Mere Morpheus

Mere Morpheus artwork

Where Is Fable On The Morpheus Spectrum? | Rate Your Agents Sovereignty

42 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 #77! Recorded: 11 June 2026 AI SUMMARY: This week covered Fable 5's rollout on the Morpheus marketplace where an anonymous provider boosts privacy and access. The new tech.mor.org UX helps builders visualize their agent's position, calculate costs and explore models. The discussion emphasized hybrid setups using small local models with Morpheus-routed frontier models for heavy tasks, highlighting the network's resilience as a business advantage over closed ecosystems with silent restrictions. Fast-growing throughput was noted with DeepSeek V4 Flash leading usage. Updates included Hermes's iMessage integration lowering friction, a web-first agent experience calling for testers and Coinbase's new MCP client. The conversation looked ahead to an agent-first web, suggesting builders visit tech.mor.org to gauge their agent's sovereignty and explore fitting paths Timeline: (00:00:00) Intro (00:02:26) Introducing tech.mor.org: mapping your agent on the spectrum (00:05:16) Alan on designing the spectrum UX and task‑by‑task privacy fit (00:10:05) Network growth: 15B+ tokens and shifting top models (00:13:38) Fable 5 in practice: $20 prompts that squash stubborn bugs (00:16:41) Ecosystem lift: frontier gains flowing to open source and local (00:19:20) Censorship, downgrades and the case for permissionless AI (00:24:55) IP claims and contributors: who owns your AI‑assisted work? (00:28:10) Positioning Morpheus to benefit from every breakthrough (00:30:06) Harnesses: OpenClaw, Hermes and iMessage lowers friction (00:34:33) Built on Morpheus: Dapit demo and one‑shot game creation (00:37:26) Early‑mover advantage: open doors for agent traffic (00:40:57) Wrap‑up: steady releases & self‑sovereign access 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.

David JohnstonWell, it's been, quite the active week indeed. I presume everybody's been having a lot of fun with, Fable and all the other releases going out.

Bowen BelmerYeah. It's been interesting to experiment with. I've been mostly looking through the pushback from you know, it's funny. The every time there's a major release now, there seems to be, like, the light bulb moment for more and more people where it's like, oh, wait. I don't control this? It's like, yes. Or or I guess, no. No, you do not. But this one I think was particularly big because of the lead up to it. You know, we talked about this a little bit yesterday,

but when when there's an update that's like, oh, it's, you know, the going from 4.7 to 4.8 on, you know, Opus or any other type of model, I don't think it sticks as much with the everyday person as much as being like, we're upgrading to Mythos and Fable five, and that just garners so much more attention and, like, hype.

David JohnstonOh, yeah. Like, this definitely hit the mainstream and those sort of naming conventions. I mean, obviously, they're brilliant in marketing. And so, yeah, it's been it's been cool to to play with, and I'm very happy to see that an anonymous provider has entered the chat on the Morpheus inference marketplace and started providing Fable five and, Claude 4.8 and all these other models,

because I think that's a great way to sort of get people off zero that are stuck in centralized AI land and at least get them anonymized use. Right? Get them using the Morpheus API for what they're doing, but still in a context that's already familiar to them. And I think that's a great way to sort of red pill them and get them moving towards open source models. So that's really cool to see. Call out to whoever did the the hard work there, but that's that's really awesome because it sort of brings all those capabilities

to the Morpheus community, but in a way that adds a bit more privacy than being doxxed and using the service directly.

Bowen BelmerYeah. Exactly right. And that kinda leads, David, pretty well into the big, piece of, I guess, UX or, like, thing put out over this past week, which was the update of tech.more.org. I'm actually I'm gonna have it pulled up right now. So if you are listening to this and either haven't visited tech.more.org before or haven't, well, since yesterday, definitely give it a look because you can now see

where your agent is on the spectrum.

