Mere Morpheus

Mere Morpheus artwork

Why Builders Choose Morpheus | Reliability, Cost & Censorship Resistance

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 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 #59! Recorded: 22 January 2026 AI SUMMARY: This week's builders update features Kyle, David, Bo and active builders unpacking the Morpheus Inference API's evolution for developers seeking reliability cost efficiency privacy and censorship resistance. It discusses the shift from early gateway iterations to the production-ready API infrastructure improvements like RPC autoscaling and why teams choose Morpheus over centralised providers that can throttle censor or shut down apps. Nima creator of DissentBot and DissentWatch shares experiences building uncensorable AI tools on Morpheus implementing staking-aware features and model choices. The conversation explores builder communities cost realities and the case for owning intelligence via open-source permissionless infrastructure that can't be turned off. Timeline: (00:00:00) Intro (00:01:31) Rebrand: introducing the Morpheus Inference API (00:05:01) Guest setup: inviting Irida TV (economics junkie) to speak (00:10:16) Centralised risk vs permissionless access (00:14:58) Open vs closed models: censorship, privacy and access (00:18:10) Marketplace dynamics: third‑party models and anonymised access (00:22:08) Model demand: Venice, DeepSeek and uncensored fine‑tunes (00:25:45) Censorship levels across providers (00:28:39) Adoption, distribution and why this matters globally (00:31:22) Framing freedom tech beyond politics (00:35:23) Open source wins long‑term: Linux analogy (00:37:02) Finding the builder community: Discords, X and meetups (00:42:02) AI writing tells and the human touch (00:45:00) Cost matters: democratizing inference for builders (00:48:56) Managing inference budgets: local vs cloud (00:51:40) Closing: doubling down on Morpheus and open source 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 JohnstonFirst, wanna say I think, it's awesome to talk about more in-depth builders and tooling on Morpheus. You know, again, the big shift for this year is we wanna talk about how people can use Morpheus and access its value as opposed to you know, we're getting past the the early days of needing to build everything now now that it's built, now that it's live. We can really go out and access, its powerful tools. So, you know, big shout out to Kyle and amazing work he's done on the API gateway

to open that up to a lot of builders. Big shout out to, Bo for launching season two of the brainpower podcast. I think that will be a great way for, builders to see what's what's possible on top of Morpheus. And so, yeah, why don't we go ahead and, kick off and just start digging in? Think, Kyle, if you wanna sort of set the stage on the latest with the API gateway, I know there's been a bunch of announcements on that side, and I think we've got some builders here today that we can bring on stage and, really talk about how it works.

Kyle (Bow Tied Blue Fin)Absolutely. Yeah. Great. I recommend that everyone first watches that, not first, but watches the brainpower podcast that came back. Really great stuff, but but we'll talk a bit more about that when when he speaks for a second, but definitely recommend checking that one out. On the API, so we are kind of rebranding it a little bit just trying to get the right messaging here. It's it's now going to be called the Morpheus Inference API. Right? The open beta API, the API gateway, these were all kind of intermediary

ways that we're kind of figuring out how we're gonna get this out. This is now what you'll see all messaging referred to as the Morpheus inference API, and that's the API access into the decentralized Morpheus inference marketplace. So just like to kinda keep hitting home on on that messaging there. Now as far as updates, we have a couple. So yesterday, I know some people reported that they had an issue for

I believe it was, like, one or two hours on on the gateway. We everything has been resolved. It was we actually like to see these here and there and and get all the corner cases out just before billing is implemented, and I'll talk about billing in a second. The the blip that we had was actually related to one of our RPCs. So we scaled up the RPC, and we have the proper auto scaling in place now. As you get more demand and very large spikes in demand, you kinda see which pipes get the pressure. And, fortunately,

we, you know, keep making those pipes bigger and bigger. And I think that some of the builders here today can speak about the reliability and the improvements in the overall architecture and and system and and how that has improved so significantly since the original, you know, open beta came out in in April or or May when whenever they started testing that. So today, I think what we wanna speak about is,

you know, first, I would say that progression, of of the use of the API and and how reliable and and how simple it has become to use. I think second and more importantly, we'd like to talk about why. And why is is the Morpheus Inference API valuable? And and why do we have interest in using this over you know? There there are tons of options available in the marketplace. Right? People aren't forced to use Morpheus. People are choosing to use Morpheus, and and why are they making that choice?

Talk about some of the, you know, specific value problems that propositions that the API has to offer and maybe which ones kind of hit home the closest for both of those. We've heard from Tales and Conquest before. We're we're super happy to have you, and we'd like to kinda dive into maybe some of those specifics. And then we have another new builder that I'll let Beau kind of come in and introduce, and you guys can get started. And then I'll I'll jump in there

and help steer as as we go.

Bowen BelmerYeah. Absolutely. Thanks, Kyle. Can you guys hear me okay? Yeah. Okay. Perfect. Yeah. So joining us today, actually, Morpheus' main account, can you please promote IRETA TV to the stage?

