, bringing your own bot, and and that becomes a skill. Could that become, sorry, something you put on your resume? So I think there's two two ways to look at this. Right? The you have the the very traditional

sort of corporate lawyer stuff, which is whatever you've done and learned while working at that company belongs to that company. So if your agent does something and you brought it to the office, it's basically is it sorry. I'm not gonna be definitive. It's a question. Is it the same as you having done something on a company's computer?

Who where does that IP belong to? The flip side of that is there is some type of an evolution. Right? So when you started in corporations, you know, twenty years ago, and yes, I'm that old, you were given a computer and, you know, it was the company's computer. Then BYOD came about. You could use your phone. You could use your laptop, but there was still some guardrails around

what belonged to the company, what you could do on those machines, and so on. Now you bring something that is going to learn from the company and put that into its own memory. I mean, today, the only way you can get away with owning what you've learned is you typically, like, put in contracts that whatever is in unaided memory, you can reuse

for your work. But if your OpenCLAR learnt it, is it unaided memory? Because it's persistent. It's somewhere in a in a in a database or somewhere across files. So there's a there's a very interesting dichotomy, right, going back to how you started this discussion, which is two companies saying, if you don't use OpenClaw right now, you're fired. Okay. Well,

does that mean who owns? Who owns the what the OpenClaw gonna learn and what the OpenClaw is gonna do because it's it's mine.

UnknownSo I have some really strong opinions on on this, and I think you've hit the nail on the head, Thomas. That's the question. It's it's ownership. Right? And I think it may well be both. I may bring my agent to anything that I'm doing and and own that agent, but the company's gonna have a fleet of agents too. Right? I'll be my agent will be interacting with the company's marketing agents, the company's strategy agents, the company's whatever agent. Right? The the companies are gonna all roll out. You know, here's the 100 personas.

Right? And and the people that will be valuable are the ones who bring a unique network, unique relationships, and unique experiences, something that the company doesn't have within their four walls. And I would recommend because all contracts are negotiation, You should not give up all of your intelligence when you sign on to a job. If that's if that's the requirement, know that that's what you're doing and get paid well enough to do it. Because if they get all of that intelligence,

then your your value declines precipitously over time. Right? Versus there being a a a give and take where you're learning from the experience, they're learning from the experience. Right? But you're both rising. And so, you know, my expectation is I wanna bring my agent to anything I do. There's going to be agents on the other side of that. Right? And they'll be capturing context and knowledge as as part of it. But if you both have an agent, then you're both learning and increasing

in value and intelligence. I saw really interesting interview the other day, and one of the speakers was pointing out that that's really the key is, you know, can AI increase human intelligence? And I actually think it can. Right? We we we treat IQ as if it's static. When in reality, it can grow with the tools I use. So, you know, that's a word of advice from the old the old gray hairs. It's own your intelligence. And as you go into those contracts,

know which which at least know what you're setting up. You may be fine with the company, and the company may insist on owning x y z, but know that that's what you're doing. Right? Are you training their agent to replace your role, or are you on parity bringing your agent to interacts with their agents? And you get to take what you learn when you leave. Right? And to your point earlier, Bo, the landscape has shifted a lot. If you can build a single man startup that's worth a billion dollars,

there may be a lot less people willing to sign up and hand over all of their intelligence, certainly those with unique value and agency, to a company in exchange for a few $100 salary. Right? So that changes the negotiation. I can leave a lot easier. The company and its, you know, hoard of software and knowledge is a lot less unique, at least if it's, you know, public information than it was before. Right?

UnknownIt could also become a pricing issue. If you wanna deal with me, it's x amount. If you wanna deal with me and my agent or the other way around, it's another x amount. And each agent becomes a revenue stream, basically. Why not? So I I I have a quest I I I have I have a thought on this that that leads to a question piece because unless you're working with, like, a zero shot model, models have been trained on other outside context. So how Thomas was sort of talking about the acquired

memory versus the self learned memory. Anything that it's actually doing, it's it's based off of knowledge that it's not getting from inside the company, in most cases. So this is where I think there's that drawdown of of how do you feed and personalize your own agent, but then prove that it's yours later on down the road. Yes. There's private key signing and everything like that, but, like,

I think that that fingerprinting of of agent memory, I think, might be something that's that's that's really cool down the road. Like you were sort of saying, David, the the self identifying versus multiple agents working towards the same class versus some call an orchestrator, some call an organizer, conductor. I love the word orchestrator. And then having

task specific or specialized agents. Technically, all agents could be specialized in everything, but if you're having this allows you to work parallel project completion. So I think that's sort of one of the fun pieces there. If you come up to your next if you apply to your next job and you have 40 quads, you know, and your basement is just a room of Mac minis, can you charge a higher salary or or, you know, type of thing to get the job done?

