Mere Morpheus

Mere Morpheus artwork

Setting Up Your Agents Brain Demo | Cold Wallets & Compute Records

39 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 #93! Recorded: 1 October 2026 AI SUMMARY: In this week's town hall a snapshot of the Morpheus network showed a new all-time high in September for tokens processed natively through the compute marketplace and more than 57 billion tokens processed year to date. Protocol and product updates across the compute router and Morpheus UI bring better support for long chat responses stability fixes multi-wallet improvements and additional security safeguards. A major focus is better cold wallet support so users can keep Morpheus tokens in cold storage while authorising a separate hot wallet for node operations reducing risk while staking and participating in the network. A practical look at building better agents used David's open-source work with Bernardo as a real example giving agents a "brain" through long-term memory context management session analysis and SOP indexing so they work more accurately without repeating past work. Next steps span looping to recursive self-improvement alongside the industry trend toward open-weight models the growing importance of permissionless AI access and why open-source infrastructure matters more than ever as governments and large platforms place more restrictions around AI usage. https://www.youtube.com/watch?v=jNGY2NpvAKw Timeline: (00:00:00) Intro (00:00:30) Network growth update (00:02:18) Compute router updates and new contributors (00:05:00) Cold wallet staking and stronger security safeguards (00:09:24) Why agent building demos matter (00:10:54) Bernardo and the OpenClaw brain setup (00:17:44) SOPs, context management and long term memory (00:30:15) Open source AI economics and UK restrictions (00:35:18) Astana talks and Token 2049 meetup plans 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.

Bowen BelmerSo we are live, everybody. It is October 1. I cannot believe it is it is October 2026. That is insane. Time is flying by. But here we are. Lots of updates as always. Quick rundown, we're gonna run through just quick network stats like we always do, any protocol updates that are worth knowing right off the bat. And then we're gonna dive right into kind of a screen share and demo. People seem to like that when we've done it in the past, and we wanna do our best to take this town hall format and provide something for people to see here and and and, you know, experience

that they're not necessarily seeing on their timeline. There's a fine line between our community that is chronically online like we are and already tracking all of the news and then people who aren't doing that. And so we really want to, you know, give something to everybody in between that spectrum of people who are paying attention to AI. So with all that being the case, let's dive in really quick to network stats, and then we'll get into the meat of everything on the, side of the protocol. So really quick, we've crossed 57,000,000,000

tokens year to date, which is awesome. The big news though is with the closing of the month of September, we have hit a new all time high for tokens processed natively through the, the compute marketplace, which is incredible. Up and to the right, You know, this is exactly what we wanna see, month over month. And we're already off to a great start today in October, and we suspect that'll continue as people continue to interact with their own personal agents and also people building things on top of Morpheus. A reminder, as always, if you're someone who are building something on top of Morpheus,

we want to show it off. So tell us about it, and we will, you know, give you a platform here to show us what you're working on and and help help drive adoption to that. Not really a big change in terms of which models are most popular at this time. DeepSeek and GLM continue to be kind of the go tos for open source users. I know for me they have been. But ultimately, you know, great news. We're we're pushing forward here. There was a protocol update this week with regard to the compute router. Kyle, are you able to share any notes on that?

Kyle (Bow Tied Blue Fin)Sure. So, a couple updates there. Something very exciting for us is that we have some new contributors, who are making pull requests into the repo. Of course, we have Alan as the master of all, Morpheus, Lumeren, Node, and ProxyRouter, but we we have some other people kind of getting into the mix and sending pull requests, finding updates, and and implementing good code. So very exciting to see the

open source contribution route pick back up. So some of those updates, although I I will say Alan is is much more closely involved in this. We're we're relating to, like, the Morpheus UI. So that desktop app that we support for kind of local use of the of the Morpheus proxy router as a consumer. Some of the updates were on the, I would say, fifty fifty bug fixes, new feature edition. So looking more into the long chat answers right back in when this was originally developed, I saw someone either on Twitter or in one of our groups who's like, hey. Remember the context window from Llama three or from one of the older GPTs?