David JohnstonHey, David. Do wanna talk about what what that means exactly? Yeah. No. I thought that I thought that was hilarious. I thought that was awesome because it's it's sort of like an easy to understand visual of where you are or where your agent is on the spectrum. Are you just using centralized AI? And, like, what are the downsides of that? Right? And so you can kinda click through almost like a journey from left to right, from centralized to decentralized, from, you know, the Panopticon of of Big AI to, like, totally private

through TEE or or local. And for a lot of people, there's probably a set of trade offs somewhere in the middle that makes sense for them. Right? Because I recognize that not everybody has $20 to drop on a powerful local rig. Right? Like, just that's not gonna be reality for most people. Okay. Well, they can still get cheaper access to models. They can still get frontier models. They can still get open source models and powerful access to intelligence

without having to buy that hardware by using Morpheus and without having to pay through the nose to Anthropic. Right? So I think there's this sweet spot where actually probably most people are where they don't wanna pay 2,000 a month to Anthropic, but they don't wanna drop 20 k on hardware, great. Here's an API that gives you those same models, that same intelligence at a fraction of the cost. And I think for a lot of people, that's gonna be really compelling.

Bowen BelmerYeah. Completely agreed. I mean, I discovered through this that even for myself, my own agent is only at a three on the spectrum right now. But I also think it highlights too, you know, we've talked about a kind of a hybrid reality too of what people's agents might look like. So for example, you might be able to, and like if you're running a dedicated Mac Mini or something like this, run some type of very small model locally,

but then call out through Morpheus to a much more robust model for, you know, the heavier workloads that you're working on. And you can kind of have the best of both worlds in that regard.

David JohnstonYeah. And and credit to to Alan, for putting all this up. I'm really loving the way tech.more.org is, evolving. So, Alan, I don't know if you wanna talk about sort of how it all came together and and what you've put up there, but I think it's it's just brilliant as a way to explain Morpheus to people.

Alan (Nomadic Rogues)I appreciate it. Yeah. It was I'm a pictures guy and I I liked what popped into my head after some conversations over the last few weeks was there's not one ring to rule them all and I'm gonna really mix references. Right? So from Spectrum to Lord of the Rings. And and kind of Dubose comment, I think that you could imagine instead of saying where is your agent? Where's your agent's task? Right? Because exactly what you said is like some things I don't care. Some things I want cheap and good enough and that's, you know, let's say you're running your own consumer node from

your OpenClaw and you're kind of a number four. When I want heavy lifting coding, maybe I wanna use a frontier model. When I'm doing something that's just good q and a, maybe I want to go open source because that token cost is quite low, but then I'm gonna send it, hey, I want you to check out this weird, you know, mole I found on my shoulder. You might want to use a t model. And so, of four, five and six kind of enables you whether it's through OpenClaw or or other front ends where you've actually got your node running.

You've got a lot, you fit the model of privacy to the task And so I think you you know, as I look more at, you know, kind of open clock, you know, like, hey, I've got these three things or four things available and you can mix and match where it makes sense to you. And that's really how this one came together was what are those trade offs? When would I not wanna use that? Because, you know, if you look at where most of the folks are, they're at step one. It's really they you know, the the big guys make it really easy.

Here's our our new really hot models. Just give us your email address which could which you could give a fake one, but your credit card which you usually can't fake. Most people can't fake their credit card. And, you know, all all all our your data are ours unless you opt out, you know, unless you tell us explicitly you don't want us to use your, you know, you as training data. It's really easy, but there are some downsides as you talk about privacy. And and on the other end of the spectrum as you know, it's all mine, you know, I've bought the hardware, I've got a network, I can do all that fun stuff.

What are the what are the variations in between? And that's how this really came to be was to be able to have a quick primer that says, what are my trade offs from convenience all the way over to the left? Highly convenient, but let's not mistake it, you are part of you may be the data, part of you may be subpoenaable, part of you is is the product and on the other side is, hey, it's really hard to create your own stuff and build your own rig but if you really want to guarantee that your privacy,

you know, that's what you're gonna be. So what are the variations in between and that's what this this game is. So it allows people to kind of explore and then, you know, hey, think I'm a number four. I think I could run my own consumer node, you know, with a frontier model. How do I get started? And at the bottom of those also, and I I don't want to miss this, is I added some helpers. So we've got a lot of things in the ecosystem like if you if you're on number four, the consumer node plus frontier. One of the other things out there in the ecosystem that builders have built that make it easier. So things like node neo for

