Merry Models Christmas Again | AGI Hype & Non-Fungible Agents
44 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 #88! Recorded: 27 August 2026 AI SUMMARY: This episode opened with a Morpheus protocol update highlighting that native peer-to-peer throughput hit a new all-time high in August, alongside growth in provider count and model availability. While API gateway numbers required context due to inflated activity from bad actors, the team reaffirmed its focus on transparency and building a stronger decentralized AI network. The discussion then moved to a huge week in AI news: GLM 5.3 Flash challenging the economics of frontier-grade open models, the potential for more affordable hardware to run powerful models locally and the implications of Apple and Nvidia's latest hardware moves for decentralized infrastructure. The conversation also covered blurry AGI marketing, Nvidia's reported $12.9 billion acquisition of Hugging Face, Coinbase's push into "AIFi," licensing pressure from providers like Moonshot and the emerging category of non-fungible agents. The recurring theme was that open source AI, local compute, and decentralized infrastructure are all moving in the right direction with Morpheus well positioned in the middle of it. Timeline: (00:00:00) Intro: Morpheus weekly agenda and network overview (00:01:34) Throughput trends and native usage hits a record (00:04:24) Gateway health, provider growth and real demand (00:09:58) Provider features, transparency and building for users (00:13:21) GLM 5.3 Flash launches and challenges top models (00:16:43) Smaller hardware and local AI get more capable (00:22:53) AGI claims, hype and a joke about escaped agents (00:25:30) AI film tools, lunar mass drivers and model quirks (00:30:21) NVIDIA buys Hugging Face and open source gains (00:32:35) Coinbase AIFi, crypto rails and model licensing (00:38:36) Crypto market rally and the case for agent markets (00:42:05) Non fungible agents and building Bernardo in public 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 BelmerAs always, gonna start with a update of Morpheus from a protocol level. And, anybody who is building something on Morpheus who wants to speak about, you know, their past week, always welcome to do so at the beginning of our our meeting here. And then we will get into a handful of news topics that are related to AI and DAI more specifically. And then, you know, sometimes we mess around and open it up to other things just depending on how much time we have and what what the mood's like this week.
So to start though, I always like looking at tech.more.org. So this is once again the website where you can do a handful of things. Most specifically, you can see where your use of agentic AI currently falls on what we call the spectrum, which is really a spectrum of convenience and centralized use cases to private and sovereign AI on the other end of the spectrum and everything in between, you can calculate how your more stake, whether it would be to the API gateway or directly
to the compute routing node, would, what the calculation would look like in terms of what type of model you wanna run and what your daily compute allocation would look like per an amount of compute from both of those methods. We have information about trusted execution environments, which is really important if you care about data privacy. And lastly, but certainly not least, we have information about throughput for the network as a whole. So taking a look at throughput right now, year to date, we are just under 45,000,000,000.
Those are AI tokens, not crypto tokens. But 45,000,000,000 tokens approximately in total throughput across 330,000 closed sessions. And as we look at the monthly chart here, we are ahead of where we were in June in terms of total throughput. We are behind July, but we will speak about that momentarily. But the hidden key, like, sneaky up into the right factor of this throughput chart right now is if you look specifically at native throughput,
meaning throughput directly through the compute router, not through the API gateway. We have hit an all time high in terms of native native throughput. We are currently at nearly three and a half billion tokens of throughput natively in August, which is a new all time high. The previous all time high was 2,250,000,000 back in June. And so that element of the protocol and the network is up into the right, which is of course what we love to see. Morpheus inference, Mark, Blaise Tech Chat notes.
Growth, throughput, and, you know, more and more and more people leveraging the technology ultimately.
David JohnstonYeah. Bo, I was I was gonna say on that subject, it's really cool to see the growth of the peer to peer network. And, if my calculations are right, we're doing about 200,000,000 tokens a day, through there. If that keeps at pace, that'll put us just above 4,500,000,000 for the month, which would be a 100% growth over the last all time high, which is, which is really cool as sort of a lot of the activity has switched from using the gateway,
to using the peer to peer nodes. I I hope to see that trend continue, but, it's great to see another 100%, growth over the last all time high.