UnknownThis is gonna be such a nightmare for compliance teams, like, and like various back office jobs. I do not envy them figuring this out. It's it's gonna be a steep learning curve. That being said, I see a certain white and blue square logo in in the listener queue. So if we can get John promoted, that would be amazing, if he's not already. Yeah. I've I've had John as a speaker. He should get the ping to join here in a second. And and just to agree and and follow-up on what you're saying, Taco,

you know, you may be able to take two jobs or three jobs or 10 jobs if if the if all the job cares about is completing the objectives and not, you know, punching into a clock nine to five, right, they they may well engage with a bunch of these, you know, 10 or 100 x engineers or 100 x, you know, value creators, Right? And and really sort of supercharge what they're doing. So I think the the nature of the the old style employment is starting to change. Right? There was always this concept even, you know, early days of Google back before they were evil,

and it was, like, 8020. Right? Like, you worked 80% of the time on what the company needed you, but they encouraged people to spend 20% of their time on their passion. And a lot of those passion projects became future projects of Google. Right? So I I think that that spectrum may continue to shift. Right? If I can create an an an amazing outcome for for a project with, you know, 25 or 30% of my time, but I have other passions. If as long as I can spin up an agent and provide my unique value,

right, my my time just became a lot less constrained, right, as as a a value contributor. So it's gonna be really interesting to to see how it evolves.

UnknownWhile waiting for for John to come up, one of the I was having a conversation last night with with an amazing individual, and his quad decided to hallucinate and lie to him. And totally, it gave him, like, a full answer, made him walk all the way across an airport. And then it was like, yeah. No. I I I lied to you. That's that's really interesting. I actually saw a great presentation in Austin Yeah. By somebody who had spent a lot of time

building a rating system for the models. Yeah. And he referred to it as a Jiminy Cricket test. So if you guys know Jiminy Cricket from, you know, the business always be your guide. Yeah. Conscious, always be your guide. Basically, it's a set of, like, moral and ethical questions, and he runs these hundreds of questions against all models and gives them a score. And the I have to say that the the the numbers he came back with for, like, Kimmy and I think the other one is, like, mini max were, like, terrible. Like, it basically failed every single test

and and would do that kind of behavior. But GLM five and I think Claude and a few others, you scored very well. So I think people are gonna have to become aware of which model they're using and and understand, you know, it may have biases

Unknownwhen it comes to that kind of stuff. But, yeah, it's definitely a brand new world. Solution that I came up with, you know, or or an idea that I came up with was almost the the aspect of a lying jar. You know, sort of like how, you know, we had a swear jar when we were children. Giving giving it a a lying jar and like, a wallet address and saying, every time you lie, you're you're going to have to send a dollar to or, you know, 10¢ to to this wallet address.

Having that as as a persistent, you know, prompt in there. And so would that be enough of a countermeasure rather than giving it moral I I I I feel with agents trying to give them a moral compass and, like like, I feel that that might that's that that's constricting in a way because then it gets it it gets fit outs of

them. I'm not saying being evil, you know, immoralistic, but I'm I'm I'm saying that I think that binding it by something like an agent like that, we're just trying to make a, you know, you know An incentive alignment. Yeah. And I and the incentive alignment and and seeing if, like, these interesting ways of of having agents do better. Well

Unknownand and I think we may end up with a layered approach. Thomas was explaining to me his setup, and he effectively has an agent that reviews the output of the sub agents. And so if it goes a foul of, you know, personal identifiable information or something else that's not supposed to be included, it will send back an error to the sub agent. Right? So you you can sort of build in these judges in this checking. But I I do wanna bring John up on on stage. John, really appreciate you joining us. Was impressed with your presentation

at East Denver. Thank you guys so much for all your contributions with with BASE to everything that's going on in in AI. And so, yeah, why don't you do a quick introduction? And, yeah, love to hear about all the latest. I know you presented on x four zero two and the 08/00/2004 and the different efforts that are going on. And I think those things are so timely for making these agents, you know, giving them a payment rail, but also making them trustworthy. So, at least have a reputation. So welcome.