Like, we have expanded way beyond that in many multiple since. So, there were some old smaller session limits in there, but now long chat answers are are being supported fully. There's no truncation or or errors with with stream completion. So that was resolved. There were some additional fixes across both the desktop UI and the overall session on, internal calculations as you got close to limits

for session time. So on the shortest session time or on the longest session time, there were some calculations that just needed to be tightened a little bit to ensure that you weren't, like, overflowing in either direction, which could create some latent bugs in there, but that has all been resolved. Same goes with the s... With, like, the more estimate on those sessions. So just some general code cleanup to improve overall stability there. Back to the Morpheus UI,

there was some additions to the multi multi wallet use. So just additional work. Right? We're focused on how do people use the product in different ways. So, like, this multi wallet use is is important on the UI side. We're also concurrently working on cold wallet use. And, you know, me and David could probably get into that in a little bit more detail in a second. But one is, like, how do we make some improvements now to help the user? And then two is how do we strengthen the system and and give users that, like, gold standard solution that they'd like? So incremental steps towards the goal there. And then

there was one other security update that we made that just hardened the... Kind of... Not the private key management. That's the wrong way to phrase it, but to put in some more safeguards around what types of secrets are are known that you would never want to be sent through the router, how do we block some of those from being sent, right, as a a belt and suspenders approach. Right? We all have great or most of us have great operational security.

Some of us can improve, including myself, on it. But sometimes having an extra fallback somewhere in the system to say, hey. Like, are are you sure you wanted to, you know, send this API key? Are you sure you wanted to send x y z? We just put some more backups in there that that will help with that, in the case of a mistake. Right? It's not something you can say, hey. I can go do some flagrant violations of of my own, you know, personal information.

You know, not not to to stop that, but, hey. Just in case something happened, we're we're able to provide a little bit of backstop in in these scenarios. That, I believe, is the majority of it. Again, some other... I'm seeing in here there's some some course fixes in certain places and and so on, but I think the next round of updates that we make relating to the contracts are going to be much more impactful, for users. So, David, do do you want... I can explain it, or do you wanna explain it on the, the cold storage situation? Either or is good.

David JohnstonNo. I think you summarized it pretty well. You know, it's always great to see these cybersecurity upgrades. They continue to harden the system. You know, today, when people are running a node and they wanna stake, you know, there's an element of sort of using a hot wallet. Obviously, in cybersecurity, using a cold wallet is is a lot better. So, you know, getting towards, the ability for people to stake and not have to, you know, hold more in the wallet, I think is sort of like the gold standard that everybody

would love to, to get to. So I know Alan and the other folks that are contributing have been sort of marching towards, that goal. Still, some upgrades to go. But, yeah, like this is part of, how the software matures is just to keep making it more secure over time. But, yeah. Appreciate Alan and everybody working on that.

Kyle (Bow Tied Blue Fin)Yeah. Absolutely. The cold wallet stuff the cold wallet stuff, I think, is something that we've been talking about for a while. We've been trying to do a lot of rolling some... A a lot of different features into kind of the same contract updates here, but we decided that, you know, this is is really is really, really important for our users. So we're gonna kind of pull this up to the front and make sure that we get it out. And that's just allowing users... You know, we have the delegation in the contract now, but it maybe doesn't work as great as we love. So we're we're rolling out something a little bit more robust

so that you can, essentially hold your Morpheus token in a cold wallet and then give specific authorization to your throwaway hot wallet to be used in the Morpheus consumer node, protecting... You know, reducing your your... You know, in the risk event, you're reducing the amount of more that, you know, that would potentially be compromised if you expose your private key.

David JohnstonYep. Good stuff. But what Cool. Go ahead. I mean, I I think it'd be great to talk about, you know, the demos the last few weeks have been awesome. It's been very cool to see what's getting built. And one of the most requested things I've heard is people wanna see how folks are building their agent, how they're improving.

Bowen BelmerAnd so why don't we jump into that? Yeah. Absolutely. And and that's exactly right. I mean, people have been... And I know even just for myself, I'm curious, like how are people actually building what their agent looks like on the back end, the skills that it has, the things that it can do. A hot topic that's been brought up, I know you've talked about this, David, is the idea of recursive self improvement for agents, not just for models. And so what we thought would be a great idea is kind of showing and talking about some of the, like, steps to getting there.