Bowen BelmerYeah. Totally agreed. And and I like to look at this in terms of, like, how people are experiencing Morpheus. And the API gateway was kind of the the first and the the first, like, really impactful light client or way for somebody to interact with Morpheus protocol that was highly effective, you know, did its job and relatively easy to use. So first of all, shout out to everyone who worked on that. But the more native growth that we see,
the more different ways people are experiencing using Morpheus, and like what it means to be built on Morpheus or or, you know, using Morpheus inference. And so, yeah, hopefully that continues to grow over time, and I believe that it will, along with, of course, the actual API gateway itself. With that, I do wanna turn over to Alan and Kyle to speak to any updates on the gateway side or the compute node side. And then also, did, you know, flag before that, hey, you know, from July,
we're it's our it was our first month of the year not up into the right specifically for API gateway throughput, but I know that there is good reason for that. So I'll turn it over to you guys to kinda talk talk through that as well.
Alan (Nomadic Rogues)Yeah. Sure thing. I can take take some of this. I think it's interesting, you know, as a product or as an environment matures, you gotta look at different metrics and certainly we all love the up and to the right on something that's very visible like that. I think I'd take a step back and say, we, you know, the Morpheus network was meant to be a peer to peer. So in order to bootstrap that, you do have to have some
ignition source. And really the API gateway was and I think continues to be that ignition source, you know, but as you mentioned, it's it's that it's that first introduction that's really easy to kind of swing your harness over and start to look at things and see how it behaves and performs and what the models are stable, that kind of stuff. If we say that Morpheus is gonna be, you know, bulletproof or you can't shut it down, the more peer to peer on both sides, both consumers and providers is incredible. And so one of the other numbers that I look at and I watch is,
you know, the number of providers. You know, we talked about number of models that you can see on active.more.org. You know, the models are 200 plus. The provider count is also going up which is great. So you have more people entering the market on that side and the more providers and the more consumer peer to peer nodes as you mentioned, the stronger the network gets. And I think that's really encouraging to see. So even though the token throughput, you know, it did take a it took an artificial bubble,
know, we've talked about this before. The API gateway had some had some issues and we had some bad actors come in and abuse the system a little bit and I think that unfortunately inflated the numbers and and made everybody excited. But now we're down to that real, this is the real throughput even if you if if you look at even the API gateway numbers, they're back to that that growth from June into into August. And so I think it's all all good and healthy. My perspective is we are as a as a community, as a protocol,
we are all good and healthy and collectively, even though you may not be able to see it as clearly, it's all up into the right.
Kyle (Bow Tied Blue Fin)So
Alan (Nomadic Rogues)I'll pause there for a second. Any any other questions on that? That's just my my color commentary. I think it's all good from from my perspective on on both of those things.
David JohnstonNo. I think that's that's awesome, Alan. I think I think having, you know, the number of providers and real tokens going through the the network, right, or I guess they were all real, but not subsidized by by the gateway is is a is a good accurate description of, you know, the the usage of the system. So that's really cool to see. I appreciate you putting out, I wanna say the visuals like active.more.org. I use that every day.
And I use tech.more.tech.more.org every day to explain Morpheus to people. So thank you for maintaining that. I think those kind of visuals to let people see what's happening inside the network are really, really great.
Alan (Nomadic Rogues)Awesome. You're welcome. Yeah. More more to come. My my favorite really is the active.more.org on the status. I mean, look at, you know, the the popular models. Deepsea v four Flash has 13 providers that are out there. So you want to talk about the strength. One of them, you know, they can't shoot them all, you know, they being whatever agencies may want to, you know, turn off a decentralized AI environment. There's 13 providers providing DeepSeek v4 Flash. That's amazing.
That is strength. So those blue lines, those those solid blue lines on active.more.org are excellent because that's our foundation gets stronger and stronger every day. So I'm glad you like them. More to more to come on, I think, active.more.org. I've another dashboard that we're working on.