a Mac or soon to be in iOS app, the j suite which is a CLI two interface, Pistachio which I think we've had focused on here on on a couple weeks ago, you know, in Morpheus skill for your Open Claw. So, what are those things that make that experience a little easier kind of the reducing friction projects that are around that. So you start with kind of clicking on the c node and then you can say, hey, how do these other things help out? And it's really about how to get started if that flavor

or that, you know, color or number on the spectrum meets your needs. And so that was the hope here was that it helps bring people in and say, hey, it's not just all all or nothing. You don't have to know Linux. You don't have to know how to code. And, yes, we send you the GitHub, but there are some good documents there to get you started. So I hope it helps. That's basically what tech.morr.org is is kind of find out more about more, you know, compare every way to run it, weigh what it really costs,

watch the network and then see how deep how deep down the privacy rabbit hole you wanna go go through. I do have real quick before before I close, we're over 15,000,000,000 tokens served by the way and we've had banner days. So, looking at the throughput which is always fun. It's pretty cool.

David JohnstonSo Yeah. I've started checking the throughput tab pretty much every day. I think yesterday was the second most tokens ever routed through Morpheus in 2014.

Alan (Nomadic Rogues)350,000,000.

David JohnstonThis is great. It, it keeps on growing. I mean, it looks like this month is on track to at least double, what it was last month, which is an insane growth rate. So yeah. No. Big credit, Alan, for putting together, tech.mor.org. You know, the way it kinda flows now is you can send somebody there. They can see what Morpheus can do for them. Oh, it's it's cheaper. It's private, more secure, or I can get access to the all these different models from a single endpoint.

That's great. And then they sort of can go down the rabbit hole of, okay. Well, that's why I should care. Like, that's why I wanna use it. Now how much is it gonna cost me? And I can pop over to the calculator tab. Like, oh, okay. If I spend x amount, I get x amount of credits. Got it. Well, you know, how does this work? Oh, put me into the documentation. Put me into, you know, how how things actually route in the background, you know, and how much are people using this? Oh, wow. Here are the most popular models. And I and I have to say it's been interesting to watch the evolution

from GLM 5.1 to DeepSeek v four Pro. Last week was the top model. This week, the top model has actually shifted its DeepSeek v four Flash has taken the crown for the most used model, which is interesting to see. So yeah. No. I I think that's super useful. Thanks for putting that together, Alan, and and maintaining that. It's really developing into a cool tool to to onboard people into Morpheus.

Alan (Nomadic Rogues)Awesome.

Kyle (Bow Tied Blue Fin)And and just on the on the token growth side, I I think that number will begin to skyrocket soon. Right? With the with that provider coming in and adding the some of those closed source models, you know, right now, one of the biggest discussions going around are, you know, I've maxed out my my cloud subscription, you know, my 20 x max subscription in, like, three days. Like, what am I gonna do? So I I think, you know, being able to go other places to get access to the same model that you might be using elsewhere

for that consistency in your code development or or your workflows, I think that's going to drive a lot of interest and a lot of usage. And last night when we saw those models, you know, being updated and adding and being added, Of course, I have some nice scripts and notifications to know when things like this happen. We kinda went in immediately and and added some of that pricing to the API so that it would be natively supported. So, you know, as Alan was talking about, you know, go to tech.more.org

and and all the to see all the documentation of what you wanna use and how you wanna use it. And we, you know, we just want to make sure that that everything that's available in the marketplace is is easy to use and then also becomes available on the API.

David JohnstonYeah. No. That's great. Thanks for jumping in with that that quick support. But I I think that is a game changer. Yeah. Just to speak to my own experience experimenting with with Fable five, I think I clocked it in at, like, $20 a prompt. And so, you know, I was really crafting and articulating and specking out my prompts before I shoved them in because it's expensive. But but to give it credit, you know, it knocked out two bugs I knew about and found two more I didn't know about and got it all the way done. And so

maybe it was worth the the $20 a prompt. Right? And so, I mean, to it's amazing to think how valuable intelligence can be. Because I threw some problems at it that some of the smaller models who hadn't been able to figure out for, like, four or five days. And I was like, I just wanna finish these these bugs and, you know, boom. It it knocked them out. So, yeah, it's it's really cool to see.

