← Back to search

Why OpenAI Should Be Scared of Open Source

The Delphi Podcast · 2026-10-26 · 75 min
relevance 100 13769 words Episode page ↗ Audio ↗
Show full episode description
Tarun Chitra is the founder of Gauntlet and managing partner at Robot Ventures, which just led Nous Research's $90 million round. He joins Tommy on The Delphi Podcast to explain why he backed Nous and Hermes Agent after months of diligence on OpenRouter data and GitHub forks. Tarun traces how crypto's first run at AI went wrong, from GPU networks to token-launched agents, and why the crypto people who dropped the token have done well. He breaks down why OpenClaw is the worst developer experience he's used, why Hermes is closer to Ubuntu than to Android, and how robotics makers built on Hermes without Nous knowing. The conversation closes on his market structure thesis: AI inference looks like crypto, OpenRouter looks like a DEX, some providers game cache pricing, and the money will flow to the harnesses on top and the inference providers underneath. Hermes, in his words, is like investing in Phantom or FOMO. Timestamps
✨ Episode Outline — click any point to jump to it in the episode
Problem solved
Explores how open-source agent harnesses could reduce AI vendor lock-in and become customizable foundations for business applications.
Benefits
  • Retain context and workflows when changing underlying models
  • Customize enterprise permissions, privacy, and database connections
  • Assess adoption through substantive forks rather than revenue claims alone
  • Evaluate inference costs using actual caching behavior
  • Reuse an existing harness instead of rebuilding agent infrastructure
Use cases
  • Investor diligence found Chinese robotics manufacturers forking Hermes and adding device-specific movement tools
  • The host built a Delphi site bot in Zapier that would run through 20 steps, analyze submissions, and send a message
  • The host used Claude Haiku in n8n to choose between DeepSeek R1 reasoning and a basic model for email responses
  • The guest tested inference providers and reported statistical evidence of providers withholding caching price benefits
  • Proposed enterprise deployment: clone Hermes, connect internal databases, and add privacy, compliance, and permission layers
KPIs / results
  • Robot Ventures led Nous Research's $90 million new round, according to the episode
  • Investor diligence lasted three to four months
  • Delphi's earlier Zapier bot ran through 20 steps
  • Illustrative caching comparison: 21 queries cost under a dollar and one cent versus a dollar and five cents
Tools / build
0:00 / 0:00
📑 Chapters — tap a time to jump there
00:00
Intro
  • Open-source AI thesis: own context and workflows while changing models underneath.
  • Robotics forks and enterprise customization underpin the investment case.
06:34
Robot Leads Nous's $90M Round
  • Robot Ventures leads Nous Research's $90 million round.
  • Hermes product recognition extends beyond awareness of its founders.
09:37
Crypto AI Was Meme Coins With More Steps
  • Early crypto AI projects prioritized tokens over useful infrastructure.
  • GPU mining conversions underestimated language-model infrastructure requirements.
14:40
Why Nous Stood Out Early
  • Nous pursued concrete research problems through YaRN, Distro, and model evaluation.
  • Distributed systems depend on many interacting optimizations, not one breakthrough.
  • Chinese model developers optimized around different hardware and data constraints.
28:14
Hermes vs OpenClaw
  • The speakers criticize OpenClaw's configuration burden and praise Hermes product defaults.
  • The host returned to n8n after struggling with OpenClaw.
35:39
The End of AI Lock-In
  • Model portability lets users retain context instead of rebuilding their agent.
  • Hermes combines learned preferences, reusable skills, and routing across models.
46:33
The Ubuntu of AI
  • Ubuntu analogy emphasizes practical defaults and compatibility with closed-source models.
  • Robotics manufacturers forked Hermes to connect agents with device controls.
  • Pi's lightweight design supports DIY experimentation and teaching harness concepts.
56:17
Microsoft and Nvidia Join the Round
  • Microsoft and NVIDIA participated in the round, according to the guest.
  • Enterprise deployments can add internal permissions, privacy layers, and database connections.
59:42
OpenRouter Is a DEX for AI
  • OpenRouter resembles an exchange connecting applications with competing inference providers.