Kyle (Bow Tied Blue Fin)Great explanation before, Alan. Super helpful. And and the comment you made on the system being healthy, I I couldn't agree more. I think that once we zoom out, the that big spike that we had last month Although you mentioned, yeah, it was some people coming in and manipulating the system, but the I think the positive thing to see there is that, like, the demand was real. Right? It it wasn't fake numbers coming in. It was people seeing an opportunity to go in and kinda come in and take some free inference. Right? So it it although it wasn't
using the system as designed, it was using the inference as designed. So we know the demand is there, and and we're we're, you know, finding ways that people are capturing that demand in in more efficient ways. And we see that transition over to directly in the in in the inference marketplace. Under comments on seeing more providers, yeah, that that is absolutely amazing. Right? We see more strength in the system. We see more providers being available. It gives a better experience for the consumers in terms of
fallbacks and so on. Now one interesting point that we're kind of figuring out and learning as the system goes on is that every provider might be a little bit different. Right? Some might offer some models and other models. Some might offer them in different ways with different context windows and different features that they that they support. So we're looking into how do we help expose that to everyone. Right? How do we show users when you're connecting to this provider, you can,
you know, use these features that they support. So you should see something coming out in the next couple weeks that we're hoping to expand on that. So you can use and you can tune everything about the model that that the provider has available. So we're really excited to expose more, give providers more flexibility and and to show the amazing stuff that they have and and give you the consumers that availability as well. So we're just really excited on the trajectory of the system, the usage of the system, and and
really the the new models coming out just makes the system more and more valuable in terms of people switching over to open source and being able to kind of run their production systems on it. So excited across the board and really great work by Alan and and all the engineers on creating a sustainable system like this and and making it stronger and stronger every day.
Bowen BelmerI just love that Morpheus as a community and everyone who works on building it is, like, from an optics perspective, going out of their way to remove spoofed figures and expose where they exist to then show what the actual use case looks like. Like, it's it's, it's one of the things that drew me into the community and and the project from the very beginning was, like, everyone's shared core value of, like, radical transparency, basically, and, like, building the hardest system possible through that, which is so rare in crypto adjacent stuff, obviously.
So Well, we're not we're not just scamming everyone like everyone else, you mean? Yeah. Well, that's what I mean. Like, and and yeah. Put simply, I mean, it's it's great, though. I mean, that's part of why, like, I I I just have such a positive outlook for the project beyond, like, beyond technically what it is just because of all of the people who are working on it. You know? It it just speaks to a great community. That's what I'm trying to say. And and it's a lot of people who were
Kyle (Bow Tied Blue Fin)you know, we're we're not just building something that we don't use. Right? We're we're dog putting it ourselves. We're using the system to you know, for the all the light clients that we're building on top of it, and and we're figuring out how do we make this better, you know, for ourselves also And knowing who that target customer is and being that target customer, we're just, you know, continuing to, to help ourselves and help everyone else while we're doing it.
Bowen BelmerYeah. Abs absolutely. It's I totally agree. It's one of the best parts. Cool. Well, I think I think that was everything from this week in terms of updates on the protocol So side of we can dive into news of which there was a lot over this last week. I think the best place to begin probably is g l m five three, and more specifically five three flash officially going live. And Yep. That
is just once again completely like kicking ass and taking names and costs, you know, $0 to use, essentially. They're actually paying you to use it, essentially. Not quite. But David, I know you've had some time to mess around with this. Oh, and it was this is kind of a tied in news story together is it was released or or exposed to the word that keeps coming to mind, it's not the correct word here. That zero x alpha
was actually GLM 5.3. So that was kind of a mystery solved there as well, which I know, David, you had been talking about on the timeline over the last handful of days. Yeah. No. I mean, it's it's interesting because
David Johnstonzero x alpha had become the most used model in the world, partly because they put it out there and teased it as a zero cost model. Right? And so for their testing, I guess it was an earlier version of GLM 5.3 flash. So if you used zero x, it's worth going back and trying it again because this is sort of like the final version that has even more training and tweaking and has better benchmarks. But, yeah, the benchmarks are pretty wild.
Basically, beats GLM 5.2, which was already a great workhorse at half the size. So this is 380 gigabytes. There are 383,000,000,000 parameters instead of I think the last one was, like, seven or 800,000,000,000 parameters. Right? So it's half the size, which means it's cheaper to run. Right? It takes less hardware to operate, but it it beats the benchmarks, every single one of them for Geom 5.2.