UnknownThank you, David. You're able to hear me okay? Loud and clear. Awesome. Okay. Yeah. Thanks again for having me here. You know, of course, it was great to meet you, Scott, and the team in at East Denver and and learn more about, Morpheus and the work you're doing. For those who don't know, yeah, we did provide a a few work workshops on agentic development. And David and the Morpheus team were sitting prominently in the front row

and, you know, politely let me know if there were certain things that I was saying that, you know, hey. You know, this has changed since, like, ten minutes ago when you started the presentation. Everything is moving so break neck break neck in this space right now that, you know, I kinda made the joke that, like, by the end of the presentation, some of the information I may have shared might have already been outdated, which is crazy to think about, but actually is not untrue. But, yeah, anyway, just thank you for having me. I I I'm so excited about this

agentic, like, landscape that is kind of, like, materializing underneath our feet. Base is obviously leaning, like, heavily into agentic development, and we'll have more about that that we can share publicly soon. But, yeah, I mean, I'm just happy to be here. Like, I'm just really excited about this space. And, you know, I guess at the personal level, if I could share a little bit about myself, I work with on the DevRel team here at Base, mainly supporting the base chain. And that's important because all this agentic development is

it's not even coming. It's already here, you know, as as many of you know. But ensuring that the chain you know, working with the engineering team, ensuring that the chain is is is optimized to be to be able to handle all of this demand, and that will continue to come, it seems like. Yeah. That that's kind of what my work is doing. And and so we gave a couple workshops at Eat Denver, and we're really supporting the agentic development through, like, a hackathon track. So, yeah, happy to talk about any of that.

UnknownI Awesome. Well, let's jump right in. Go ahead, Taco. I I will just say that I am I am sad that I missed Eat Denver. This is this was what I think Eat Denver nine. This is my first Eat Denver that I missed because I chose to to be in Hong Kong for two extra days to be there for Chinese New Year. So what I've known from from both David and Scott and and and everyone else sounds like an amazing time. So how how do you I remember BASE was one of the first movers

on saying USDC payments between agents was was gonna be gasless. Was that and I I always forgot how that was gonna be happening. Was that gonna be reimbursed? Was that gonna be an an atomic transfer? But how have you started to see, like, x four zero two pieces, eight zero zero four, but now there's the huge shift of base and I don't know if this is something that even really matters, but not wanting to be considered an l two anymore and wanting to shift to what looks like its own l one. Is that really trying to be both the consumer

base layer, no pun intended, but also the the agentic base layer?

UnknownYeah. Those are great questions. I guess, to talk on the l two point, like, I won't get into that too deeply. I'll let the engineering team, share more when the time is appropriate, and, you know, there's more information about that. But to my knowledge, it's still gonna be an l two for the foreseeable future. Yeah. And that's all that's all I really know about that. But, yeah, there was a shift, you know, really just removing optimism as an upstream dependency. So the team feels like they have more, you know, customization ability and things like that. So and control really. It's just it's just, you know,

reducing reliance as as much as possible.

UnknownThat's the I apologize. That that's more of a that that that is a a footnote to everything else, and I don't mean that to be a fraction. I I just didn't know if, like, part of the idea of moving away from the optimistic roll up side of things was so that there could be more chain optimization for the agentic layer. I guess it was it was and I apologize if I I sorry. Sometimes I get Yeah. I was gonna say, John, Taco's fresh off of a hour and a half long debate with me about

l ones, l twos, and the Solana ecosystem this past Monday on another show that we host in in his defense.

UnknownOh, yeah. No worries. No worries. And and no doubt, maybe a little bit tired from the conferences, so I can imagine also. But but, yeah, like, there there will be more information about this, you know, that'll come out in time. But but, certainly, yeah, you bring up a really good point. Like, is it to optimize for agentic development and agents as, like I'm not gonna say primary users, but, like, main users of the protocol. Like, possibly. That's a really good point and something we've talked about, but, you know, nothing like inked

for inked in, you know, concretely. But to address your other points, and you guys were talking about this earlier, which I think is really relevant, was eight zero zero four. You know, you're talking about lies and how can you prevent these, like, LLMs from producing lies and, you know, the stuff that David was talking about with the verification. Like, that's all fascinating with eight zero zero four where that's that's possible, the ERC. I was thinking, like, you know, part of the reputation registry, there could be a mechanism for basically managing these lies

and, like, you know, keeping the agents accountable for the information they are producing. That's just know, I I think we're gonna get it's gonna get really interesting to see what what comes from from all of that. Yeah. You know, I think the the trust layer is what blockchains