Thomas BorrelFor sure. Yeah. Hold on just a second here. While while we do that, I will You should I really you should you should be able to see that right now. So got the invite there if you if you want to add. Oh, there we go. Looks looks like you're on there. So you wanna test it off? Kick it off.

Bowen BelmerK. I'm muted. Not he's he's not maybe he's on mute. It's okay. I will do a proper in speakers. Alright. I'm off you. Oh, there we go. Perfect. Awesome. So joining us today, I do not know if you want to be doxxed, so I will not dox your real name, but I appreciate you jumping on with us. This is economics junkie, aka iRitaTV, aka the founder and operator of DissentBot, and also a very active community member in the Discord and the social channels.

When I first mentioned that you were gonna be jumping on to the spaces today, some of the comments across developers and other people in the community were, that guy beats us to the punch when something breaks before everybody else every single time, which is honestly a huge benefit to the community and ecosystem. So thank you for that.

Irida TVYeah. Yeah. Happy to help. But No.

Bowen BelmerI mean, we'd love to have you on to talk about, you know, what is going on with with DescentBot and everything else that you're working on, and also just to hear about your experience so far using Morpheus, for inference and then the other ways that you might be using it. And also any pushbacks or or suggestions for the community. I mean, this is really meant to be an open forum of all of the people, for the most part at least, who are, you know, most active in making this project happen. So, yeah, we can take it whichever direction, you know, you'd like.

Irida TVYeah. Yeah. I can just start, you know, talking. Maybe my my full name is Nima. I mean, real name is Nima, so I'm okay with being doxxed, by the way. I built a website called descentwatch.com, and I launched it in around 2019. The idea of that site was basically to curate the content from all the censored and canceled and persecuted journalists in the Western world.

And recently, I just you know, I built a bot in front of in my website, auto poster account have been banned in pretty much all the major social social media, but still allowed in Telegram, NoSCR, and Gab. And it's not not sure if I'm getting anything leaving anything out. Discord also, auto posters working. But

so I think, you know, when I was when I put together this bot, Morpheus was the perfect choice for inference because there's no censorship possible in their decentralized network and all that. I heard about Morpheus first, like, I think in 2023, actually. We I also run a podcast that that's my handles. I read at TV. I have David on there, like, almost two years ago at this point, I think,

to talk about Morpheus. And yeah. I mean, you know, there was just a little hiccup in the API, I guess, yesterday or so. But other than that, most recently, the API has been very reliable, very useful, and it's you know, in the beginning, there was lots of hiccups. And I guess, you know, like you said, I was beating you to the punch because I was really into using it. You know, anytime it wouldn't work in, I would let you know. I hope that has,

you know,

Kyle (Bow Tied Blue Fin)contributed to hardening the API a little bit. It's it's it's been so important, and I think that really like, that's what and the reason why I kind of let in with the progression is is exactly kind of leading into your experience. Right? Like, we launched this thing first. We're like, hey. We just built this. It's going to break. Right? It's it's we're halfway there on productionalizing this. Let's put this out and let let some people use it and test it. And, you know, the people like you and and Tails and and Conquest here,

you beat it up. Right? And you're like, I broke it this way. I broke it that way. And and that was critical feedback to, like, really getting it to the level that it is now. And I think that, you know, that that's not a bad thing. That that's exactly what we're going for, and that's that's why we wanted to put it out there, you know, for free for for all that time to kind of make sure that we're giving you the value also while not just going on asking everyone, hey. Come help us be a test. We're like, we gotta we gotta work together on this. So I I think it was,

you know, really, really helpful from all the work that you did, and and we're, you know, we're we're proud of of where it has become. And I think that if we could talk a little bit more about that, maybe the decision or maybe the risk that you have in choosing an AI provider. You know, when you're talking about content that is questionable, not questionable from what people are saying, but questionable about people being canceled and censored on on other platforms, and you wanna to be able to provide this this uncensored

and unfiltered access. Just kind of what happens if if you use one of the alternatives, like, you know, GPT or whatever else, and just talk about that, like, decision making and and vision and value.

Irida TVLast part cut out because it was just dropped and put back in with the last sentence. The,

Kyle (Bow Tied Blue Fin)you know, the alternatives and the choice of more Yeah. Yeah.