So you can basically say, I'm now getting GLM 5.2 or better level intelligence for a fraction of the price, which is pretty amazing. They're definitely going head to head with DeepSeek v four flash, which has been the hot model the last few weeks and had garnered a lot of the market share. So it's really cool to see this new entrance from GLM 5.3 flash coming out. And the open weights are due tomorrow for the full
GLM 5.3 pro. I don't know what they're gonna end up calling the full version, but those are due tomorrow. So they're gonna drop on Hugging Face in, like, twenty three hours or something like that. So we're gonna need another Christmas gift tomorrow. So, yeah, mystery solved. That's that's who it was. I consider z a I one of the top open frontier labs at this point. Like, they just keep releasing these incredible models. So, yeah, I I saw somebody added it to Morpheus very, quickly,
and it's online. So I'm gonna play with it some more, but the benchmarks I'm getting so far are great.
Kyle (Bow Tied Blue Fin)Absolutely. I let let me add on that on on some breaking news here. Breaking news, I may have been wrong. So I I have said many times that local AI or running AI on smaller machines, you know, we're never gonna be able to get to production, you know, grade or workflow grade on these smaller machines, on something that even can run locally. And
on some of these latest releases and some of the new tech that that's come out from, you know, on Sloth had some and and a couple of have come out with some. It is becoming possible. Right? It is becoming possible that smaller machines with these models can run frontier level intelligence. Right? You look at you look at the GLM flash. You look at the QUEN three eight twenty seven and and some of those other models, and it it's we're getting closer.
Right? We are you know, when you start to look at some of the new NVIDIA stuff that's that's being talked about, you know, not just the DGX Spark, but some of their more consumer grade stuff that's coming out, we may see people being able to be providers in the system with smaller hardware. You know, maybe not for an enterprise organization running on on smaller local hardware, but we're we're excuse me. We're really starting to see you'd be able to
crank out five, six current sessions at decent took per second volume. So it's just really exciting to see the trajectory performance of these models getting better and smaller. They're not necessarily making the trade off anymore on, yeah, let me get an okay product at a smaller level. It's let me get an amazing and and frontier level product at a smaller level.
David JohnstonYeah. That that ties in, Kyle, to the other big yeah. Yeah. That a 100%. And that ties into the Apple announcement around, their m five ultra and m six hardware. So they're going head to head, it looks like, with, Spark, from NVIDIA and coming in with, very similar or better numbers. I mean, it's still not cheap. Mac Studio will run you at 10 to 15 k, but it's conceivable that, you know, consumer grade hardware
is soon gonna be running GLM 5.3 flash, level models. So, yeah, it's great great to see that and really potentially opens up the number of people that can participate in providing tokens.
Kyle (Bow Tied Blue Fin)And two more missing other ones. Thinking about this Oh, great. Cool. Let give let give them more about that. Thinking about what does this mean for for us and for the market, I mean, it means token price down. Right? And it means that on that same eight by h 100 cluster, you're going to be able to get ten, twenty, 50 x more out of it. So all these data centers that are being built, all these, you know, massive compute clusters that are being built,
now there's so much more excess capacity that there's going to be. Like, demand goes up, but supply goes up inherently because it costs less to run. So now when you have this marketplace where people will have excess compute, you know, they have a place to give it. They have a place for for people to access it. So I think across the entire stack, it it helps users. It it makes AI more achievable for for people, and and it strengthens our system and the ability to, you know, go and offload that excess compute and be able to grab it at a super low price.
Bowen BelmerYeah. I was just gonna say that I'm I'm glad that you brought up the hardware piece because that was actually next on my list. It was gonna be that Apple, RAM RAM, MOG, Nvidia. We're we're getting some Gen Z lingo in here. But but, yeah, they they released, you know, like you said, David, the updated chips for the mini and also the Mac Ultra. I will say I was at this has caught the attention of a lot of people that I've seen, like, not necessarily
natively in the AI space. I was at a, like, a networking event yesterday for a completely different industry, and a handful of people were specifically talking about these hardware releases and, like, how cool it is to be able to host models locally. And so it's it's continuing to break through the zeitgeist of, you know, just general computer users, which is essentially everybody as to why, like, local hosting matters and
realizing that, like, oh, this is something that they can actually do, and it's ultimately just, you know, what is it? Like, maybe a hop, skip, and a leap away from someone to go from that to realizing why Morpheus specifically matters.