  • Reported caching discrepancies can erase apparently cheaper token pricing.
  • Investment thesis: harnesses and inference providers capture value from token traffic.
[SPEAKER_00] There's a sense in which the rise of open source and open weight models in the last six months, like the immense rise. If you look at the fingerprints, people from crypto are all over that. OpenClause is like kind of the worst developer experience of like anything I've used personally. I think that kind of thing is like the difference between product engineering and like backend engineer. [SPEAKER_00] I would find these people who are like, I'm some random Chinese hardware robot manufacturer and the interface to my like little robotics embedded hardware thing was like fork termies. One thing I was kind of looking for was like, are there people who are lying about how much they're giving you in cash token? Like I ran these experiments and like tried to make sure I went through all the different providers and basically found there were some providers doing this. I think that market structure looks like crypto. All the money goes to the top and the bottom. [SPEAKER_00] What is the top? The harnesses. What's the bottom? Inference providers. Hermes is like investing in Phantom or FOMO, right? [SPEAKER_01] Hey, everyone. I just hosted Turin, who's the managing partner of Robot Ventures and the founder of Gauntlet on why he led Noose Research's latest round and his thesis on Hermes Agent. He's an early investor in Noose and spent three to four months digging into usage and the developer community and got extremely bullish on the opportunity. We're three-time investors at Delphi, huge supporters of the team, and I'm a vocal Hermes power user. Truett opens with the history of crypto and AI. We trace how early Noose has been. [SPEAKER_01] Hermes models with tens of millions of downloads, yarn for longer context windows, synthetic model evaluation, and distributed training through Distro. Now Hermes Agent. His point is that those years of research and infrastructure were in a detour. That's why the team understands the hard problems underneath the product. That's why his OpenClaw comparison goes deeper than the interface. His analogy is that OpenClaw hands you the parts and asks you to build the car. [SPEAKER_01] You're choosing the models, managing the context, and figuring out how much reasoning to pay for. Hermes was designed or had a different philosophy. Learn your preferences, generate skills automatically from how you use it, and make it easier to share context and route work across models. The It Just Works experience comes from research depth, meaning real product taste, with hundreds of decisions and optimizations compounding from thousands of users and those contributing to the repo. [SPEAKER_01] His bigger thesis is that Hermes could become a front end for AI, like Phantom or FOMO in crypto. You don't need to understand the infrastructure to use it, and as models improve, you shouldn't have to rebuild your agent or leave the context behind. The evidence that really stood out during his diligence was he found robotics manufacturers building on Hermes without Noose even knowing it. They were just trying an assistant. They were choosing it as the foundation for their own products. And the enterprise opportunity takes that a lot further. [SPEAKER_01] Imagine businesses having their own version of Hermes, connected to their datasets and workflows, with their own permissions, privacy, and compliance requirements. They don't have to build everything from scratch or accept a one-size-fits-all assistant. Turin explains how Noose is actively pursuing that enterprise market alongside its consumer community and discusses Microsoft and NVIDIA joining the round and how they can help lead enterprises to Hermes. That's the bigger opportunity when you unpacked. [SPEAKER_01] Not just an agent you use, but an AI foundation your business can make its own. The models underneath can change while the knowledge, context, and workflows you've built remain yours. It's why Turin sees Noose as potentially the most important open-source challenger to the centralized AI companies. My personal thesis is that Hermes allows businesses and people to own their entire AI stack as an asset, [SPEAKER_01] which doubles as providing them unique outcomes versus using the same models as everyone else and getting IPvarm. Satya and Adela from Microsoft, now it's Garth from Palantir, have voiced this thesis vocally. I hope you enjoy the episode. Hey, everyone. Welcome back to The Delphi Podcast. I'm your host, Tommy Shaughnessy. And today, I'm thrilled to have Turin on. Turin is the founder of Gauntlet, managing partner Robot Ventures, and has way better glasses than I do. Turin, how's it going? [SPEAKER_00] I appreciate it. Great to be on. And I think everyone has to go through their own glasses journey. I will say, the first time I ever really started looking into crazy glasses was I happened to randomly be in France 20 years ago. And there was this like, I