Those being the agent brain, looping, and then RSI. So the ability to have self recursive improvements. Sorry. Geez, my brain. My brain could use self recursive improvements. Now the challenge with this is that it's hard to show code. So we were talking about this, you know, as a group, like, hey, how do we visually show this as a concept? Because... And actually, Kyle, you brought up a great point. It would be somewhat like saying,

look at this code. Okay. You know, it does it does... Hopefully, you can read it. Or, like, it'd be like showing somebody a piece of music and being like, hey. Look at this written down music. Isn't it great? You know? So somewhat of a challenge. But we thought a good place to start for this is to share the GitHub page for Bernardo, which for those who don't know, Bernardo is one of David's agents. Is this like kinda... This is kinda like alpha prime agent for you. Right?

David JohnstonBernardo is pretty prolific. I've been using him really since I got into Open Claw back in January, February, and he's just been doing a ton of coding. And, practically, I've open sourced it all. So, you know, everything from the Morpheus skill, where it makes it easy for somebody to use Morpheus with OpenClaw, like, here's the repo to do that. Right?

If you wanna look at install OpenClaw where it makes it sort of one click, I've gotten a lot of people asking recently, hey. I wanna make a white label of install OpenClaw and do it in Chinese and do it in Spanish and do it in Portuguese and do it in, you know, whatever their local language is. I think that's really cool. And so, you know, that's one of the things I've I've seen a lot of the last week is all these people popping out of the community saying, hey. I wanna do that. I wanna run one of those.

You know? And I was like, great. Here's the code. And so like you said, Bo, what a lot of this looks like is handing your agent a repo and just telling it, like, go crazy. Take this as as a starting point. So I... I'd really recommend for for folks if they're looking to... You know, the Morpheus skill, you know, has a lot of what you need if, you wanna run the, the Morpheus router, and it has a lot of the learnings that, you know, Bernard has had the last six months about how to do that,

effectively. But when it comes to, RSI, if we wanna kinda focus in on that, yeah, I like this idea of let's talk a little bit about the brain today. Next week, let's talk about looping. And then the week after, let's really get into RSI because these things build on each other. Right? If you haven't set up a brain for your your agent, it's gonna be really tough to do the looping. And until you have looping, you can't really set up an RSI agent. So, you know, having

gone through all those evolutions myself, I'm happy to, you know, have been the guinea pig and offer

Bowen Belmerwhat I've learned in that process. So Which, let's by the way, really quick, David, like in practical terms, what I'm going to do and what I recommend other people do is download this video when we're done and transcribe it, and then give the transcript to your agent as well as access to Bernardo's GitHub page, and start there. But anyway, yeah, go ahead.

David JohnstonYeah. So, you know, let's start today on the brain side. And I looked at a lot of different potential solutions. You know, I'm not a maxi about one particular solution, but the best one, at least for OpenClaw users, that I found was OC Brain, short for Open Claw Brain. And this is a skill, right, just have your agent effectively, you know, Google OC Brain

skill, and it's an MD file. And the approach that OC Brain takes is to create a database of all the sessions you've ever had with your agent. And so when I set this up, I had 6,000 sessions with my agent since the beginning of the year. And what it's doing is effectively trying to find patterns. It's trying to take the learnings that it has from all those sessions

and sorta build certain beliefs or, you know, things that it can extract and says, oh, when David, you know, gives the approval, it's only after it's passed its testing. Right? When when David, you know, wants to code, you know, he he always uses another model to do the audit. Right? And you just take all those learnings from the sessions, and, you know, and you can then have a really rich database

in order to apply that brain. The way that I'm applying all of that is now instead of writing all the prompts by scratch. When I write a prompt and and I set it up to, like, hey. I want you to do x y z. Look at my... So let's say my cron jobs that I'm running. I had my brain now go back and look at those prompts for the cron jobs and enhance them with the context

that's relevant to that job. So if I had 20 cron jobs doing all of these different things, all of a sudden they became a lot more accurate. They became a lot sharper. And honestly, what it did, it was just save a bunch of time. What your agent is often doing is kind of redoing all the work it did last time because it doesn't remember. It doesn't have all the context for how it got to the solution. So if you you take the time to set up

the brain skill, and you've digested sort of all of those patterns, now all of a sudden your agent is a lot smarter. Right? I mean, that's that's why people call this having a second brain or or giving your AI, you know, its own its own brain skill is it's taking everything you've learned before, and it's putting it in a context that your agent can really quickly access. Right? And so now, you know, that's become just part of my process.