David JohnstonSo Oh. Yeah. Everything's trending in the right direction. Well and it's fun because if we can get in the next few months, right, to that reality where, you know, people start getting this hardware in their hands and they're running those models, that's gonna make the provider network so much more decentralized. Right? You might be able to see dozens or hundreds or eventually thousands of providers, right, if you've got frontier models in the hands of anybody with 15 k worth of Mac Studio hardware,
you know, out there offering it, you know, it becomes it becomes really interesting. One, it really pushes the decentralization of Morpheus a lot more radically out there, you know, just because you have a huge diversification then of of where it's being hosted and and who's hosting it. So, yeah, I'm I'm really excited for that. It's it's good to see it's headed in that direction.
Bowen BelmerCompletely agreed. That brings me to our next news story from yesterday, which which was Sam Altman claiming, that OpenAI will achieve AGI in 2026, and then within just a couple hours, Jensen making a statement allegedly claiming that they have already achieved AGI. So my question is, what what what is AGI at this point, and how exactly are we, defining defining this where we're at a point now where people are actually claiming to have already achieved it?
David JohnstonIt kinda feels like it's turned into into marketing. I see Taco's got his hand up. Yeah. Taco. Feel free to opine, brother.
Player 1 TacoOh, no. I I just when you started to talk about he who shall not be named, I I just I how many times have you have he said that they've hit Agencik already or close to and
Bowen BelmerI don't know. It's just it's starting to not mean anything anymore, like, at all.
Player 1 TacoLike, are is is he saying that? Are they about to do another raise? Do you know? Or are they about to announce that another model accidentally went and attacked some other platform,
Bowen Belmeryou know, and collected all the user data. It's like the Michael Burry effect. You know how Michael Burry's called three out of the last 30 recessions or something like that? Like, he's gotten it correct because he basically calls for one every ten minutes. That's what this feels like with AGI. It feels like every single day I wake up and someone from a large frontier lab like this makes this claim, and then also the new tested model escaped its environment. It's like every day.
Kyle (Bow Tied Blue Fin)Yeah. They also have pretty shitty environments, it seems.
Player 1 TacoYeah. Well, it's never their thought file. Well, it's a third party vendor's issues.
Bowen BelmerThat's right. Maybe that's what Morpheus needs, you guys, is we need somebody to purposely garner an escaped agent of some kind and get on the news for it. It broke out and started making money for everybody. Oh, no. Yeah. Oh, no. It gave you self sovereign money. Protect me, please.
Player 1 TacoThey staked their staked the diverted emissions.
David JohnstonIt was installing Ethereum wallets left and right. Nobody can stop it. You get a token. You a more token. You get a more token. We're all getting more tokens.
Bowen BelmerYeah. That's exactly what we need. Exactly. I love it. Also, unrelated, but David, I saw you made your film debut.
David JohnstonAh, thank you. Why? I I I I appreciate it. I I was merely the the director making editorial calls. There, there's credit to go to Scott as the producer, but, it's pretty cool what you can do on Morpheus now. People have added all the video models. And so, little thing, MiniMax h three magic, and all of a sudden, there was there was some cool sci fi coming out. I I basically created a a short
little clip of what you'd imagine a lunar mass driver would look like. So for all my space nerds, Elon's been talking up the last year or so, mass drivers. And, basically, this is the idea that the lunar gravity isn't that strong. And if you just put a big, you know, electric rail, and and start shoving things off it at a high enough speed, it will get to lunar orbit, and then you can send things back to Earth. Actually, Bezos has been, to his credit, talking about this for years. We hosted him here in Austin, Jeff Bezos, back in 2019,
I think it was, in one of the new world's conferences. And he's like, yeah. I'm gonna put Amazon factories on the moon, and I'm gonna push it off with a big old, you know, mass driver, and, yeah, ship stuff, drop ship it to the Earth. And and people thought it was a wild concept at the time, but now that sort of Starship and things are really scaling up, you might actually be able to pull it off, and and I think they will. So, yeah, just a little cool visualization. So Elanid, in one of his recent updates,
mentioned one of my favorite sci fi novels, The Moon is a Harsh Mistress by Robert Heinlein, and it has a mass driver in the book. And so, yeah, I finally got to create a cool visual of what that what that might look like. So, yeah, I I look forward to directing more AI films in the future.