Irida TVWell, from my website, such a striking use case, I think. I mean, I don't wanna it's first of all, I literally can't get useful responses to the types of questions that people ask in my platform if we were to use ChatGPT or some kind of mainstream platform like that because when you ask what controversial topics, let's say you ask for vaccines or elections, it just gives you stupid

canned talking points. And so that even from a qualitative standpoint, I I don't wanna use an AI model that that renders the bot useless, basically, for my audience. And then also, of course, you know, I anti censorship free speech, Morpheus is a is a perfect network a perfect AI network to power such a a chatbot. It's the the the perfect system

to resort to in this situation. Yeah. That's And then also to monetize, by the way, you know, I know I I came across this recently that you made it much easier now to list the stakers, and I immediately implemented that because now, you know, anyone using Morpheus, they can check the wallet address just using public URL and see if they've staked to their stuff net and then give

unrestricted or preferential services to them. That's what I'm doing on the same button now. I got my first real staker actually a couple of days ago, and that that that's really you know, it's just another thing that I love that you you made it so much easier to obtain that kind of information. Because previously, we're going, you know, through this dunedune.com proxy. Yeah. And that wasn't really that great, but now it's just, you know, perfect.

Kyle (Bow Tied Blue Fin)We we really love

Thomas BorrelI'm sorry. Hold on. Hold on. Let me let me interrupt there just a second. How do you spell that how do you spell your website again for for anybody that might be listening?

Irida TVIt's descentbot.com. Descentbot.com.

Thomas BorrelAwesome. Thank you.

Kyle (Bow Tied Blue Fin)Yeah. I really love to hear that on, you know, the staking side, and we'll get to the other side in a second. But, like, you have really invested in the full Morpheus platform. Right? And you're kind of using it as design. You created a subnet. You said if someone stakes this subnet, then they get a certain level of access on my product. And and now you got your first staker. Right? And that or the first,

you know, staker in in the in the new version of this. But Mhmm. It it's such a and you're able to use the new systems to be able to act to accurately see how much they have staked on a programmatic basis rather rather than the June dashboard. So, it it's great to hear that, you know, the work that we've done on trying to simplify that experience is actually being put to good use. Right? That that's always something great to hear and and that you're having a good experience.

Now getting back into the comments before of, you know, centralized versus decentralized, open source versus closed source, that's kind of exactly what we are hitting on. Right? And it's it's not just the censorship and and the access, but it's also it it can't be turned off. Right? And it and it's the permissionless part of it. Right? Even if GPT was answering you,

you know, Sam Altman or team can say, we don't like this. Right? Our donors or our funders don't like this.

Bowen BelmerCan you know? Wait. Wait. Wait. Sam Altman wouldn't do that. Not Sam.

Kyle (Bow Tied Blue Fin)They can unilaterally decide to shut you down and fill your business if that's the only source of interest. And there's risk there, when you're dealing with something that might not be completely aligned with what they like. So so if you could just talk, I guess, a little bit about about that and that risk, and I I just think it's something that probably relates really closely to you that that we see a lot of people relate.

Irida TVYeah. I mean, they introduced this centralized risk. You know? It it I would be building my whole platform hoping and having to beg to the these these mega corporations to be able to stay alive. And and with Morpheus, I don't have to do that. And I actually built my app with Morpheus and Venice from the get go, so I will never even,

you know, have that kind of dependency. Never run into that situation where somebody can shut me up. They don't like what it is that I'm doing, which is exactly what they do. You know? It's exactly what they will do down the road. And it's just this trend of these mega mega corporations consolidating all this incredible power in, like, a few people that's very dangerous. And, you know, David has been talking about this for over two years now.

I I love what Morpheus is doing for for exactly that reason.

Kyle (Bow Tied Blue Fin)Amazing. You know, we feel the same way, of course. The we've seen a lot of this, and I think Bo talked last week about the use of different models in different countries. Right? And you see open, you know, close oop sorry. Open source models being used in a lot of the world because of the inability to access. And it's not just from a location basis or region basis, but it's also from a content basis, like in your case. And,

we just love that you're kind of the the perfect explanation of of why we think that that's valuable. So we we Well, also yeah. Also,

Irida TVthe it seems like at this point, you can let's say you wanna use GPT. You can you can use Morpheus and and get, you know, GPT LLM through it. So down the road, I mean, I think it's, like, gonna be so irresistible to use open source AI even if you wanna use a mainstream element.

Kyle (Bow Tied Blue Fin)For sure. I think yeah. The so I'll speak to that very briefly. The so since Morpheus is, you know, a a permissionless marketplace, people can come in and offer any models that they choose. And the benefit of even if you did use GPT, for example, through Morpheus, you get the benefit of the anonymization and and the proxies where there is no tie to an individual user, right, through

that connection. And and I think, you know, Claude will be added on there at some point. GPT will be added on there at some point amongst others. It's that, I think, the the bigger value there is is the privacy aspect of you're able to get access to some of those in a in a middle ground where maybe some data is being used for training, but it's not tied to you. It's kind of like, I I I'm going halfway there as far as my my

sliding scale on on privacy. And I personally choose the open source route in in many cases. I use the closed source route in some cases where where I can, but I think that sliding scale is important where we we don't need the the Lamborghini privacy for everything maybe, but, for, you know, some cases, we can kind of branch off where where valuable.