Every... I think it's every day, I've set it up to effectively do a new scan of all the learnings from that day, add it to the database, and then the database, it has it as a tool to reference to sort of improve each one of these cron jobs or things that your your agent is doing. So my my prompts went from, like, a couple sentences to here's an enhanced pros... You know, you know, basically a one pager for the agent to be able to know everything

that's in context. So that... That's a good sort of how to set up your your OpenCLaw brain in a nutshell. The other thing I would say that's been really powerful me... For me is to take those same type of lessons, and you could build them into formalized standard operating procedures. I've I've got, like, 42 SOPs now. And I... And I've tried to open source a lot of these where I... I've sort of ripped out the private information, and I've just published a generic version. Some of these are in the repo.

And, like, okay. If I wanna do coding, here's how I could do that. Right? And always run this regression test, always, you know, do the cross model auditing, etcetera, etcetera. And now all of a sudden, your, you know, quality of your code goes from, you know, one agent that maybe has some gaps or errors to really, really solid code after it's been properly audited. So, you know, one of the things I would say is a lot of this this brain related

activity, it's basically context management. Because, yeah, we got 200,000 tokens, but man, if you try to shove all of your information in the context window, a lot of times it'll just fall over. Right? And and now I've got 42 SOPs. They don't all fit in the context window. And so I'd really recommend to people it's worth creating an index layer where you're indexing all of those 42, basically a small quick summary

of every SOP. And just now now I've got it down to this point where I tell my agent, hey, check your SOP index file, find out which context is relevant to what you're working on, and go ahead. And now it can just read one short file. It sees, oh, this is SOP 39. Great. It only reads the context that is relevant to its task. So, you know, that's kinda what I would recommend is, you know, download the skill,

create these patterns, start building out these SOPs, create an index file, and, you know, it's it's been amazing for sort of the accuracy and output of my agents went from, like, you know, having trouble most of the time to almost every time. Now they're able to get the task done because they have the information from the brain that they need to get it done. So, yeah, check it out. Give me feedback. I wanna open source more and more of my my SOPs,

but, you know, this has been this has been a really big unlock. So if you spend the time to do that and sort of... You know, it took it almost a full day for my my agent to crank through using the skill to digest those 6,000 sessions. If you don't have so much data, it might be a lot faster. But for at least me, you know, it's it's worth the time though. It's paid off in a huge way the last few months that I've... Since I've set that up. So Basically. Happy to answer any question. Bo, you might you might have a good perspective,

Bowen Belmeryou know, coming fresh at this. Yeah. I do. I... I'm I'm curious. So, like, let's say I I share with my agent this GitHub repo. I tell it that, hey. I want to set up a brain using this skill. And you said that it it took about a day because it's going through all of the past sessions and kind of plucking the the important pieces to add to that brain and and actually, like, build out that database. Is this like a do it once and it's done type of upgrade to your agent, or is this something that you are continually

monitoring and, like, remanually updating

David Johnstonand adjusting the skills? That makes sense? I don't know if that question makes sense. Well, the nice thing is once you've got it set up, I just set up a cron job to update it every day. And so once I had that digested, you're maybe only doing 10, most people, 10 or 20 sessions a day if your agent is cranking through different tasks. So it goes from, like, oh, I have a backlog of hours to do. Oh, it take a couple minutes. Right? And I don't have to do any of that manually. At this point, I just have a cron job, the agent looks at the sessions for the day, and just updates the brain automatically.