randomly walked into this conference that was like the eyewear conference, and then that changed my life. [SPEAKER_00] So like I said, you never know when it'll hit you. You might suddenly, next time Tom's on, he's going to just have something crazy. He's going to have some steampunk glasses. [SPEAKER_01] Were you there for the eyewear conference, or is it just random? [SPEAKER_00] Random. Random. The thing is like, I was like just out of college, and I was like, ooh, I see free food. It sounded very stupid. I didn't know if you addicted to the glasses. And then I was like, oh, okay. This conference, I think, didn't really have that many people, and they over-ordered food. So there was like, here's some free food. You know, sometimes having a party, you never know what it'll create, what monsters it creates. [SPEAKER_01] I have these expensive Avulux glasses. They're like migraine glasses, but they look really weird when you wear them, so I don't usually do it. But they're amazing. [SPEAKER_00] You know, I will say, and this is a very controversial opinion, because a lot of people get quite offended when I say this. But I think I never would wear transitions, because you look like a serial killer when you wear transitions. But there are a lot of people who wear transitions, and they're always like very self-conscious about it. But like, okay, if you're in New York, you take the subway, or anywhere in the world, you take the subway. [SPEAKER_00] Every time you come on the subway, you look like you just came out of one of those spy movies, and you killed someone in the subway. It's like, I don't know, I can't do that. [SPEAKER_01] I feel like it's hard to trust somebody who wears transition lenses. Like, your life is that dynamic? You can't buy two? I don't know. [SPEAKER_00] Yeah, exactly, exactly. Like, sunglasses is totally fine. Not sunglasses is totally fine. But like, the middle thing, I don't know, something about it is like, I'm either a card shark or a serial killer. I'm not sure which one. [SPEAKER_01] Well, founders know if they pitch you or I to not wear transitions lenses or progressive lenses. [SPEAKER_00] There you go. [SPEAKER_01] Auto red flag. [SPEAKER_00] Auto red, yeah, exactly, exactly. [SPEAKER_01] Well, Turin, I'm really excited to talk with you. Robot Ventures just led Noose Research's $90 million new round. So, huge congratulations. Maybe we could start quickly there, and then we're going to dive into the eras of crypto and AI. [SPEAKER_00] Yeah, definitely. Definitely appreciate it. I mean, you guys, as well as us, we've been there for the whole journey. And News has sort of been through a lot. You know, I think a lot of people, the interesting thing, so there's oftentimes this type of advice I always hear. And, you know, that changes in venture capital over time. You know, I think in the 2010s, the advice was like, you want the company to be much more famous than the founder. [SPEAKER_00] You want, you know, you want the brand of the company to be known much better than the founder. Because that means, like, the founder is not just, like, wasting all their time on social media. This is, like, early social media era. And then in the pandemic, everything changed into the, like, the founder is the only marketing agency you have. So, like, you need them to always be on social media and have a public presence and image and whatever, right? You know, like, founder first marketing. [SPEAKER_00] And that sort of flipped this script where you, like, saw a lot of people who ran companies where they're more famous than their company. And then, fast forward to now, we're kind of in this, like, interesting thing where sometimes there are companies who have products where people don't even know who makes, you know, they have so many sub products that no one even knows the company, let alone the founders. And I think Deuce is a great example of a company like that where tons of people know what Hermes is. [SPEAKER_00] Hermes, Hermes Agent, you know, the very popular harness. I mean, their social media is extremely popular. And their social media is popular not initially because of Hermes, but because of their amazing design skills in the beginning. But then, obviously, they made some amazing products over time, but they've done a lot of different things. [SPEAKER_00] And I think the interesting thing about Noose is, you know, their evolution over time really reflects how crypto kind of had to learn what its place is in the new AI world. And a lot of that has changed over time. You know, what people thought crypto and AI would look like in 2023, basically, like, post-ChatGPT. Feels like forever ago. It has evolved so much. Right, exactly. [SPEAKER_00] And, like, in general, it's kind of funny because obviously everyone in AI is the I don't think about you at all meme about crypto. But there's a sense in which the rise of open source and open weight models in the last six months, like the immense rise. If you look at the fingerprints, people from crypto are all over that. [SPEAKER_00] And