Kyle (Bow Tied Blue Fin)Yeah. My my personal experience with it is that it's absolutely amazing. Like, it's really it is it feels much more finely tuned towards the Claude coated agentic coating. It seems that it it handles these loops way more cleanly. It's able to predict kind of predict the results that it's that it's going to receive and and plan around that. So, like, I'm gonna run a test. If this happens, I'm gonna do this. If that happens, I'm gonna do that as the next iteration. It just feels like a

way smoother, iterative coding model in my experience so far. Kind of one of the I forget the older model names, but there there was a really big jump between two of the prior Opus models. I think it was maybe Opus four and Opus five or Opus 4.5

David Johnstonor Yeah. I think 4.5 was that magic moment. What was that? December 25? And and really kinda solidified with 4.6 in in January, February. I was like, yeah. This actually gets over the threshold. Yeah. It kinda feels like another one of those jumps here six months later.

Kyle (Bow Tied Blue Fin)Yeah. For for sure. It it it just it really is it feels like a much cleaner experience. And, you know, my excitement around the around that runs to, you know, maybe all the people who go and spend a lot of money to retrain their models off those, and we'll probably be the open source beneficiaries of that very soon. So excited that, overall, when AI takes a a leap in intelligence and and capabilities,

it it benefits the entire AI, you know, ecosystem and industry. So, you know, good things flow down, it seems, and and I'm just really excited about that.

David JohnstonYeah. That's really awesome.

Bowen BelmerOne thing I wanted to touch on too is, you know, we were just talking, just a moment ago about, you know, tech.mor.org and the throughput. And you'd I I think it was, it was either David or Kyle said, like, hey. You know, we're likely to see that, you know, skyrocket here. I mean, I definitely think that's the case even just based on the various projects that I'm aware of who are building businesses using Morpheus inference. You know, obviously, like on a on a personal level for a lot of us, we have the

the vision of like the personal agent that executes specifically for you and has your own personal best interest in mind and which is very much I mean, it does exist and is gonna continue to and prosper. But also, we can't ignore the people who are also going to build literal businesses where maybe from the user experience of of their end user, they don't even realize that they're using decentralized inference through whatever it is that they're interfacing with, but they are getting the benefits of that. And, I mean, we're already aware of a a couple projects who are doing that, and I know that there's ones that I'm not even aware of. And so that, I think, as some of those really go to market, is where I think we're gonna see a massive spike as well just in,

token throughput and and just straight up demand for more token.

David JohnstonWell, that's that's a good point, and it's worth touching on for a minute is one of the benefits of Morpheus is it's much more resilient because it doesn't have a single point of failure. As as fun as it is to use, Fable five and and Claude, honestly, it falls over a lot. And if you look at their uptime, it's like a rainbow. Right? Like, offline, partially offline, online. It's not even 99% up. So if you're building a serious business, it's really tough to have something that's constantly breaking.

You know, engineers and developers will kinda put up with it and, like, you know, go take a coffee break and come back in an hour and hope it's fixed. But if you've got a critical process, you want multiple options. And if you've got Claude, but, you know, also the open source models and Grok and all these different options, all of sudden, one of them falls down, no big deal. You just reroute to another available model. Right? And I think that's part of what people are gonna need in order to industrialize,

you know, AI as a big part of, you know, critical back ends and not just, you know, something to play with.

Bowen BelmerYeah. Absolutely. Well and and, also, I think we should highlight too, there has been quite a bit of criticism, of the rollout of Fable five, not necessarily from a a tech perspective, although there has been some criticism there as well. But it's a lot of things that we talk about frequently as to why permissionless, open source, private intelligence is so important. On on the side of performance, I mean, there's been reporting of lack of transparency with things like silent downgrades and altered outputs,

when people are using it for, like, their own training and things like this. Not gonna focus too heavily on that just because I don't have, you know, specific data to support that, at least not in front of me. If someone else does, I mean, definitely feel free to speak to it. But some of the other criticisms, though, not so much on the actual performance side, are things like gatekeeping the use case. So there has been reports of, restrictions on scientific and technical work. So, basically, just certain topics that Fable five seemingly won't touch.