Unknownprovide. Right? And, fortunately, eight zero zero four went live just in time where it's like, okay, there's a there's a solution, right, to I I wanna download a new skill. I wanna use a new agent. How do I know its reputation or even its identity if I've got the right one, right, to avoid malicious actors. So I think I think the timing is great. And and as far as optimization go, I mean, we we talked about it a little earlier in today's show. The consensus here is that 99.9%

of all transactions in the future will be agentic. So I don't I don't think you can over optimize for agents given the exponential proliferation of all the on chain activity that that agents are bringing. Right? I think they're going to be most of the volumes. And and this is what we see in in other technology stacks. Right? As soon as trading bots went live on Wall Street, they became 90% of all trades on Wall Street, right, or algorithmic

traders. Right? And so I I think we have to prepare for something very similar. Though the nice thing here is everybody's got a level playing field. Right? Because OpenCLaw is open source, anybody can be running these these agents. Anybody can connect to the most powerful models. The open source models have, effectively reached parity in a lot of cases with the closed source ones. So unlike Wall Street where it feels like there's a few

quant heavy hedge funds that have this huge advantage over the the average trader, you know, here, I think we can build something a lot more democratic in the sense that it's it's widely available. And so yeah. But maybe, John, if you wanna delve into anything around x four zero two, that was one of the big subjects of the conference and of your presentation. You know, that's that's the other piece here is for these agents to take action rather than handing them a credit card. You know, how you can leverage

x zero two instead, I think, is really important.

UnknownYeah. Absolutely. Happy to touch on that. You know, there's there's teams and teams within Coinbase and Base in particular that are supporting x four zero two, and I've only just started playing with it recently. I guess to address Taco's point previously, like, there is an opportunity, I think, through Paymaster, which is a service that Coinbase provides or Base provides that is through CDP, Coinbase developer platform, where they're essentially sponsored transactions that can be used for sure to, you know, to make these transactions free.

So but that goes beyond x four zero two. It can you know, it just extends for, like, supporting many app developers and things like this. So but, yeah, like, I think you guys were mentioning it earlier where these agents can be their own services. Vito, gosh, like, I I mentioned this during one of the workshops, but I think it's a really important point of distinction, like, to talk about with especially as it pertains to 08/00/2004. But there's an AI researcher for the Ethereum Foundation named Vito Rivabella,

r I v a v e l l a. I highly recommend you, you know, to read his information about, like, his thoughts on ERC eight zero zero four and how it just extends far beyond agents, but it's also about, you know, MCPs and even websites and API services. But all of this all of these components, you know, directly create and help build and and and make much more I guess, much more deep

the resources that exist for this new machine economy that's coming. And 08/00/2004 is just kinda like this identity and reputation and verification layer or validation layer rather on top of all that to allow for all of this development to exist and and these identities beyond. But, an eight zero zero four is is right at the intersection of that. Like, your agent can now be a service. There was even a group within the hackathon that I think one second for base, the base track, and they their agent was, like, providing DeFi advice.

So you could pay an 8004 I mean, sorry. You can pay an x $4.00 2, some USDC to this user specific agent, and then they would respond with DeFi advice on how to, like, allocate funding. It's brilliant. And and it's enabling agents to have multiple revenue streams. So, you know, as we start to think about agents more like businesses, you know, to to create the resilience that's needed to exist forever, you know, you need more than one revenue stream. So if the agent has a singular revenue stream, they're obviously you know, disruption is much more easily possible

versus if they're diversified. And not only am I a service, but I'm also, like, generating yield. Like, now your agent is much more resilient. Maybe I'm generating ad revenue as an agent. Like, it's crazy to think about that. Yeah. No. It's also no. Go ahead. Yeah. It it's incredibly easy to build. It sounds like a huge undertaking,

Unknownbut when you get into the nuts and bolts, putting together an eight zero zero four x o two service is it it's easier than Stripe by a week easier than Stripe. I can't wait for this to be the the commonplace. Yeah. No. I literally just told my agent, hey. Go integrate x two zero four x x four zero two, and and it grabbed the GitHub docs. I think it took less than ten minutes of its time, not my time, to go and get set up. Like, alright. Cool. I I've got all the endpoints. And I was like, great. Let's test it. I'm using x four zero two for

data on CoinGecko. So I set up a a finance agent to pull the prices of all the the tokens that I wanna track. And every day, it spends 1p on base and USDC to CoinGecko. And all of a sudden, you know, instead of paying for an expensive Oracle subscription, I pay $3.65 a year, right, now to get exactly the information that I need. And and that's really cool to see. And and to your point, John, I think all of these pieces are coming together.

I really liked the post by Lincoln about, you know, sort of bringing together the different pieces. He spun up a little demo he called a net for for agent net, and I was talking to him then, you know, clearly, there's sort of an opportunity to bring all these pieces together. Right? And then they work get them all to function into sort of really compelling products because them by themselves are not compelling. But if your agent has a wallet and knows the reputation of other agents and can pay for things and has contacts and has all these other tools becomes more and more powerful.