Thomas BorrelQuick question. Yeah. Well, actually, just I'm gonna do an add on to that, and then quick question. What what the I think, and what you were saying about people can provide any models that you have. Not only is it just the models, but they can also add you know, because they can set the pricing. You know, some of these some firms, what we're seeing and and I think, you know, Brian's got a couple of these questions, you know, have excess capacity that they've negotiated. So they've negotiated negotiated specialty rates, and then they can pass that through because it's it's it's excess,

you know, it's it's excess it's excess tokens that they otherwise wouldn't use. And then they can say, okay. Hey. I've got 20% over this month. Let me go ahead and monetize that. I've already got it, you know, I've already got I've already got a rate negotiated for this set of tokens, and so then they can, you know, they they can it can be feedback into their bottom line. So just to add add on to that. But It's a great question.

Kyle (Bow Tied Blue Fin)It's it's a great point, and, I think there's probably two key areas that are a little bit different about that that are that are both super important. Right? Even think about some with a clawed max subscription. Right? They're like, alright. Well, I'm using more than the pro, but not as much as the max. I'm gonna offload some of my capacity there. Right? I have a contract with OpenAI for $25,000 a month that I'm going to spend, and I'm only spending 20, but I wanna monetize it. And if you extrapolate that even more to open source,

you have people who are saying, I have a data center that's running at 71% capacity, and I'm comfortable going up to 75% capacity. So I'm going to host and offer those models to monetize something that I previously wasn't monetizing, and that's kind of what we're talking about from a provider standpoint. Now when you have all of that funneling into the system, you know, who gets the benefit? It's the consumers. It's you're able to get access to lower cost AI

because it's, you know, the use of that excess capacity. And so it's really a you know, it's it's a flywheel where consumers benefit and then providers benefit and then consumers benefit again because of the increased number of providers who are entering the system. And it really just continues to scale up and continues to provide more value as the ecosystem, you know, becomes more mature and and scales. So that's a great point, and thank you for for

highlighting that.

David JohnstonI'm curious. What are some of the models that you're seeing the most demand for?

Irida TVFrom my side, mostly, I I see I mean, Venice, uncensored, and I've seen deep sea. I don't have necessarily a lot of traffic yet, so it's you know, I don't have a huge aggregate of data. But when Venice happens to be the default choice, a lot of people use that, And it's just, like, in my case, it is a very good choice because you're dealing with

controversial content and uncensored content. That's what I'm yeah. Yeah. That's what I'm seeing. They definitely have a good,

Kyle (Bow Tied Blue Fin)fine tune of that, I believe it's the Mistral model. I I don't remember exactly that. It's a dolphin, I think. Oh, yeah. Yeah. It's dolphin. That they have forced the ability to have even more uncensored content. So I I agree. I think that's a good uncensored route. Maybe not the most intelligent model if you compare it to something like, you know, 250 to 500,000,000,000 parameter model. But for the specific use case of uncensored, I think it's a great option for for a lot of people.

Irida TVI mean, I've played around with a lot of the other, you know, really high intelligence models, and many of them, I would say, are also pretty well uncensored. And, yeah, they do some of them do provide a little well structured responses and better formatting in some cases. So, yeah, I I agree there's probably some more intelligent models. I've also so the way I've instructed

the chat to respond is that whenever, let's say, there's a nostr discussion on the topic, and it links to the nostr discussion. And it was driving me nuts for a while, but I figured out it's something how to word the instructions to where it really nicely format all response, incorporates all the sources, and even links to notes to discussions if if there are any.

Kyle (Bow Tied Blue Fin)I missed the final part of the Oh, can you hear me? Yeah. Yeah. I could hear you. I could hear you.

Irida TVYes. Where did the instructions to where the models, the output, incorporates all these sources from sort of some of these Yes. Yes. And censor journalists and also, and also incorporates no sir discussions on the topic if there if there are if there is one.

Kyle (Bow Tied Blue Fin)Is that relating to the the web search aspect of it, or is that relating to something else? No. I'm sorry. I wasn't exactly I'm not following a 100% on on the question. That's why I'm just struggling a little bit. Well, I wasn't really asking a question. You talking about how it does that search and how it does those references. Okay. Got it. Got it. Got it. Sorry. Yeah. Apologies for that. No worries. I I agree. The I agree. I think the web search aspects and and some of the new web scraping aspects are

you know, there have been incredible improvements on on that front.

Thomas BorrelWhich model just out of curiosity, which model are you guys seeing as the one that's most censored where you look at it and you think, oh, that model whatever I stick in, that model's not gonna give me the actual answer back. Are you what which one are you which one is the the most villainous or centralized at the moment? Which which one's flexing the hardest on its on its user, but it's on its sharecropper users? That's a great question. Name names. I

Irida TVthink off the top of my head, I wanna say I've actually been kind of disappointed with the Venice Larch on on that at times because it it would resort to, you know, more mainstream narratives. But overall, from the models available through the Morpheus API, don't know if they're fine tuned or whatever, but I feel like they're all pretty good in in that regard.