I think that comes from really understanding the lessons from crypto and, like, applying them elsewhere versus just, like, shoving a token into an AI thing and hoping it works, which is, I think, how, you know, a lot of the crypto AI stuff started. [SPEAKER_01] So, Taryn, you mentioned, like, a lot of early crypto in for projects back in 2023, especially, like, the rent-a-GPU ones. The idea that if you're running, like, a big Bitcoin mining network, you could kind of, like, flip a switch and, like, you know, serve AI inference, host models, things like that. But in hindsight, it does seem really naive, right? Because the models, the open source models weren't even around to serve, right? We've only, like, recently, the last year or two, gotten really good open source models from China. [SPEAKER_01] So, it just seemed, yeah, it definitely seems a bit naive in hindsight. [SPEAKER_00] Yeah, like I said, I think a lot of people saw the success of existing miners converting their operations. And then they also saw, to some extent, stable diffusion as an example of someone who was running a mining operation converting their existing GPUs into making a model. [SPEAKER_00] I think what people didn't realize at that time was that language models would just have a lot of idiosyncrasy is that, like, image models didn't. And, like, no one really wanted a decentralized image model, to be honest. Like, no one seemed to care about that that much. Um, and I also just think it's, like, it came from, like, crypto is an open marketplace. And so, people are inherently lazy. [SPEAKER_00] And they're, like, in the sense that they're just, like, well, hey, I could launch a token for saying that I'm doing this thing that, like, probably won't work. But, like, the token will still have some valuation. So, I think in the beginning there were a lot of, like, disingenuous projects that were, like, just, you know, meme coins with more steps, as I like to call certain types of projects. I love meme coins with more steps. [SPEAKER_00] So, I mean, it's, like, just, I mean, it's like, I'll tell you a lot, but it's kind of, like... someone who had, you know, worked in sort of centralized hardware stuff 15, 16 years ago. And then, you know, at the beginning, I was always like, oh, the decentralized things are all scams. And then after like four or five years, I like figured, finally understood sort of [SPEAKER_00] a little more of like, why do blockchains work? Like what is different about them? Why do all these kinds of things that you're taught in school about distributed systems and like grad school classes? Like, why are they violated, but still seem to work? And there's a sense in which because I was in crypto for a while, but also have been on the other side, I was like, you know, there is a sense in which there's like a 1% chance of like something, some architecture, some real [SPEAKER_00] kind of technical advancement, being able to get over some of these problems for decentralized training. Maybe I was a little less optimistic on inference. I was much more optimistic about. And I think the thing that stood out to me about Noose was like, they were one of the only people who is, who are straddling, straddling the line between wanting to try to do, you know, in 2023, this kind [SPEAKER_00] of like decentralized training, but also taking a more scientific approach to it. So they like, they wrote a bunch of research and did a bunch of experiments, some with, you know, one of the co-founders of OpenAI in the early days on how to do low bandwidth training. And at that time, I think, you know, realistically, there was only OpenAI and Anthropic that had real language models. [SPEAKER_00] There were a lot of like really shitty language models, like a llama probably just came out around [SPEAKER_01] then. Um, and have the, uh, I have to, I actually have it in my backpack. I've never thrown it out. Just the, the distro preliminary report, like distributed training that we're doing it. It was, uh, I think, I think one of the things you're getting at too, and kind of one of the things we saw early on was the team was just really forward looking, right? Like they were doing like synthetic evaluation of models. Like everybody tries to like gain these benchmarks, but they were doing that [SPEAKER_01] in a synthetic way early on. I think Jeffrey did, uh, yarn pretty early, which was like to increase the context length back when you couldn't put that much into an AI model. Then they did this distributed training run. I forgot if it was 40 billion parameters or something like that. Um, so I think to your point, like they definitely, they never launched a token. They never did like anything in for wise on, on the token side. It definitely seems like they kind of delved into [SPEAKER_01] a little bit, but also understood kind of the opportunity and where they could, could thrive [SPEAKER_00] there. And I think they were being rigorous at a time where like everyone who is like talking about the decentralized side was a little bit of a scammer