Things like biology, chemistry, other AI research, medical analysis, which again is just a reminder that while these tools are fantastic in in a lot of regards, and they are, you're also you know, you're you're kissing the ring. Right? Like, you're you're able to use it to the capacity that you're being allowed to use it, not as a general open tool where, hey, this thing exists. You know, I I can use it for what I see fit, not necessarily what,

a board of directors or CEO or whoever is, you know, ultimately making that decision gets to make. And and this is especially poignant and important as we look to the not so distant future where Anthropic specifically is likely to be a publicly traded company. And so now you don't just answer to the centralized providers in terms of like who those decision makers are. You answer to them as they answer to shareholder value,

which is like a completely different proposition. Have you guys seen any of these pushbacks? I mean, those were kind of the big ones is, you know, restricting on certain topics, certain tasks, and some of the, like, anticompetitive behavioral concerns that people have had.

David JohnstonYeah. I mean, it's already been in memes. Right? Like, the the guy talking to the woman at the office meme, right, where it's the attractive business guy, and and Claud's like, oh, you're so sweet, and then it's a scientist. And she's like, HR, please help. Yeah. Yeah. It's like, you know, clearly, you know, anybody interested in doing that technical science is really annoying to basically be censored. And I think what happened today is Claude kinda or or Anthropic basically sort of backpedaled a little bit. And they're like, yeah. You know, failing

silently and not telling you we're censoring the output probably wasn't the right approach. We're going to degrade the service and just shoot you over to quad 4.8 automatically if you ask something that triggers, you know, that kind of reaction. And the community is like, well, thanks. I mean, it's still annoying. We can't access the intelligence, but at least you're not failing silently, and we don't know why it's broken. Now and now we're just being sent over to 4.8. But, yeah, it's gonna be really interesting because what it what it signals is,

one, Dario put it out another CEO of Anthropic put out another message about, oh, there needs to be regulatory. You know, these these models are getting so powerful. We can't let everybody use them, which one is marketing tactic, but two is is him trying to go down regulatory capture avenue again like they did in in '23. Right? And so, you know, I think there there definitely needs to be alternatives in order to push back on that. Right? And Jeff

Bezos or or or Beth Jasos, as it were, was the the head of the EAC movement was tweeting basically all yesterday about that. And I was like, hey, man. If we need to hold another EAC news, Morpheus rave in San Francisco, we will pull another 2023 and and push back on this because that was sort of an epic turning point back then. And I posted a few photos from from that event, which was was pretty epic. And so yeah. No. Like, the censorship is is real, and it's gonna keep coming up.

I'm hopeful that Grok will be more open. Not only are they open sourcing models, and and Elon says they'll open source 4.2. Croc 4.2 is really powerful by the end of the year, but they're generally, like, we're trying to seek truth. Right? And truth maximizing is is an interesting framework. And so if they don't engage in that type of censorship, I think it'll put pressure on Anthropic and others who are trying to restrict use. The other approach is the OpenAI approach.

They've said, hey, we'll help you with your scientific invention, and oh by the way, we get 50% of the IP value. You discover a new drug, that was our compute, that's our IP. Which is wild to think like to access math, you're gonna need to pay this enormous tax to these companies. And, I think the only way to to check that is to have pressure from alternatives.

Bowen BelmerOh, yeah. Well, okay. First of all, when it comes to the downgrading thing, yeah, like, I I did see as well that they had made a statement, but it's still a trust me, bro. Trust me. We're gonna we're gonna stop doing that. You guys are right. We shouldn't have done that. Sorry. You know? So there's no, like, actual, like it's like it's like when a governing body does an internal investigation on itself, and it's like, we found no wrongdoing.

David JohnstonYeah. And it was fine. And and we've made no reforms, but we totally won't do that again, and we did nothing wrong. Yeah. And you should definitely trust us.