Thomas BorrelI was I was throwing in the centralized, so Quad, ChatGPT, Gemini. Oh, yeah. Okay. Yeah. For me, I was I was I was doing probably the ones that canceled you the fastest might be one on the list. I guess Yeah. Trying to get a picture on that. What what do To me, like, the most ridiculous one is from ChatGPT first. For sure. Because of the you know, immediately,

Irida TVit flashes warning red messages. Can't answer this. You know, can't talk about that.

David JohnstonWell, I think and that's that's for me the the big thing is, you know, you're using any of those models, you don't own that intelligence. Right? You're renting it from some big tech company, and they can take it anytime they want. Right? To to use Chris's analogy, you know, you're a sharecropper. Right? But the landlord gets all the benefit. And so for me Mhmm. What motivates me so much about this is I wanna own my intelligence the same way I wanna own my Bitcoin, same way, you know, I wanna own my house. You know, I don't want to, you know,

owe owe it to somebody else. Right? You know, in that kind of system, it's to really make progress. Right? Because whatever progress you make, they're just, you know, taking that value from you whether you know it or not, whether you see it or not. You know? And so, yeah, personally, I'm pretty encouraged, you know, looking at some of the numbers. I think Venice users are getting close to 2,000,000 and people using the app to power 20,000,000 plus users access to open source. So, you know, to to see, you know, Morpheus and and sort of all the ecosystem that's built and built on top, it's it's pretty encouraging to see that there's that much demand for for what's what's getting built. So but, yeah, I remember that that early

interview, and, it's so awesome that you've since then, leaned in and built all this on top of Morpheus. Like, that's that's what we're all hoping for. Yeah. I'd I'd love to hear more about kind of the distribution strategy for DescentBot.

Bowen BelmerI mean, if there is one, like, are are you actively marketing it to people for various use cases? Like, do you have a network of people who would benefit from it? It sounds like, you know, obviously, your own journalistic efforts are are largely part of the motivation. But, yeah, what what else is going on on that side of things?

Irida TVYes. Not not much at the moment, to be honest. I mean, I have it listed on my website, descentwatch.com, as the bot component to it, but I haven't run any ads or any campaigns or anything like that. It's been kind of like a a hobby project of mine. You know, this is the first time I'm really talking about it anywhere. So I would love to do more. You know, I just haven't been haven't been very proactive in marketing so far.

Bowen BelmerOkay. Yeah. Well, no. No worries at all. I mean, I you know, it's just one of those things that I I don't think we talk about this a lot, but ultimately, I don't think people are gonna realize the importance, unfortunately, until they have to. But but then once it happens, it happens all at once. We actually we just talked about this on on the podcast episode we put out yesterday. We had Tina Baker Taylor, who is a cofounder of Venice, on the show. And we, like, form discussed a lot of these topics, and and one of the things we talked about is Irving based in The UK.

I mean, there are literally people who have federal agents showing up at their door because of social media posts. And Yeah. I don't know. I mean, I hope we'd never see that in The United States, but I also, you know, unfortunately wouldn't be extremely surprised if we did. And as those things accelerate, I think people will come to realize why stuff like what you're building and why Morpheus and Venice and these other things are so so crucially important. I also think that, you know, when we talk about the centralized providers, we can even take it another step further because on top of the fact that

a lot of people are predominantly, if not only interacting with the centralized providers, I think a lot of people are mostly interacting with AI in a work capacity. And a lot of the time, that is like a centralized model wrapped in whatever their employer's compliance team and tech team decided to throw on top of it, which is even worse because now you're getting Sam Altman propaganda and your employers on on top of it, which is just, you know, horrendous.

Kyle (Bow Tied Blue Fin)Exactly. I think a a good comparison as we go into this the weather situation of there being a huge snowstorm across The United States this weekend is like I was just thinking of the comparison of, like you know, you go to the supermarket and you buy a bun you you buy a bunch of of food and you buy this snow melt and you get your shovels ready before the snowstorm. Right? And it's so that you're prepared. And

seeing what's happening in The UK and, Bo, you made the comment of, you know, you hope this never happens in The US, but if it does, you know, you you wouldn't be surprised if if something like that do happen as far as information censorship. Like, I wanna be prepared. Right? I I wanna have my stuff together so that I'm using an uncensored and permissionless platform already. And if that day comes, my companies and products and services are not going to be impacted by it. Right? I'm I'm ready for that storm because I'm I'm using something permissionless, and I'm using something open source and noncentered already.

So just kind of kinda close it down a little bit with with, like, an analogy of something that that we can kinda relate to.