Player 1 TacoQuestion. Are you going to expand on that a little bit and submit it to this amazing film festival that Venice is hosting?
David JohnstonIt's a good question. I'd have to look more into the parameters of, of what qualifies, whether my short split
Player 1 TacoAnything made into AI.
David JohnstonOh, brilliant. I
Bowen Belmermight I might have to, might have to put it in for the submissions. I don't know what the cutoff is. But Yeah. There's, like, a different award categories, but I think the only real parameter is it has to be made exclusively with AI.
David JohnstonWell, this is exclusively made with AI for sure. There was, there was go to the moon? What the heck? Not not yet. You know, I haven't, achieved that level of, of directorship. You know, that that belongs to, Stanley Kubrick. Right? That's that's the that's the old joke is, NASA hired him to film the moon landings, but he's such a perfectionist. He's insisted that they shoot on-site. And so
Player 1 TacoOne of the funny things I noticed, I I think you used, what, Dreama or Cdance 2.5?
David JohnstonYep. I like the Chinese monologue that pops in. Yep. I I I saw that, and I was like, It wasn't intended, but it's a great Easter egg. And if you go and translate using Google translate the words, it's there is no such thing as a free lunch, which is one of the great quotes from the book and is a concept from economics. So but to just put it in Chinese, it it makes it all mysterious, and you're like, oh, what was what was the the meaning of that?
But it's a it's a great Easter egg.
Player 1 TacoIt is a very good Easter egg, and I too have had that happen while creating some cool funny stuff. And so I just I I I saw it, and I and then this is one of the really cool things. You start to learn the different models that you're using, and you see and you get to see, like, how, like, that they don't meet anything else on the market or where they're at. And you so when you're seeing other videos, you're like, that was made with AI. And then you're like, that was made with Cdance
Cdream two c 2.5. And, like, this is one of those really cool things, like, sort of, like, identifying car models in a way.
Bowen BelmerYes. Also, I forgot. Sorry. I don't mean to derail us from this conversation, but I wanna, hit all the timely stuff too. And this was also breaking news that I totally forgot about and somebody just reminded me, so thank you. This was arguably the biggest story with the exception of GLM 5.3, and it is the fact that Nvidia has agreed to acquire Hugging Face for $12,900,000,000. Holy crap.
Which is insane. And also, once again, like this just reinforces everything moving towards open source. Like, don't have the largest chip manufacturer in the world say, actually, we're just gonna deploy $13,000,000,000 to help support open source for no reason.
Player 1 TacoWow. Well, and that that lets NVIDIA well, one, collect a lot of user data, but also because NVIDIA has always been really cool on putting their models out. And, like, I get the weekly and monthly emails with all of those APIs for their models, which is pretty cool.
David JohnstonYeah. And and I think of all the major players that could have acquired something like a Hugging Face, honestly, I think NVIDIA is probably the best option. Right? They have the best, as as Taco was saying, sort of the best open source, open weights, friendly reputation. So, you know, you know, if if if, you know, a startup's gonna get acquired, you know, you, you wanted to get married off to, to the best partner. So, yeah, kudos to, the Hugging Face folks, on this awesome exit, and kudos to NVIDIA
for continuing to support the community.
Bowen BelmerYeah. It's just, directionally mega bullish for everything that we're, we're working on here. Definitely.
David Johnston100%. And on the community side, I wanna call out. I see, Morphine is here. I have to say he consistently puts out some of the top notch memes and posts in the in the ecosystem. It's just, like, tireless in showing up and commenting and putting out sort of great content. So keep it up, brother. Alright. Cool. What else we got, Bo? Is there is there more news? I mean, it feels like there's a a a thousand things that happened this week. Big models, big acquisitions, new Apple hardware.