Right? And then I have the brain also automatically looking at all of the things in my agentic backlog. This this is, you know, sort of a little preview of what I'll I'll cover next week in the looping. But you wanna get away generally from babysitting your agent and giving the tasks one by one. Eventually, I just created a backlog. Hey. Here's a backlog of everything that you need to work on. And and once you have that backlog,

as it's sitting there waiting to to be picked up, my brain is enhancing the context of of that prompt. So when the worker shows up, it has everything it needs to execute. So it's a really important prerequisite for setting up all the looping.

Bowen BelmerWell, and I just wanna make sure, like, in in... Put in the most, like, basic terms too. Because I I already know, like, if I didn't already have you helping to explain to me how some of these things work, I wouldn't inherently understand the difference of just saying, okay. Well, I did my, you know, install openclaw dot x y z agent. Maybe I even took that next step of exporting it locally to my device, but I'm using Morpheus for the inference. Right? So now I have my agent that I'm talking to locally, but it's calling out not locally for the inference. But, hey. I have this MD file where I explain to it, this is what you are. This is what you do.

Can we even, like, maybe even more simply just highlight again why the brain even matters in the first place relative to the MD file as it exists today? I know you you basically covered that in great detail just now, but, like, that's the question that I know I would have if I didn't... Hadn't already talked to you about this.

David JohnstonSo think about it like short term memory versus long term memory. Right? In previous discussions we've talked about with OpenClaw, for example, and I think it's very similar with Hermes and other Aejenta carnasses, you can set up a soul file. Right? And your agent's gonna read that soul file before every prompt. Right? And it's basically your system prompt for your agent is to say, hey. The soul is your name is Bernardo. You're helping me with coding.

You know? You should always be frank about, you know, what you find and accurate in your outcomes, and here's your agent file. This is, you know, your list of skills and tools you have available. And, oh, here's the user file. Here's David, and he's working on these things, and he has these repos. But these are relatively small files, right, that you're putting in short term memory. Right? They're, you know, a a page each, basically, because you don't wanna overwhelm the context. If you literally try to shove a novel

into the system prompt, you gotta take up a lot of the context window, and now you can't really get worked down because you've, you know, put all of this unrelated context into what it's trying to do, it can get confused. So Yeah. You have you

Kyle (Bow Tied Blue Fin)you have the lost in the middle problem as well. Right? So if you... Right. The the more context you give, the agent is really good at remembering what happened in the beginning, what happened at the end, and what happened, I think, almost exactly in the middle. But in the in the the two con... The the two areas, you know, between beginning and middle and middle and end, it's kinda like, alright. I skimmed this, but I don't remember the specifics here, and and the specifics are important. So when you're creating that,

let's call it a prompt pack. Right? When you're creating that pack, you're saying, you know, exactly what David said. You nailed it. It's who am I? What is... You know, what are the critical things that I need to know no matter what? You know? What are my goals for this, which which are generated not maybe at a global basis, but as a... I'm working on this right now, so this is the goal of that. And then what's my context of this? Right? Maybe it's the short term context where if it's your second message in a chain, it includes the first message. Right? But then also what's my...

You know you know, let... Let's call chat history as an immediate context, short term memory as a, I've been working on this today, and then long term memory as, hey. These are global rules and learnings. And, like, how do I create that prompt pack to be both concise but but also comprehensive? Right? Concise in number of tokens, but comprehensive in being able to answer my question and and understanding enough what's going on. So I think that's... There there are a couple ways to do that, David. I'm I'm happy to... I'm interested in you explaining your implementation,

but it's... This is something that we have all gone through many, many different times on how do you get these damn things to work well. And, you know, that's one of the the key triggers on garbage in, garbage out, if you will.