Bowen BelmerYeah. For sure. Well I actually I think that this would be a really good at least within The US portion of our community, you know, of which we do have a lot of, you know, people all across the world working on Morpheus. But I I think as a community, we should really so crypto and definitely freedom tech has been kind of, you know, viewed through the lens in The United States of, like, a more conservative ideology,

especially politically because of, you know, things leading up to the last election and then also, like, the actual election cycle itself. But I really hope we can continue to frame it as, like, good for everybody because this really should be, like, a non like, any individual person should should want this, you know, no matter no matter what it is your opinion is on, like, any other issue. If if we recognize AI in general, it does exist and will continue to exist, like, this is ultimately good for everybody.

Because my big fear is, as we know, like, these tides, you know, turn and and, like, the pendulum swings, And there's so much extreme discourse right now in things where the spotlight is not on, like, this type of thing, but it can change quickly. And when it does and when, like, a dissenting viewpoint now has some type of, like, power and control again, I don't want the spotlight coming back on,

you know, what what we're doing and, like, what people in in DAI are doing in general. Obviously, that's the whole point of building DAI. Like, that's why you build decentralized antifragile systems. But I but I do think that there's this kind of social in The US affiliation between, like, freedom technology and one side where the reality is, like, we have to convince people or not even convince them. Like, help them realize that, like, it literally doesn't matter who you are. Like, this is this is a net positive for you, whatever it is your belief is.

Irida TVAgree. Yeah. I think it's, like, I don't know who said it, but, it has to sometimes it has to get really bad before people, realize how how bad centralized options are before they resort to decentralized. But then the other hand, as David has pointed out many times, right, now everyone uses Linux. Almost all servers, even corporate servers in the world are Linux. So in the long run, the open source will win out either way.

Bowen BelmerTrue. Yeah. No. Ultimately, yeah. Totally agreed. Is there anything else that, you know, you would benefit from with regard to DescentBot? Like, if Morpheus could do x y z thing that it can't currently do or maybe x y z thing in a different way from your perspective? Like, if you can make a change, what would it be?

Irida TVYeah. I don't know I don't know if the image generation is available yet through the Morpheus API. I'm also using the Venice API, and, you know, people can switch between Venice and Morpheus on the distend, but so that would be pretty cool if I if if if we could have image models in addition to the language models or I haven't seen that yet in the, Morpheus API. Another thing that would be nice when I pull the language model,

names into the drop down. In Venice, there's a, you know, friendly name, human readable name, and then the kebab case language model identifier. In Morpheus, I think I only have the Kebab case identifier. If I could have a nice human readable name in addition to that with space separated, that would be great. If it's already available, then,

you know, I I haven't seen that yet. So these are certain things that, you know, come to mind.

Kyle (Bow Tied Blue Fin)Sorry. I'm just writing this down. Give me one second.

Thomas BorrelI have a quick question while we're we're waiting on that, and this kinda goes to both Joseph and Nima. Where where are you guys you know, as you guys are building in the open source, you know, Freedom AI, Where where are you guys finding other people that are that are building like you are? You know, are they in any particular forums or discords that that you guys are finding brothers brothers in arms in this fight against,

you know, censorship? And where yeah. Where where where are you finding where where you've where do you find your people? Like, where where where are they hanging out? Where you found the value?

Irida TVThat's a great question. I I I mean, the closest I have is or the the only thing I have is probably the Morpheus Discord. And I other than that, I've been doing it all by myself pretty much. I haven't really gone to meet meetups, which I wanna do, you know, all these things to meet people and and exchange thoughts and ideas. I wanna do all these things. I just haven't, you know, gotten to it yet. I guess if I could re

Kyle (Bow Tied Blue Fin)reframe the question a little bit, I think it's more of what where do you go for, I guess, non Morpheus information. Right? It's like, are you're you're building a a very interesting product or you built a very interesting product. Like, where where are you looking to get your users, I guess, or or where are you hanging out to get the top latest information in AI? And I guess, is that what is that what the question was was focused on?

Thomas BorrelI think the question for me, it's it's more about, like, you know, the you know, you're you're you're, you know, you're building something right now. You have a specific use case. You've you've kind of experienced getting shut down. You can have the benefits of that. And and and, you know, Joseph, as you're building as as well, where where are you going from you know, where where are you seeing other builders that are similar to you? You know, are there are there any discords or are there there projects where you're, you know, you're able to resonate, you know, help solve problems together? Can you guys kinda have similar

similar tales of and and challenges that you're navigating. Are are there any discards? Are there any subreddits? Where where do you find the resources? Can you you know, said you kind of doing it on your own. Where where do you find these? Where do you find the most value?

Tales & ConquestI think for us, definitely, I I would echo what Irida said here with the Morpheus Discord, the Venice Discord, as well as any DAI events in the past couple of years. I don't know if there's really a subset of deep end evangelists on Reddit or or some of these other forums. So I'd be keen to to learn more about where you guys kinda scour your brothers in arms. But

for for us in the Tales and Conquest side as we've been building this game, we've we've been pretty heads down and just focused on the community around us that's built in Morpheus and in our the sister project of Venice. But yeah. Where do you guys kinda find folks hanging out these days that are interested in this?