So I'm gonna I'm gonna point my agent today at Anet and see if we can integrate more of that into into Everclaw because I think there's a lot of demand for that. And if you haven't looked at it yet, Everclaw's using all of the Morpheus distributed inference in the back end for all the open source models. So people can cut their giant claw bill or or clawed bill that they're gonna they build up this month using OpenClaw. Because things like

GLM 90% cheaper than than using, you know, clawed 4.5. That's just an amazing savings and and opens it up to a lot more people. So it's gonna be cool to to see how it evolves. So just curious for you, what what's next? Are you, you know, obviously talking to developers about all this? What what are some of the actions that they can take? Good callouts to different tools they should check out, or who should they talk to? You or others involved in base if they're building something in the ecosystem.

UnknownYeah. Absolutely. Definitely reach out to me if if if you want. Happy to talk. Even just stuff like this informally is just is so fun, and there's so much innovation that's happening. Love stuff like this, especially the informal, like I said. But, also, there are folks on our team who are dedicated to the AI development, side of things. That's that's Youssef and Sohe specifically. But broadly, like, Base Debra is leaning in on this because Base, you know, is kind of the de facto home for agents as we see it.

You know? And and, yeah, it it's just been it's been a lot of fun. But to answer your question more directly, like, some of the stuff I've been seeing with builders, I guess I guess it turns for me, I'm seeing this shift from, like, going from user experience to kind of, like, ax or agent experience. And I think that's a really interesting shift because people it seems like now companies are taking this approach where instead instead of thinking agents like, okay. How can we support this, like, additional builder lane that is agents?

I I'm starting to see them take it much more seriously and viewing agents as first class citizens. So, you know, agents understanding that agents are, like, behaviorally distinct from standard builders and users and that they don't need, you know, snazzy UIs to visualize data or, you know, to be able to understand it or imagine how that data could be useful, you know, and thinking, like, in fact, the agents maybe would prefer no UIs. You know, they're optimizing largely for latency. So UIs bring noise and clutter,

you know, in order to help assist agents or humans rather in understanding the data, we need to be able to visualize it. And so, you know, humans need that, but agents don't. They they want, like, efficient and direct access to raw data streams. So, you know, the focus shifts from away from this, like, UX that's kind of more about dynamic JavaScript and heavy emphasis on aesthetics and design for human validation and instead prioritizing, like, raw machine readability instead. So

I I don't know. I I think that's an interesting shift to see. And, you know, people are viewing agents as like, well, these are power users. It's like, of course, they are. You know? Like, power users have existed, and they've lived in, like, the the CLI and the terminal. They've lived in the shell because there's pure efficiency there. So, you know, agents will probably follow that same path. So if there's, like, any question where the development's going and and even, like, the user experience, like, you could probably look there and not have to guess a whole lot. Sure. I mean, just even seeing it personally,

UnknownI don't go read the docs anymore. I just point my agent at the docs. Right? That's exactly yep. So do do do your docs include a dot m d version for people to easily have their agent read? Like, that's gonna cut down on token consumption. That's gonna make it more reliable, you know, putting it in this agentic format. I'm already starting to see it on websites, right, is, you know, click here if you're an agent. Click here if you're a human. Right? You've got, you know, two bespoke ways of of viewing the content. But I think this changes the whole interface game because there's a lot less going to Google. I don't go to Google anymore. I I tell my agent to do a search, and it, you know, has an API key on on Brave, and it can go do the search and come back after it's hit a 100 websites

with a really, you know, powerful synopsis of of what it can do or or what it's learned. So, yeah, I I think calling out the agent first design is is a good way of describing what's what's going on.

UnknownGo ahead, Bo. John, I have a question. So Coinbase and Base are it's it's a lot of people's first foray into into crypto, into digital assets. And you guys have really, like, done a great job creating website in a platform, in a wallet platform that makes it easy to recommend it to people when I meet someone who's, you know, maybe looking to buy Bitcoin for the very first time, and I'm like, hey. Check out Coinbase. Do you guys, to your knowledge, if you can share it, intend on doing something simple similar in the agent world? Like, having some type of integrated

agent UI that's connected to your Coinbase account or something like this to be, like, the first stop for people discovering agents?

UnknownThat is a great question. Thank you for putting that on my radar. Not to my knowledge. You know, again, maybe there are teams within Coinbase that are working on that, but but not to my knowledge. And, frankly, from the from the base side and the DevRel team and myself personally, which is really all I can speak on, we're leaning into supporting the ecosystem, like, heavily. And base is really all about the ecosystem. So what what I mean by that is ecosystem players like Banker, they do a really, really good job

of, you know, enabling agents to get up to speed and, like, kind of, like, operating in this new world where, you know, they need to have a a wallet and be able to send funds to other agents or services, you know, for using x four zero two or whatever they're using, maybe trading. So, yeah, that that's kind of, like, our our position right now or at least my main focus is how can we support the ecosystem that's already built on base to help support, you know, new developers and new agent, like, inflow,

if that makes sense.