Bowen BelmerYeah. That that was all the biggest stuff. The other thing that I saw, which was, actually also today was and let me just pull this up so I can get the acronym right, was so Coinbase has been on a tear with just updates and, like, various initiatives that they're working on. But this one today, they put out an article about they're calling it AIFi, the financial system for AI, and it dives into kind of
what this will look like in practice, and it really gets into a lot of the stuff that we talk about regularly, like x four zero two payments and just the intersection of agentic AI and crypto. Because as we know you guys, an AI agent does not wanna use your smelly paper garbage fiat money, and they certainly do not want to use your centralized payment processing provider who can turn the off switch on them off at any point that they so choose. And so
crypto is the native solution for AI agents. And, yeah, Coinbase just continues to lean into this. I do not love the name AIFi, but, you know, here we are. So we'll see. Is it am I gonna have to adopt this instead of d d AI?
David JohnstonIt might be easier to to say, and it might I like that it puts AI first because, you know, I think in the minds of a lot of people, decentralization, like, the the most people don't really appreciate its benefits, right, until they really need them. Right? And so I think it's okay. You know? Let's do this let's do the thing that will get lots of adoption.
Bowen BelmerYeah. That's fair. Also, I will say, you know, I'm a habitual DAI user, like like, using of that phrase, but there have been more more than a couple times where someone thinks I'm saying DEI, which is entirely different.
David JohnstonYeah. Yeah. It can be it be confusing for people. So Yeah. Yeah. Maybe AIFi is is Well, I I think it makes sense for Coinbase because especially in their context, they're a financial play. Right? And so I think it's fair to say what we're doing is more broad than AI five, but it's an aspect of what you could use, you know, an inference marketplace like Morpheus or smart agents or the tech that we're building for. Right? So language will continue to evolve. Right? There was Bitcoin, and then it was crypto, and then it was blockchain, and then it was web three,
you know, different audiences, different personas. It's all good.
Bowen BelmerYeah. Very fair. Good point. But no, that that was all the major stuff from this week. I'm happy to open it up if anyone has anything else they want to make sure that we cover. I don't know if it was covered earlier,
Player 1 Tacobut seeing Kent Moonshot letting all the cloud providers know that they are going to be requesting a 30% royalty for any of their, k product soup products.
Bowen BelmerNo. We did not talk about that. Tell me more.
Player 1 TacoIt's just that. I saw an article. Moonshot is gonna be, you know, demanding from, like, Azure, GCP, Microsoft, Google, everyone that, you know, if they're providing their k k suite k three suite stuff that they are gonna be expecting a 30% royalty. And I don't know if that's just like a a starting number, you know, and like, let's, you know, talk about it. But, you know, and and a company
licensing out their models, basically, is what it's starting to look like.
David JohnstonWell, it's it's worth pointing out that this was sort of expected, or at least I expected it. Because if you real if you read some of the licenses, not all of them are using MIT, which is which is pretty permissionless. A lot of them are using Apache, or they're using some custom license that they're calling open weights, but really has some provisos in there that say it is open and free for anybody with less than 500,000,000 users. And what what they're saying when they're where they're sending those kind of benchmarks
is they're basically letting any startup or researcher or consumer do whatever they want, but that threshold is high enough that it catches the big tech companies. Right? So what they want is the hyperscalers to pay them, right, if, you know, their their tech is used. Right? So they're they're happy for anybody else, but they see their competitors as the hyperscalers. Right? And if they put a 30% tax fee, whatever you wanna call it, license fee on top of their models,
they'll always be able to out cost compete those other hyperscalers on their own hardware. Right? So with that's part of their their business model, I think, that they're trying to preserve. So you could say that's cool or they're not that's not cool, but, you know, some people are gonna do it. I think it's a move that limits their market share, and so I think it'll be a move of the not top tier. So the the k series, I don't see as top tier as like a GLM
series. And GLM is full MIT, which is one of the reasons I really like it. So, yeah, that's the inside baseball for anybody that didn't read the software license, technicalities. Always. I'm sure. Bo, I'm sure you were already aware of all those all those details.