Bowen BelmerAlso, yeah, you bring up a good point, David. I mean, this is like, you know, during the .com era or like other early Internet eras, you'd hear these stories of people who are, you know, working on a startup at home, but using like a computer owned by a company, and then all of a sudden it's this huge IP dispute because they weren't using personal, you know, their own personal hardware. That's now gonna be the case. Like, you go start a business and you use, you know, Claude to to help you do it. Do they have a claim to that? I mean, obviously, they shouldn't, but are they going to? I don't know. Well, they literally insert themselves

David Johnstonas a contributor on GitHub if if their agent or their AI is involved in writing the code. So you can go onto tons of different GitHubs, and Claude is credited. Like, that's the path towards exactly what you're describing. Yeah. Which is just sad, but, unfortunately, the reality. Also, the the playbook that they're playing from a regulatory perspective, it's the exact same playbook as FTX.

Bowen BelmerI mean, let us not forget had had Sam Bankman Fried not, have happened to be stealing retail funds for his own personal gain and his hedge fund's personal gain, he was like on a warpath to quote unquote work with regulators to establish crypto framework. And oh, by the way, they're gonna structure it in a way that FTX happens to check every single box that needs to be checked to be a government approved major player in the exchange space. We're just seeing that exact playbook now with these centralized AI providers.

David JohnstonYeah. 100%. The yeah. He his basically, his goal was was to ban DeFi, and The US would have been his his only playground for f FTX. So it's good that Eric Voorhees and others called them out, right, on on one, the hypocrisy of of what he was doing, and then people sort of realized the Ponzi game that he was playing. Because fortunately, on chain makes things pretty transparent, and it was it's clear the funds weren't there when people started digging in. Like, he would have gotten away with that fraud in traditional finance. Oh, but that's that's that's get away with it on on My point, though, is on the AI side of things, it's the exact same playbook

Bowen Belmerlikely without the fraud piece. It just happened to be the case that there was also fraud going on, and so it ultimately, like, killed that that whole component of what they were trying to do. But presumably, that that doesn't exist on the AI side of things, or at least, like, it's just a totally different, you know, categorically different type of business. But the playbook of regulatory capture is totally the same.

David JohnstonSure. 100%. So yeah. No. It's it's crazy. It's it's interesting. Like, things are continuing to accelerate. And, I mean, it's kinda normalized that we get a giant new advancements in AI every few weeks, you know, and that the capable capabilities grow like this. But, you know, if we position Morpheus well, the open source project and the protocol and the community will benefit from every one of these breakthroughs.

Right? You know, we don't have to be, like you said, at odds, you know, necessarily with the advances. We're just offering an alternative way to access them. Right? And so we want everybody in the open source world to get a huge benefit every time there's an advance. And and that those advances tend to flow down pretty quickly. And and I wanna call out Google too. I mean, Gemma four was a big milestone, and now there's basically a frontier level model that you can run locally, right, at a fairly small size.

So, you know, keep in mind, today's fable five frontier magic is, you know, four months from now, the local model you'll run on your hardware. Yeah. Right? And not even not even have crazy hardware just like a normal laptop. Yeah. It's now the worst frontier model you're ever gonna see as we look into the future. Exactly. Very cool. Well, I know we got other stuff to cover. What's on the list, Bo?

Bowen BelmerYeah. I mean, those were the biggest couple things. Let me pull up. Sorry, guys. The the one of the other things that I thought was worth mentioning, tangentially related, but I know a lot of people who are using Morpheus are doing so through either Open Claw or Hermes as their, what's the word I'm looking for? Not holster. Harness. Harness. Thank you. Yeah. Harness. Yes.

Kinda like a holster. But I just saw that there was a there was a announcement between Hermes and PhotonHQ, which was gonna allow you to do an integration with iMessage. And I thought that was actually pretty big. Now obviously, from a privacy perspective, you know, you'd be now using an Apple product ultimately to converse with your agent. But similar to what we were saying before with the where does your agent fall on the spectrum, I mean, I think that there's,

again, like, different u and and maybe, actually, someone correct me if I'm wrong, but different use cases for different levels of needed privacy. I look at this as a very good thing because it's just further lowering the barriers of entry for people to start using personal agents. Like, it it doesn't seem like having to use Telegram or Signal or something like this is a huge barrier, but I can tell you firsthand, I I can think of, like, at least two people in my personal life who I've tried to get set up with agents running on Morpheus, and they simply did not wanna use another text app that they weren't already using regularly.