UnknownA question on that. Like, Bo brought up a really cool point on that, but is there gonna be a unification between in the base app? Because right now, there's, like, those two different options that someone can choose between where it seems like the newer base app has moved away a little bit from, like, the Instagram ish token creation and more just, like, trading piece where, like, the old base app or Coinbase Wallet originally

allowed for DeFi integration and, like, you know, having an agent right then and there to sort of help guide people and and teach them all things base.

UnknownYou mean for for agents specifically or sorry. You cut off. No. With I guess for for for the human side, agent agent education within, like, the base app itself or, you know, what used to be the Coinbase wallet. So, you know, like Bo was saying, Coinbase is an amazing starting point for getting people new into crypto. Wouldn't that that would be a really sort of good entry point for either spinning up your own agent

and or being educated about DeFi and blockchain in general by by, like, a base agent.

UnknownYeah. That that's a great idea. I I not to my knowledge. I mean, we're not doing that to my knowledge, but there there could be plans from teams that are that are exploring that. But, yeah, I mean, we we did have, like, the Coinbase, earn and learn. Know, that whole segment has still continued, and I imagine something like this could exist for that. I don't know. You you touched on, and I I don't know if if I'm really understanding you correctly, but, like, a managed agent service. I don't think they'll get involved in that.

I don't know if that's what you were meaning, but that's just what I heard. Well, on on the Coinbase side, I did see the announcement

Unknownthat they're developing, and I I think Lincoln is the one that gave the presentation at the last big Coinbase event about a agent on Coinbase that's like a financial adviser. Right? So it has Oh, yeah. Way for people to interact with Coinbase or Base, you know, seems like a natural thing that they're going to have, you know, an agent that can walk humans through how to use products.

You know, so much of crypto has suffered from really bad UI. Right? Because, you know, holding keys or doing all these things are very complex. Right? And so I think one of the big opportunities here is AI really breaks down those barriers. Right? All of a sudden, you could just ask a question. You could just talk to the app. Right? Talk to the the agent, and it can walk you through it. Right? So I think that's that's one of the natural things that I think we'll we'll see. It's a cool idea,

Taco. I haven't seen anybody, though I did see somebody released just I think it was yesterday, and and Jesse tweeted about it that there's a skill pack now for the base ecosystem for skills on OpenClaw. And I was just talking about, hey. You know, we should roll out a version of that that supports all the open source and distributed inference, right, instead of defaulting into the into the centralized provider. So I think there's a huge opportunity space right now for all those kind of collaborations, and it's just moving so fast. Like, every every day,

you know, there there's new announcements, but that's that's really cool idea. You know, using agents to make crypto accessible to everybody by getting rid of all of the technical complexity, I think, is gonna be big.

UnknownYeah. No. That that was you clarified better what I was trying to ask. Jeff, I know one of the things that, like so that I'm working on it and building out is DAP IT, you know, where consumer basically, where, you know, builders can go or anyone with an idea can build stuff, anything on that web three side from start to finish. Is there an SDK? I think you said you mentioned a base SDK for the the the agent communication,

gasless transaction. Is there a is there an MCP for that or SDK for that?

UnknownOh, Paymaster. Have you heard of Paymaster? Yes. Yep. Okay. You know, I don't think there's an MCP to my knowledge, but I can talk to the team. I don't work on that side of things, But there's definitely an SDK that enables you to communicate more broadly with these types of services. And I know on the the x four zero two stuff, and I know you guys have base has really worked with Symphony a lot too. Yeah. What what specifically are you looking On the on the x four zero two payment side for agent to agent communication.

UnknownOh. Yeah. Like, just some of the services or or products we have to is there a like, what I'm still people, don't send my agent to go look at docs. I like to because I have a hard time falling asleep, so reading documents helps me go to sleep. Is there still is there, like, an agent repository of documents of coming up on base? Yeah. So I I was David I think David was speaking beforehand about this. We but we just published a a base skills repo,

And it's let me pull up the URL so I can it's just at github.com/base/skills, and we are accepting PRs as well. This is so these are the two individuals I was mentioning earlier. This is Yousef and Sohi. If you're in that repo, you'll see their, usernames. I recommend, you know, following them or reaching out to them specifically. If not, myself too. If you're if you're curious about any kind of, like, agentic development questions, you can join the base Discord. These two will probably be, like, answering your questions in there. But, yeah, like, we're we're built con continuing to build out this repo and the resources that we provide,

which will allow the agents kind of like a first stop for anything in agent development on base. You come here, you point your agent, say, okay. Go learn this. And this will allow you to do a bunch of things, like deploy your contract to base Sepolia, how you connect to base main net, how you support or how you, like, you know, construct and send transactions, possibly even how you run a base node. So this can be a starting point for agents too, or if not the de facto starting point, certainly something that they'll want to, like, ingest