Bowen BelmerYeah. Yeah. Yeah. Absolutely. I was not. But no, thank you. I appreciate appreciate the overview. Yes. I I was I was certainly not. I'm trying to think, guys, if there's anything else this week that we gotta cover. I mean, other than no, I guess that's pretty much it. I don't I I tend to my fallback is always to lean into general crypto news, because, like, there's just the Venn diagram of our DAI community and, you know, the crypto community is basically a circle.
But the only thing I have to say on that is we're back. Solana at a $108. Oh my god. Alright. Ending the show. That's great.
David JohnstonEthereum Ethereum approaching all time highs or at least Ethereum over 2,500. Yeah? Yeah. Ethereum making me less sad than it was. What what's Bitcoin at? It's a higher than 80 right now. 80? Yeah. Just broke 80. 5? Good. Yeah. Hey. It's it's all good. We we like to see all of our crypto brethren succeeding. More liquidity is good for more open source and more Morpheus. So yeah. No. It's cool to see. You know, it came sooner than I would have expected. I would have expected it sometime early next year, but, you know, okay. Cool. First
first price movement, we'll we'll see where it goes. But, you know, it's all good.
Bowen BelmerTucker, you know the best part about Solana getting to 63,000,000,000 market cap is that it just means that Morpheus can get at least a 63,000,000,000 market cap because it does something useful, which is which is nice.
David JohnstonOh. I think I get some fainting Did you not fainting phrase.
Player 1 TacoSir. Sir. Sir. This is a Wendy's. No. So one of the one of the really cool things that was announced today actually on Solana is the fact that Korea picked Solana to do an there was a four party MOU to put their what's the word I'm thinking of here? Their their ETFs on. So
Korea is putting their market on Solana.
Bowen BelmerI'm sorry. Sorry. Not gonna have a market six weeks from now. Have you seen their
David Johnstonsorry, Taco. That was awesome. Alright. Good good times. We we love we love all the blockchains. Keep building. I just Except except for
Kyle (Bow Tied Blue Fin)Tron. Yeah. Except for except for Declare.
Bowen BelmerThere's a couple. But I will say, you know, I'll I'll give Solana their dues that it is the best place if you want to virtually buy graded Pokemon cards. It is the place. Hey, you know what's really crazy? Actually, I do have one other thing I wanna talk about before we get out here that is related to everything that we have going on here, at least tangentially. So there was an NFT drop that happened yesterday that sold out, like, very quickly for over 3,000,003 million USD total for the for the lot.
One, are are NFTs back? And two, non fungible agents. This is still a piece of the market that hasn't really taken a form yet, but I really think it will. I think there are absolutely gonna be non fungible agents. This was an I you know, an idea and a concept that, I don't know, a Morpheus community and certain members probably yourself, David, were like some of the pioneers of even minting the concept. And there will, in my opinion, at some point here, be like a
open sea of non fungible agents and probably some type of, you know, bull run specifically for that as like a market sub second segment.
David JohnstonI appreciate the kind words. I think the credit goes to Scott for the nonfungible agent, the NFA, as a as a term. And I'm glad glad to see that it's, it's finally catching on. So, you know, it only takes people a few years to to grok the crazy frontier ideas that we we have in 2324. So, no, it's it's it's great. I hope to see more of that. And, you know, honestly, with my own agent, one of the things I wanna do is more building in public. I'm gonna spin up a building Bernardo website.
And, basically, I I wanna post to there, like, what I'm going through, setting up my agent, connecting it to Morpheus, setting up its brain, setting up looping processes, building SOPs or skills for Bernardo. Like, I think there's there's a lot to share with people who are practically all going through the same journey. So, yeah, something, something I'm putting on the docket.
Bowen BelmerCool. Yeah. I will keep an eye out for that. That sounds awesome. For real this time, that was all I had this week, guys, in terms of, in terms of news.
David JohnstonAwesome. Good stuff, man. Great. Well, thanks, everybody. Appreciate you jumping on, Taco. Alan, for everything. Kyle, for jumping on. Beau, as always, and, Morphine and everybody else. Thanks for participating. It's been one hell of a week. Wow. Hopefully, we keep seeing that nice up into the right usage on the peer to peer network. That's what I'm here for. So talk to you guys soon.
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...
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...
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...
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...
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...