David JohnstonYeah. It's it's a real barrier. And that's that's where, like, you have to meet people where they are. IMessage actually isn't that bad on security. I mean, you've got some dependencies with Apple, but, you know, it's not Google trying to mine everything you do. Right? You know, they have a fairly good record of at least trying to to keep that private information private. Right? And so you gotta meet people where they are. There are billions of people on iMessage. Right? We shouldn't leave them out of the AI,

you know, open source revolution because, you know, of their chat app. Right? And so, you know, 100% agree. And and just to give credit where credit's due, Hermes, those guys are killing it. Oh, yeah. I I think they're doing three, four x the volumes now of OpenClaw over OpenRouter. Right? And so if that's any indication, they've captured a huge amount of mind share, market share, and so it's really cool. And and just as a reminder, the Morpheus skill works with Hermes. It works with OpenClaw.

All the Morpheus tech is, sort of agent or harness agnostic. Right? So use whatever you want. Connect it to Morpheus and you can superpower your agent, with all these models and all this tech. So, and on that front, I'm continuing to make really good progress with install open claw .xyz, and I've lined up a number of testers. I've I've got enough tokens to probably take 42 testers. I thought that was the fun magic number. I've

got about half that many people have reached out to me, I still got a few spots left. If you're interested in being a tester, shoot me a message. But the vision here is a web first experience for agents. You can sign in with Google, you can pay with a credit card, and boom, you push a button, your agent is launched and you can use it right on a website. Nothing to download, nothing to install, you never touch a terminal, you don't have to be technical in any way shape or form, and you can access it on the web. So it's been cool the last few weeks, this is a partnership with Manifest Network. They're providing the CPUs that are hosting

the agents and obviously Morpheus providers are hooking up the inference, right? And so those two are really powerful combination and hopefully make it super easy for anyone to have an agent because I've got a ton of friends like you're saying who want an agent who wanted an open claw, want a Hermes, but they can't get over that technical barrier or they don't have the right hardware or whatever the barrier is like I want to take those barriers out of the way, Right? And so last call for testers, I'm gonna probably be putting this into production next week because it's looking good. It runs, and I've gotten some successful

screenshots out of of enough testers. I'm feeling pretty good about it. But hit me up if you wanna test it out, and, and I'll get you set up.

Bowen BelmerPerfect. Yeah. Everyone definitely reach out. Little Easter egg there with the 42, David. I like it. Those who know will know. Exactly.

David JohnstonExactly.

Bowen BelmerI I did wanna I guess, I don't love self promoting, but while we're here, I do wanna promote that we did put out the next the latest episode of Built on Morpheus as well, speaking of Manifest, because Dapit, which was who we were featuring on that latest episode, has a couple guys involved in Dapit who are also involved in Manifest. And you should definitely check it out. It was a very cool demonstration that Achilles gave as to, dapit.io's capabilities.

Won't spoil too much, but I will say he one shotted a video game using Morpheus inference, using the Dapit platform, the creation of a video game. It was it was very cool. Did the whole thing live. Another piece of news that actually just broke while we were on live right now was Coinbase just officially announced their Coinbase for agents, which is, it looks like, a client in MCP. So that is interesting. I'm curious to see how

how we can integrate that into our ecosystem as well. I don't know if you guys saw that. It literally just happened while we were talking.

David JohnstonI I hadn't seen that yet. That's really cool. And Coinbase, to their credit, have been doing a ton of work in this area for a while. They released their whole builder's kit for for people building with agents, a bunch of wallet tools, and stuff like that. But I'm glad to hear they've sort of, you know, given it a brand and made a big announcements around you know, I think this whole idea of agent first is really important. I was at an event on Friday

here in Austin, had 300 agent builders in the room all talking about what they're building on on Claude and in the ecosystem. And one of the coolest presentations was by Josh Baer, the creator of Capital Factory, has, like, gone down the AI rabbit hole the last six months. Right? Long time investor and CEO, but sort of put his coder hat back on and built a project called agents I think it's agentsfirst.dev. And you can go to the website, just pop in your URL, and it'll give a rating to your website

for how agent friendly it is. Like, do you have an MCP server? Do you have, you know, a a .robots text file? Do you have, you know, markdown files available? Like, all the sort of check marks for like, is it easy for an agent to use your service through your website? And I think that's really important, you know. He pointed out that it was 2009. There's a very influential book that came out, basically, mobile first. And it was like, hey, guys. You you need to everybody's moving to mobile. You need to pivot all your websites and everything to to work on mobile. I think we're at a similar point right now with, you know, people rebuilding the web for for agents.