David JohnstonHonestly, for me, it's mostly on x. I've dabbled on their platforms. I haven't found them on on LinkedIn, which, you know, continues to be very corporate. Right? Reddit definitely has a a different culture, though. There's a good active, Venice subreddit. There's a few others. I think it would probably be worthwhile to lean in some there because that's actually where a lot of the training data is coming from. And so if you want your world view to be included

in the future AIs of the world, it definitely pays to put it on something like Reddit, especially because Reddit's willing to sell their data to all comers versus something like x, which, probably is only getting, I would guess, trained, into Grok as opposed to, you know, giving that fire hose of data or at least they're charging a lot of money to give that fire hodges of data to others. Right? And so, yeah, you know, as the community grows,

a lot of times I resort to going places in person, meetups and conferences and stuff like that. But for me, I'm mostly living on x these days.

Tales & ConquestUsually, LinkedIn's only the place we go to to to learn how not to do AI.

David JohnstonSome of my favorite memes are, hey, guys. Just to warn you that the LinkedIn people finally heard about this. It's

Thomas Borrellike,

David Johnstonwait. Or or wait till the LinkedIn people hear about this. Yeah. So Isn't it funny though how,

Bowen Belmeryou know, speaking of LinkedIn and, like, there's just an endless there's AI slot posts everywhere. Right? But, like, they're particularly bad there, and also on Facebook, I would say. But isn't it funny how, like, we're slowly learning to read the language of AI writing? Like, I feel like I can pretty accurately discern if some type of, like, medium to long form piece of writing was was written by AI or not. But and it's only been around for what? Like, two and a half, three years that regular people have been using it?

Irida TVOh, yeah.

Tales & ConquestIt it is interesting. And so nowadays, we we that use it to help us cut Brian right in a format that doesn't look like AI slop. So we have to tell the AI to remove the m dashes, the prefixed unordered lists, and

Bowen Belmeryou you it's very easy to spot them for sure. For now. I mean, I'm sure very quickly we won't be able to tell the difference. But, yeah, for now, it just it does make me laugh because it's it's still, like you're, like, learning the language of being able to tell. Images are getting They're really

Tales & Conquestthey're getting great as well. You know, the funny thing that I find refreshing and I'll I'll look to for more of their content is if one of the posters has a misspelled word. Because then I know it's a human.

David JohnstonProof of human.

Thomas BorrelThat's been noted and now added into a a model where you can just add, like, occasionally misspell a word. Yeah. Grock Grock is gonna start making spelling mistakes. Exactly. No no Em dashes. Add a misspelled word.

David JohnstonYeah. I I had almost never seen Emdashes used until the recent spike in AI content.

Kyle (Bow Tied Blue Fin)So I didn't even know what an Emdash was.

Bowen BelmerI know. Right? I I did the opposite. I used to use them all the time specifically because it would, like, set your writing apart usually, and then it did the exact opposite. It was like, never mind. You're you're a horrible writer because now you've an M dash. You didn't write this.

David JohnstonYeah. It it copied you.

Bowen BelmerWell, I don't think it was me specifically. Certainly not. But it was it was it was definitely, like, something that I utilized and then had to completely change, especially my my old blog that I used to write Yeah. On real estate stuff. It was like, I had to completely change the writing style because I realized it looks automated now.

David JohnstonYeah. I mean, somebody pointed that out to me. They're like, wait a minute. You didn't write this. There's Emdash. It's like, what are you talking about? They're like, oh, AIs use all the Emdashes. This is fake. I was like, okay. That didn't even occurred to me, but good good to know. I guess I'll stop using them. But yeah. Yeah. You're having to, like, time stamp stuff and be like, I wrote this 2019.

Thomas BorrelYep.

Bowen BelmerThere you go. That's funny. Cool.

Tales & ConquestI love to circle back to the question I answered earlier as to, like, why a builder might use Morpheus. And I know in previous spaces, I've talked about the kind of decentralization and privacy aspects, but cost today is a big factor. You know, I was I follow on x, like, one of the guys behind the Eliza projects, the developer Shaw, who's now building, you know, a RuneScape style MMORPG

game, study using Claude to essentially vibe code through it. And he shared recently that his monthly Claude bill was $35,000. So the more we can do to democratize AI and put it in the hands of people running it on their own machines, I think that will drive down cost. And that's one of the reasons we're really excited about Morpheus and this API as it goes live is we think it's gonna be way more competitive for our needs at Tails and Conquest because we don't need

frontier models. We we just need quality open source models that are are cheap and scalable.

Kyle (Bow Tied Blue Fin)Yeah. I I was just talking about about Shaw exactly yesterday, and and I saw that same post, like, $5 or whatever it was. Crazy. And I think we're really close, right, to coding models being there. And, say, every week, see new models come out that push that gap closer and closer and closer. And I think there's most applications right now are,

like, open source is sufficient. I think there there are some applications where you may choose to to use the, you know, what I said, the Lamborghini, but they're becoming more few and fewer and far between. So and I think we we perfectly agree on that, which is great.