or like, like I said, meme coin with more steps, like scam and meme coin with more steps are like, it's like, it's, it's a very fine line between the two. Cause like you could be well-meaning, but just don't have a good idea, but people still give you money for it. Right. And that, that's a meme coin with more [SPEAKER_01] steps. Well, what got you over the line early on? Cause like you're, you're kind of a longterm investor here too. Like, did you like you're, you're seeing like you come from a rigorous technical background and you're seeing a lot of distributed stuff in AI that doesn't really make sense, but this sort of, this sort of did to you early on. Well, well before Hermes. So like [SPEAKER_00] in this era. Yeah. Yeah. I think it's like distro. Okay. So they yarn as well as sort of some of their early research on like different types of embeddings for better compression, just sort of stood out to me versus like the normal crypto AI companies at that time who, because like they were like focused on very specific, well-defined problems versus like [SPEAKER_00] promising the moon of like, Hey, this system, this complicated system that has a million moving pieces will get perfectly right on the first shot. That's kind of like, you know, right. Any distributed system has like a hundred times or a million times more moving parts than like the centralized version oftentimes. Um, and you know, most of the computer IP people are like, Oh, we're just going to like, you know, take an existing blockchain and like somehow [SPEAKER_00] just like force the node node operators to run the, the kind of the software. And that's like, well, none of them can actually do that at sufficiently high bandwidth or at a level that, uh, was useful in the sense of like, okay, if I'm doing it at like 0.01 token per second, who cares? Um, and the other thing was like, there were a lot of things like BitTensor that were like kind of in this vein where they like really jumped too fast to the token and [SPEAKER_00] decentralization piece. But then the like utility of the actual projects was kind of weak in BitTensor's case. I think they found a medium ground where they kind of became this like funding platform, like launch pad effectively. And like, so arguably there's some residual value from that, but it didn't really feel like that was like a technical innovation. Noose were one of the only people who, you know, Noose and prime intellect realistically were the only people doing real research and [SPEAKER_00] like focusing on well-defined specific problems that would potentially, if you know, like every successful distributed system, um, is success by like a thousand little optimizations. It's not like there's one big idea that's like, oh, okay, this big idea automatically solves everything, right? Like if you look at even just Bitcoin going back to the original paper, like there's a bunch of different [SPEAKER_00] concepts that are need to work together in order for the thing to work, right? Like first of all, you need the hash functions to have very high collision resistance. Um, you need this difficulty adjustment to like not be gameable. Like the idea that I can based on how many people join and leave the network, it can like adapt to them. So like you can dynamically scale, you need the inflation curve to be, uh, not too steep and not too inflationary also, right? [SPEAKER_00] Like, you know, it needs to have some balance between those and you need the node software to have like sufficiently low bandwidth, right? That's why Bitcoin blocks are so slow. Um, and like each of these is like a separate technical problem oftentimes in, you know, from a different type of specialty, like a person with a certain specialty knows each one of those. And usually to build these decentralized and distributed [SPEAKER_00] systems, you have to bring those all together. So you need kind of a, like a wider breadth of knowledge rather than depth. And so I think the thing that noose and in a lot of ways prime intellect at that time had were these where they were like the only ones kind of doing, doing that. I think the difference to me for me, for noose was I think prime intellect maybe wasn't focused on the [SPEAKER_00] training side as much to, I mean, I know they did do a training run, but like, you know, they kind of hedged their bet and were like, we're also going to be like a GPU reseller at the same time. Whereas noose was really focused on this research piece. And like, that was sort of what stood out to me. Um, I was, I was like, you know, in order to ever have decentralized training or decentralized inference [SPEAKER_00] be a viable concept, um, you just need these 10, you need these 10,000 tricks that when you get there to get them all together, they compound and give you the final product versus like a one-shot thing that like, because like the ladder is like real needle in the haystack and often never exists in science and engineering. Um, which is what, what I think the early crypto AI stuff is like, there's a lot of [SPEAKER_00] hopium. It, it, it, it, in some ways it's a little bit, it reminds me of 2017 ICO boom, right? Where it was like staple a token and take something that exists already and say it's decentralized, like decentralized Uber. Like, you know, well, that's what I thought, I think that's the thing about new sets to that was like, they were actually trying to do the real thing by doing the hard work and not launching a token and not making promises that were crazy. [SPEAKER_01] Yeah, no, I'm, I'm with you. I mean, it is kind of funny, like mixing some hindsight with, with what we're seeing today in terms of like the token side, the distributed network side, I'm pretty sure for both news and prime is sort of abandoned, right? Like we've kind of from first principles figured out like, Hey, that, that doesn't work. Right. And that's fine. Right. They went through it, they tried it. And now it seems like there isn't a full focus on, uh, decentralization [SPEAKER_01] or, or distribute like distributed networks, but there's a huge focus on the open source side. Um, and that seems like it's going really well. I mean, Hermes has, you know, millions, maybe tens of millions. I'm not sure if downloads around the world. So I'm kind of curious, like how you link the two worlds together. Cause like you saw distributed centralized networks and then they were solving really hard problems, but now they're just sort of kind of the champion for open source. [SPEAKER_00] Yeah. So I think, you know, news, obviously one interesting thing to think about, right? So in 2023, everyone who's building open weight models didn't exist or existed the way like news stood, which was like a GitHub and a group of people and like maybe a little bit of funding, not that much. Right. Um, and arguably like the decentralized training people were competing with the Chinese model [SPEAKER_00] creation companies. And I think, you know, if you look at deep seek, you look at mini Mac, you look at, uh, uh, GLM and Z.ai. If you read the papers of like the model development, again, all of these things boil down to like, there's 10,000 possible tricks you can do. And you have, you only, you can only pick a hundred in your model. How do you pick the best hundred little tricks, like little optimizations that you can compound to, to make your model faster or better or whatever. [SPEAKER_00] And I think the Chinese companies rightfully focused on data collection and like basically making their own Mercore style stuff, except for much cheaper than in the U S and that allowed them to just like get better economies of scale while also doing a lot of these optimizations. But their optimizations weren't focused on like lower bandwidth or lower power requirement or make [SPEAKER_00] it decentralized. Their requirements were focused on like, and we have shittier chips. How do we make it run faster? And so it's, it's still a resource constraint, but it's a different resource constraint. [SPEAKER_01] It's pretty innovative though. Like the deep sea was nuts, like mixture of experts, second rate [SPEAKER_00] hardware. Yeah, for sure. No, no, no. I'm, I'm not at all knocking that. I'm actually just saying like if we look back in history, when news and prime intellect and other people are doing decentralized training was the same time. These Chinese model companies were also doing, Hey, we're going to build our own model. And they both started with the resource constraints. The decentralized training people start with the resource constraint of, well, it's decentralized. We don't know how much computational resource we're getting. Uh, and we have to pay for it and [SPEAKER_00] meter it and attract it somehow. So that's, that's like one type of resource constraint on training. And the, the, the Asian companies were like, well, we actually only get shittier chips and, and like resources and worse data centers and worse data. So given those constraints, what can we do? Right? Like, and so the, if you look, both sides may have a different bag of tricks, but out of the, you know, 10,000 possible optimizations, they both chose kind of like different subsets of 100. [SPEAKER_00] And the Chinese ones, their hundred was just better. Right. And like, that's why they got to, to building, you know, world-class models faster. But as we've learned from the kind of agentic world and the kind of recursive improvement stuff, once you have a good enough model, you just start increasing a lot faster and it becomes much harder for people to catch up to you. And so I think what you saw was like a lot of the early infrastructure people, [SPEAKER_00] a lot of decentralized training people by 2024, 2025, uh, kind of starting to pivot out. Right. And that was also when we started to see the first versions of like trading agents and early agents. Now, those things again, we're a little bit just like meme coins with more steps and a lot more [SPEAKER_00] replies on Twitter. Right. Yeah. And, and, and, and in some ways those probably set people back because like they were correct that