Bowen BelmerYeah. This just makes me think of last week we were talking about x four zero two, and there was kinda like this weird narrative going around that, oh, look. It had this spike in usage, and now no one uses it. And it's to me, it's just so goofy because to your point, David, like, everything is going to be built agent first. I I I do not see a reason to not build everything agent first going forward.

David JohnstonYeah. Because you're just gonna tell your agent what you want to do, and then it's gonna go and find tools. So the point that Josh made in his talk was, guys, it's like 2007 when the iPhone came out. There were no apps. Right? They released the App Store after that, and if you were one of the first apps, you got a ton of users because you were the only app in that category. Right? And so right now, there are not a lot of MCP servers. There are not a lot of public agents. There's not a lot of agent first stuff. And he said, hey. I I put out this website,

you know, one day ago, and, you know, I post I I created a bunch of agent first tools, and within one day, I had 3,000 agentic visits. Right? And and it's it's low hanging fruit. Like, a lot of people have in their Cloudflare settings that they block all agents. Like, that's how people used to treat agents was, oh, it's just a worthless bot. Because before those bots were just scraping information, they weren't actually making purchases, but now they're driving purchasing power.

And so you gotta change your perspective. You gotta dig into those Cloudflare settings and say, hey, am I am I blocking agents? Like, I don't mean to do that, but that's like the default for for a lot of infrastructure in the web.

Bowen BelmerYeah. Well, and that's just, you know, selling to something that's not human directly is a very, very interesting, just like societal shift. Yeah. Like, I I think of it as, like, you know, search engine optimization was so big over the last decade or so. Now agent optimization will inherently be a role that, like, every business has to have or in some capacity has to have that.

David JohnstonYeah. A 100%. It's just gonna be table stakes.

Bowen BelmerReally crazy to think about. Yeah. That was all I had this week in terms of scheduled topics to talk about. Is there anything specific that somebody wants to bring up or or, you know, have part of the discussion this week?

David JohnstonI'm good. Alan, Kyle, I see we've got some of our our regulars. My favorite meme and image creator Morphium

Alan (Nomadic Rogues)is is online. Good to see you, man. Hey, Alan. Hey. No. Nothing else for me. It's been a good been a good chat. Continuing to be interested in what the

Bowen Belmerwhat what Fable brings to brings to the world to see how that progresses. So Yeah. For some reason on my timeline, one of the biggest metas right now is people pretending to one shot video games with Fable. I don't really get why, but are you guys seeing that? For some reason, I keep seeing people post like Skyrim and being like, I one shotted Skyrim with Facebook. I don't don't know why, but I'm seeing it, like, constantly.

David JohnstonClicks and baits.

Bowen BelmerLike I guess. Yeah. Because they're talking about it. I guess that's why they're doing it. Yeah. A 100%. 100%. Goofy. That's it. That's great. Well, awesome, guys. Thank you so much. If nothing else, then thank you everybody for joining as always. You know, look forward to doing these every week. And, yeah, it's it is funny how quickly we've grown somewhat accustomed to these, like, massive releases. But as they continue to roll out, we continue to have awesome contributors to Morpheus that, you know, provide

ways to access these various tools and do so in a way that is considerably more self sovereign than anything else offered by centralized providers. And as a reminder, if you want a really good way of measuring how self sovereign your current agent or workflow is, go to tech.mora.org, see where your agent is on the spectrum. Because one thing I can guarantee, your your agent is on the spectrum. So official diagnosis here, your agent's on the spectrum, find out where and how to move it across that spectrum depending on what your goals ultimately are.

David JohnstonAwesome. Good stuff, Bo. Appreciate everybody. Keep being awesome, 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