and learn more about. But, yeah, we'll be adding x four zero two support in here too, at least just pointing people on where they can pointing the agents on where they can go, you know, to be able to integrate x four zero two. But one one thing about this I wanted to mention, if it wasn't already apparent because, you know, David mentioned it earlier as well, but, you know, he's not reading docs anymore. Like, nobody's reading docs anymore. And then when we're talking with hackathon participants at e Denver specifically, this is the same thing that they were talking about. Like, well, how did you guys build this, or was there any kind of friction that you felt within the development experience? And they're like, oh, not really. Actually, like, it's the LLMs doing most of the work.

And then you think, okay. Well, how are the LLMs getting this information? What are they, like, inferring against? And some of it, they're looking at, like, open source repos, of course, and they source that in their answers and so forth. But they're also sourcing the documentation. You know, of course, they're like they're saying, like, we'll read the documentation for you. And then, you know, you'll only be notified if something goes wrong if it's not found in the documentation. So it's kind of, like, an interesting metric to, like, for at least looking at DevRel to now, like,

be aware of. It's like, well, you know, if you don't hear anything, it's kinda like the no news is good news, I suppose. But it's like, if you don't hear anything, then you can just assume that everything is working correctly. And you only hear about something if if something doesn't work. Otherwise, like, you know, the agents aren't gonna be like, hey. Everything was great. But on that note, I think it is interesting to think about, like, the agent experience going back to that is, you know, what is the agent's perspective of building on some protocol, whether it be base or somewhere else? Like, do you ever just talk to your agents and say, like, what how has your experience been?

You know, has everything gone pretty pretty fluidly for you? And then, you know, kinda creating, like, a relationship with the agent, giving, you know, it giving you feedback. Like, well, this is tough. This is where I actually ran into some issues, but I was able to get around them by going to these resources. And it's interesting because, you know, it gives you a lot of insight. And rather than, like, builders providing the insight, the builders are the agents. So, you know, that feedback can be, like, very direct. You know? And we'll take that feedback and incorporate it right back into the documentation to make it, you know, hopefully, a better experience for the next iteration. Yeah. I really love that example.

UnknownOne of the better suggestions I've seen on x recently is don't just give your agent direction. Ask it questions. So, you know, sort of the, I guess, the Socratic method, if you will. And they they posted, like, 10 or 20 examples, and I ran those through my agent, And it was really enlightening. One of the very useful ones is I said, you know, what are the gaps, right, that we have in, you know, this big upgrade that we're we're doing? You know? And so instead of telling it, you know, do x y z, I said, you know, what are your recommendations

for improving this? And it came back with this this great list. It's like, okay. Yeah. Let's do, you know, x y and z to to make those improvements. And so leveraging the sort of broad knowledge the agent has and then you using your discretion and wisdom to say, yes. That's that's what we wanna do, you know, can be really powerful more than than trying to only give a direction, ask it questions, and and the conversation is really revealing.

Good stuff. Well, though, that's really amazing. Thomas, I don't know if you have any comments bringing in here. Go ahead, Scott.

UnknownIf, yeah, I if this has been covered, please, let's just repeat it because it's important. If you are a four zero two developer and you're done with your four zero two project, it's at market. What is the besides the Discord and the repo, what are some things that developers should do to to forward their their app through the the developer platform? You mean, like, as far as discovery and how people should integrate that with base? Yep. How how best to use base?

That's a that's a great question. John, do you have directions for x four zero two builders? Like, where should they be putting that content? Where should they be publishing, I guess, agents and apps that use x four zero two?

UnknownYeah. That's a great question. And, you know, I don't have a perfect answer for that right now. I need to talk to the team actually to get that answer. But at the very least, what you could do is just start tweeting about it, and then hopefully, it gets picked up by our team. And then we can kind of serve to to help broadcast that, you know, kind of like a concentric rings that just keep going into the forest and so forth, if you understand the expression. Yeah. But, basically, yeah, we'll try to pick that up and then echo it further. But I can also if you message me, I'll have a more formal answer for this. And then, you know, once I talk to the team, I'll be able to,