Bowen BelmerOh, jeez, guys. That was a lot of dead air. We need to really jump in here.

Thomas BorrelOh, I think I was I was, like, sitting there. I'm like, I started talking, I'm like, I'm mute here. So I that that probably is the that's probably the best thing is to make sure that you're off mute.

Bowen BelmerYeah. No worries. Did you have something?

Thomas BorrelI I I did not. I was trying to fill the dead air just like just like you were there. Yeah. No worries. That's that's what I'm here for, dead air connoisseur. I I think you you you brought up the 35 the 35,000 a month. You know, it's like it's you know, I think as my kids have been have been have been coding. Right? And so they've gone from you know, I think it kind of exploded from on the Claude code side, and, you know, they started discovering, you know, retro games. And so, yeah, they're like, hey. Let me go build my retro games. So it I think at the first day, they got Claude code,

then they installed it. And then I think within two days, they had run through all of my credits by by noon. And so I was getting throttled, and, like, my monthly credits were, you know, were throttled out. And then they then they switched over, and then they're running through their Morpheus and their DM. So they really are a bit like crack addicts. So you that 35,000 a month, it's like it goes from a from a, oh, hey. It's not it's a nice dinner to, you know, I'm paying for a car. I'm paying for a sports car. And now, you know, hey. You're paying for a Bugatti and a and a a and a vacation

a vacation place in Vail, and I is is what Shaw's doing. So it'd be great to see kind of on the topic of, you know, Shaw spending that. They're, you know, they're in and Eliza, they're building solutions. It'd be great to to to see about doing spaces with with some of the other guys that are, you know, building the space and building solutions for, you know, for for everybody, you know, tooling for everybody to use so that you can get that inference cost down. Because if you use those tokens more efficiently, if you're, you know, kind of be able to get your stuff done,

you know, it allows for those it allows for some of those costs to come down. So just more of a more of a general comment than a question for anybody.

Bowen BelmerWhat On my side. Platform are they building games for? Like like, simple computer games?

Thomas BorrelYeah. No. I mean, they're they're they're doing it locally, so then they then they integrated their you know, the controllers. You know, they're doing it. They're they're using OpenCode, tied into Morpheus. So when the API went down, you know, I think one of the one of the early teens came in is like, hey. Something's not working. And and, you know, Kyle got a bump. You know, like, hey. Hey. Hey. This went down. But yeah. So so doing that, and and now I've got it all the way to where the kids are yeah. They're all looking at what they can do locally because I think, you know,

anytime you're ending up in the AI, you know, they've participated in Morpheus. They participated in Venice. You know, you start looking at you you start seeing those inference costs come in. And so the more especially if you're engaging with something like Sisyphus or some of these you know, your Ralph, you know, you can set you can set a fairly complex problem in, and, you know, it's gonna then the computer's gonna work for ten hours. And then it's gonna run into your API, you know, the the the your rate limiting. And so you really have to start looking at start to look from a budget perspective where where you're using that. And if you can use that locally,

great. You know, we're open in this household, we're open source fanatics. So I think, you know, open code, try to do locally, you know, run it through Morpheus, run it through run it through Venice. I think, you know, the image generation is not not there on on the Morpheus side. But, yeah, as as as to to get just going back to, you know, we we don't wanna rent the taxi. We wanna try and own the car. And I'm trying to we're trying to do I think Brian talks about this a lot is having a twenty first century family, like, trying to get the kids to understand these concepts.

Like, how do you think? How do you frame problems? And then, you know, build from first principles. And I think for us, you know, for, you know, for first principles are, you know, you wanna own you you wanna own, you don't wanna rent. And then you also want to secure your position. You know, land goes up. I think, David, you said this on the the the, you know, you said this on the on the the the brainpower podcast. Right? You're talking about the, you you,

inference is, like, the new new new oil. Right? And you wanna you wanna own the oil well. You don't want to be, like, trying to buy the gas. You know? Yeah. I I forget what your analogy was there, but it was I I you know, you when you and Tina were talking, I think it it it it really resonated.

David JohnstonThanks. Yeah. No. I think that's that's the key to all of this. So and that, I think, is what motivates a lot of us. But it's been really awesome to, sort of bring together the builders who can do a lot more of this, going forward. You know? And a lot of what we've been imagining is is now real. Right? The open source models have caught up. The tools have gotten extremely compelling. It's no longer a tool. It's become this critical thing in our lives, whether you're coding or writing or transcribing a meeting. And so, yeah, I think, we're just at the beginning of this this whole process. So, you know, I'm gonna keep on building, and, my whole thing for 2026 is just doubling down on Morpheus and open source

and pushing all of this to the next level.

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