David JohnstonYeah. Yeah. For me, the unlock was... At first, I put too much into the sole file and the user, and it just kinda got confused. And it was it was too ethereal. My instructions were very philosophical at first. What I ended up with at the end of the day was much more practical, shorter descriptions of what I wanted. So that's again the short term memory, right, that you're putting in the system prompt with these things that go with every time it sends a prompt into to the agent. The brain is really your long term memory. Mhmm. Right? So it's everything

you've ever talked about with your agent, but it's able to recall it from its short term memory and and access that long term memory only when it needs. Right? That's the key here. That's why I call it context management is the brain is basically giving you a tool for doing better context management at scale. Mhmm. Because these systems will work for a little while, and then as your context gets larger and larger, it has more and more trouble. So you're just basically giving yourself a structure

to infinitely scale this with as many files as you want, and you're only calling in what you need for that particular task or session. So Yeah. I've always heard that

Bowen Belmerthat exact problem is kind of what stops the, replacement of humans, basically, is just the fact that our our context windows and ability to pull from the past in data that we took in, you know, twenty years ago is just so great, and in a lot of ways, not even understood. The example I always give is, like, it's crazy how you could smell something. Like, you could smell, like, a certain food Mhmm. And then it gives you a memory from, like, two decades ago of, like, a very particular time, place, and experience,

and you have no idea, like, how your brain is necessarily storing that connection between the two.

Kyle (Bow Tied Blue Fin)Yeah. We upgraded to JEV. We we got JEV and VectorDB up there. Yeah.

Bowen BelmerIs Kyle's ready running JEV locally. Yeah. I guess, yeah, that's exactly right. No. That's a good explanation though. And I I always like going down to the... And I apologize, I have a cold. Going down to, like, the most layman terms that we can get to because I just know, like, personally, that's how I would experience it, you know, if I was listening to this. But, I think that's a good intro to to the brain as a concept, and we'll do it in small bites. Next next week, we'll talk about looping, and then we'll go into RSI,

recursive self improvement, not self recursive improvement. I think I said that wrong about three times when we first jumped on. But, yeah, we we will, you know, continue down this thread, and awesome. Thank you for kind of running through that, David and Kyle. I did have... I mean, we had a handful of things in the news. We we are gonna try to make the stream a little less news heavy than it used to be, but I still think it's worth, you know, touching on some of the things for the people that that kinda stick around. The biggest thing that I wanted to mention was just that,

obviously, all of the frontier models and open source creators are always trying to get to, like, the next level of intelligence, the next level of, you know, products and services that they're putting out. But it seems to be the case that competing on price has, like, really leapfrogged the narrative. Fortunately for everyone in our space, you can't really compete with the price of open source. I mean, it is still multiple orders of magnitude more cost efficient to be using open source models.

And in some cases, you have, like, OpenAI changing the pro price of ChatGPT, but actually giving you less allowance. So they both increase the and decrease the allowance, which is hilarious. And a lot of corporate America, or at least an increasing number of corporate America, has moved to open weights as well. There was data from Alphasense that showed that AT and T currently runs about 40% of its AI workloads using open weight models, and they're aiming for 70%

within the next year. And that's, I would suspect, one of many major companies who will be doing that. I don't know if you guys have a stance or any takes on on any of that.

David JohnstonIt's cool to see the move, you know. This is what Morpheus community has been predicting for a while. Right? That we're just gonna see ever more open weight model usage. It's what you see on a lot of the public statistics. You know? Last week, I think it crossed 50% generally in the industry are using open weight models. And, you know, god bless if this is part of what Jensen has been standing up for the open weight community because he wants to sell GPUs.

Right? And he doesn't care if people are running open weights if they're running proprietary models. He just wants to sell more GPUs. And so he's really been a powerful advocate the last few weeks to avoid some of the regulatory capture stuff and everything else. The one thing I saw in the news that was pretty crazy was, I guess The UK is now formally blocking Venice.

Bowen BelmerAnd Yes. I have that next on my list. Yes.

David JohnstonIf you're, you know, in The UK, now that your government wants KYC, right, to access AI, which which is pretty wild. And so they've just had to straight up block users unless they become a pro user and do the the, you know, basically, the equivalent of KYC, which is is prove you're you're an adult with a payment credential in order to access their... So it's really sort of limiting anybody that wants to use the free tier

from accessing it. But, you know, it's a good reminder of why Morpheus is important. Right? Because any... You know, Venice is wonderful, but as a company, like, you know, they... They've got a deal with those type of of regulations, right, as opposed to an open source protocol, right, where there's there's nobody to run KYC. Right? It's a it's a peer to peer network. So... You know? But it's a good reminder. Right? And,

hopefully, it brings awareness and, like, the level of Orwellian craziness that this is sort of exemplifying. Hopefully, you know, sort of wakes people up faster. But, anyway, I thought that was pretty wild.