agents were going to take over, but they were like the shittiest possible agents. Right. Cause like you fast forward a year from when AI XBT and AI 16 zero at the peak, that's when cloud code codecs, et cetera, we're just like having the huge growth curve, [SPEAKER_00] not just in users, but just like in capabilities and like stuff they could do. Um, and so I think again, crypto kind of right idea, like, like in the same way decentralized training was there like right idea of like having resource constraints might make you make a better model or like a competitive model. Um, but you have to have the right ones in the same way. It's like the AI agent thing. You have to kind of have the right thing. And like, you don't need the token piece in that in a lot of cases. [SPEAKER_01] It's kind of awesome to kind of go through memory lane before we get to where we are today. Cause it sounds like what you're saying is like in crypto land, we had people really approach AI with well-meaning intentions. Like they wanted, people wanted to train models like Newsome Prime, but the technical choices they made just like weren't competitive with the Chinese research at the time. And then like 2024 to 2025, [SPEAKER_01] like we had AI agents, but they weren't managing our lives or trading our portfolios. They were just like applying to us on Twitter. So it kind of seems like we're getting close each time, but we're not exactly on target with kind of what the, the big picture is. Yeah. And I think the interesting thing is like [SPEAKER_00] people in crypto who, uh, pivoted, um, have into like non-token AI have done really well. So like a great example of this being, you know, news, to some extent open router together, together AI actually was going to be like a decentralized inference. I mean, when they started, yeah, but it was a little [SPEAKER_00] bit different than that. They weren't like trying to be a network, you know, like they were trying to launch a token, like, like, like the people were trying to launch a token. I think that the ones who pivoted had this kind of now open routers sort of weird. They were definitely not trying to launch token, but, but in the sense of like philosophically, I think they came from this crypto view. And again, open router is a great example of finding the right optimization can get you [SPEAKER_00] under resource constraints, can get you very good performance in some of these systems. And in some ways that's like, that is the same ethos that decentralized systems have of like, I want to have the highest capacity while having low resource usage because the lower resource usage means I can have many more people participating. Right. And so I think there's a lot of shared ethos in the Chinese open source model development and the decentralized crypto development. They just like [SPEAKER_00] kind of went different routes. And I think that, to be honest, the crypto people didn't have as many, I would say the talent was not as good. And then also like they just chose the wrong optimization. Right. Now you look at something like news, the interesting thing, you know, first era they were doing decentralized training. They're going to launch this thing on Solana. And then, you know, they were kind of trying to figure out what to do. They're doing a lot of [SPEAKER_00] research or doing a lot of private inference stuff, working with kind of optimizing models for running in cities. Um, which, you know, what, what you actually see like near doing right now, Venice, stuff like that. Um, and sometime in 2026, you know, I think they saw open claw kind of have this moment at the end of last year, early this year. And one thing about open clause, it's like kind of the worst developer experience [SPEAKER_00] of like anything I've used personally. And, and I don't mean this in like, Oh, really? I'm not trying to be kind of an asshole. Cause I'm a noose investor. I'm trying to basically say like open claw feels like an Android because it's like, it makes you have to make too many decisions like to use it. Right. And it's like, if you, you, you look at the comparable centralized products of codex of cloud code of muse now of Grok bot. Well, you know, like all, all these things, [SPEAKER_00] they all have this thing where like my grandmother who doesn't know what AI is could use right. And, and open claw was like decidedly the opposite. It was like, tell me what inference agent you want. Tell me whether you want sparse attention or not sparse attention. Tell me like, um, your reasoning level, how much reasoning model choice, context choice stuff. Right. It, it, it, it had this Android feeling of like, you give me 10,000 options. And yes, [SPEAKER_00] I could choose a better thing than the defaults, but like you kind of hamstring who can actually use it. [SPEAKER_01] But I mean, the crazy part, I mean, maybe like give you a quick example. I was building an agent on N8N, which is like the, you know, more complicated version of Zapier. Um, and that Zapier is where I built our first, uh, bot in 2024, 