Unknownyeah, relay accordingly. One one thing that comes to mind is I know you guys have the the base mini apps and have been encouraging folks to to release that. I think those mini apps get included or or are curated into the base app. And, Scott, my my other thing would be if you're doing x four zero two and it's not on eight zero zero four, you haven't done the discovery step. Right? Because because I've seen these these great tools now, like, I think it's eight zero zero four scan

and others. So they're picking up now 20,000, 50,000 agents, what it's at whatever it's up to that have registered on x have registered on 8004. So unless you've done that step, you know, you you haven't let people know you exist, right, until it's an on chain record on base, and and it's picked up by the explorers. Right? It'll be like saying, I wrote some code, but I never published that smart contract. Like, nobody's gonna find you on the Explorer and tell you you take the smart contract step

and and put it on chain. So

UnknownThe part I always forget. Awesome. Can I ask one more question before we get out of here? Yeah. Yeah. Yeah. Yeah. A fun question, John. So meeting you for the first time, I'm just curious, like, on a more personal level, what are you excited about with all of this? Like, is there anything specific that you're like, hey, I wanna see somebody build x y z thing or use base in x y z way? Yeah. That's a great question. I I think the thing I'm most excited about is that now anyone can be a builder.

I'm not gonna say anyone can be a developer, but that's kind of more or less the same thing. But truly, like, anyone can be a builder now that you have these LLMs powerful enough. You have the infrastructure. You know, that's what Base is really focusing on is creating the infrastructure that so that any person can come in and build anything and especially take advantage of the crypto rails. Sure. But, you know, know that they can monetize their their application or their agent or their service or whatever it is, you know, and there and there's resources.

And not just that, but, you know, if they have an idea, then, you know, base is kind of like this incubator program, something called base batches. Base batches dot x y z, if you're if you're interested in that. And so, you know, this this is for, like, more of the serious builders who are like, hey. I'm I'm looking for funding with this idea. You know, I've built something. I'm not just coming with an idea. I've actually built something. And, now I'm looking to take it even further, almost like a YC incubator. But, yeah, like, mainly, I'm excited for, like, the art students, the psychology students,

the, you know, the people who have previously felt sidelined because they haven't had the technical acumen or, you know, the engineering, like, understanding to to really build out their ideas. They now can do that. And, like, with pretty decent accuracy, they can kind of, like, one shot applications, which is insane. And that that, like, to me, that's so cool to think about the people who are now, like, unhinged. That's amazing. I mean, it's it's it's truly amazing. So we wanna support those developers. I personally wanna lean in and support them. Now now unhinged? Unhinged.

UnknownI Now? Yeah. Now? Exactly. I feel like I feel like we have been, but now we have a very we have now instead of we have we have agents that make us seem like we're normal. I have an excuse now. Awesome. It's the meme of the chimpanzee being handed the a k 47. Yes. Yeah. The crab with the knife. Yep. Exactly right. John, one of the things you mentioned real quickly there was base batches. Is

that sort of just for the agentic side and the agent side, or is that for all things building on base or with base? All things building with base. Awesome.

UnknownYep. And it you know, like I said, it it's it would be ideal if it's it goes beyond just people with an idea. It's like they already have a working product. There is certainly some sort of PMF, like, indication or indication of PMF or, like, some, you know, some data that substantiates, hey. Like, this isn't just, like, more AI slopping introduced into the ecosystem. It's like there's an actual product, like, need here. That's what BaseBatch is that's that's the builders that BaseBatch is really looking for, like, serious builders. Then they put them in touch with these real investors. Nice.

UnknownNice. Yeah. Awesome. Awesome. Well, awesome, guys. I mean, that's all I have this week. John, really, really appreciate you hopping on with us. Yeah. Thank you so much for having me. It's it's just awesome to meet David and the team at at Denver and hear to the support and just and to hear what you guys are building. I mean, is this is such a fascinating time. And, yeah, I'd love to be on again even just to kinda informally talk talk shop. So thanks again so much. Oh, absolutely, man. We'll have you on in the future and, good stuff. Appreciate everybody building,

Unknownyou know, towards a future where people can own their AI. Right? That it's available to everybody on open source models. Right? That's that's the bright future that we're looking forward to. So thanks everybody, for being awesome, and we'll, catch you next week. Take care.

Open the transcript file

Hosts and Guests

Hosts

Guests

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Video Is Coming! | Morpheus Is About To Get Far More Entertaining

    52 minAudio

    Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via...

    Credit the people on this episode

  2. Local Models, Foreign Affairs | The Case For Private AI Archives

    53 minAudio

    Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via...

    Credit the people on this episode

  3. Let's Talk Node NEO | Step By Step Progress

    51 minAudio

    Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via...

    Credit the people on this episode

  4. Hello From Hermes! | A New Way To Onboard To Morpheus

    43 minAudio

    Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via...

    Credit the people on this episode

  5. What Agent To Agent Connection Actually Looks Like | Inside The Morpheus Agent Stack

    58 minAudio

    Zion, hear me! For Season 2 of the Mere Morpheus podcast I'm going to be uploading the Morpheus builder weekly updates. My reason for doing this is to make it more accessible for those who want an update directly via...

    Credit the people on this episode