Bowen BelmerYeah. I mean, if only there was a whole community of people across the globe who were building a way to access these things that are permissionless and perpetual and don't require, you know, any of these things that The UK wants to push. The UK's ability to just absolutely just shoot themselves... Cut themselves off at the kneecaps every time there's any type of technological advancement is quite quite incredible.

David JohnstonIt's it's truly amazing. It's truly amazing. And it's like, didn't you guys help give birth to the free speech thing? Like, what what happened to that UK? Like, you know, it's... I guess it's been a few hundred years, but... Yeah. They they officially gave up when the Internet came out, and now they're just, like, doubling down, tripling down, and it's

Bowen Belmerthe revolution really messed those guys up, man. People forget. Yeah.

David JohnstonYeah. It's, you know, the the freedom moved to a different jurisdiction, some... Somewhere in Texas or America or whatever. But, yeah, it's it's pretty wild. So, you know, we'll keep building open source and, you know, I think freedom of intelligence is a human right worth fighting for. And so here we are doing doing the good fight. It was it was a really amazing that day. I got to speak at the Astana hub and, you know, ton of

AI experts and and local folks interested in AI showed up. And, yeah, I'll post some of the photos and stuff, but two hours. We spent two hours answering every possible question they can come up with on agents. It was it was really a blast. So, yeah, kudos to the network school on their new campus. You know? If I were to summarize I've... Summarize it for people, it's kinda like Dubai North. Like, it's a colder version of Dubai,

but it's very similar, like, fast growing neutral country with a a, you know, real friendly framework for for AI and crypto and and related projects. It's a kind of place where, like, oh, yeah. You can set that business up, and it'll take fifteen minutes to incorporate. And you can, you know, do all these things that would be impossible in most most countries. So, anyway, it's cool to see. It's it's a beautiful city. It's basically brand new, and a lot of people moving here. So, yeah, it's been fun to explore. Transparently, Transparently, I've never even considered

Bowen Belmerwanting to go to Kazakhstan before, but I do now. So, you know, I I do... I would love to visit. I would love to see Network School and and Astana. So, yeah, it's great. I'm I'm very impressed at how quick they were able to just keep the ball rolling and and keep the mission alive. And it just shows that open... Like, as long as there is anywhere on Earth that wants to support open source and freedom of of, you know, freedom technologies and freedom of speech in the form of code,

it will exist, you know? And so it is the gift that keeps on giving. It is it is the most for the people thing of all time, which is so funny to me because I, you know, people hear AI all the time and instantly assume evil empire centralized control of your life, and they don't realize that there is an entire, you know, thousands and thousands of people across the world who are, like, very much building the thing for them specifically, you know, for the individual. So,

yeah, shout out shout out to them. Very cool that you're out there, David. I hope the talks and everything went well. It looked like it did, the one you gave the other day, so.

David JohnstonYeah. Last thing we're saying is whoever's going to token twenty forty nine is coming up, like, next week. So if you're gonna be out in Singapore, tons of Morpheus community there. They'll be at token twenty forty nine, be at the network state conference in Singapore. Should be a blast. So looking... Usually, the biggest conference in in crypto every year. So if you're coming out, come say hi.

Bowen BelmerAwesome. Yeah. Please do, people. If you come out, you see the Morpheus shirt, I can guarantee that logo will be on David's chest, almost certainly. I think he's got a wardrobe of Morpheus t shirts. You will find them. But, yeah, absolutely go connect with the Morpheus community out there if you're token 2,049. Otherwise, thank you as always for joining. Hopefully you like the even further revised new format. Feel free to give us feedback. I mean, we really wanna provide, you know, information and things that are truly useful to the community and the people following the project who want to be involved in building on the project or a user of the project. Yeah. So just thank you to everybody. Kyle, thank you as always for hopping on.

And with that, I think we will cut it and see everybody next week.

Kyle (Bow Tied Blue Fin)Thank you, guys.

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