2023. Like you come to Delphi site, it'd run through 20 steps, analyze, send us a message. So I thought, hell, let me make a personal agent on N8N. So I was doing that and building it. And I remember, you know, deciding, you know, using Claude Haiku to decide [SPEAKER_01] whether to go to R1 reasoning on deep seek or a basic version to answer my emails, things like that. And OpenClaw came out the same weekend and I stopped using N8N, which is basically where you have to draw arrows between services and build like this manual process, used OpenClaw and hated it and literally went back to N8N. It was just so clunky and hard to use that I just couldn't, I couldn't figure it out. I couldn't like get excited about it. Yeah. I think like the interesting [SPEAKER_00] thing is like OpenClaw in some ways is like the right idea for a product, but like the wrong execution on it. And like, there is a sense in which someone had this tweet the other day that I kind of stuck in my head a little bit, which is, um, like in AI, it seems like second mover advantage is actually like, there's like a real thing to second mover advantage and that you learn a lot [SPEAKER_00] from the mistakes of the first mover. Um, and like, you could argue that I would say like Claude getting a high uptake, you know, late 2025, early 26, right. Where like kind of eclipsed OpenAI, obviously OpenAI has come back with a vengeance recently, um, was because it was just easier to use and it like had more human like touches so that like the average person who does not technical could kind of make a product with it. Right. Like, you know, there were like reporters and [SPEAKER_00] journalists making web apps and dashboards, right. Without having to like know a lot of the technical language of like what to say to make the web app. Right. Like you really had this no code to think, which I think obviously OpenAI came back and caught up, but I think, you know, why did Anthropic have this opportunity is, was like, they learned from the lessons of OpenAI. I think arguably you could [SPEAKER_00] say the same is true for the open weight models. I think ZAI and Kimmy at like Minimax just like learned a lot from DeepSeq where like DeepSeq did a really good job of like super optimizing, like inference time, making the model small. Um, but where ZAI and Kimmy kind of, um, did a better job was like changing how they're thinking about reasoning, changing how they're like doing resource [SPEAKER_00] usage, changing a lot of the kind of like attention, uh, uh, topology, stuff like that. And I think the second mover thing is kind of very real here in that OpenClaw had clearly the right market, right? Everyone was excited for this agent concept. And again, crypto tried it a year earlier, but because they shoved the token in it, like no one was able to use it. Yeah. The other thing too [SPEAKER_01] about, I mean, like, I think OpenClaw deserves a lot of credit for like creating the category, right? For sure. And, uh, yeah, definitely like a lot of credit. I think the hard thing though is like you, you'd get like, you get people in like this weird AI psychosis with it where like they'd build stuff and they'd spend like 10 hours fixing it and then they finally fix it. But no one else could use it. [SPEAKER_00] Exactly. That's like the problem. Like they made something that was great for them, but then they, then they, but they couldn't ever get anyone else to use it. And so then no one else could be like, oh yes, this is awesome. I remember when OpenClaw came out, I want to say it was like December or January, uh, like end of 25, early 26, where I went to this meetup. That's right. [SPEAKER_00] Which is like an OpenClaw, like hackathon type thing. And it felt a little bit like going to like a crypto hackathon where it was like people building their weird little project of like, how do I like get my cat to like, you know, when it eats food out of the cat dispenser, it like sends me a USDC payment type, you know, like some, some, some weird ass project that makes [SPEAKER_00] no sense. And you're like, does this have a market size of one? Yes. And so the thing was like all these people who did these OpenClaw hacks and they were like presenting them or whatever, but like almost all of them were the worst demos in history. Like none of them, their demos worked first of all. Secondly, everyone who was like, oh, I thought that person was cool. Like, can I like download the setup? And it was like, no, actually it's like too hard to fucking transfer it to you. And so [SPEAKER_00] I think like the irony is like going to the OpenClaw events showed you a lot of the reasons, like love, what was like wrong with it from that perspective. And so I think Noose, you know, now, now coming back to Noose again, you know, kind of, they, especially Technium in particular, because it was just like his side project to start. It's nuts that it was his side project. Just insane. He just, he was just like, I love OpenClaw as a concept. [SPEAKER_00] But like, it's unusable. And because Noose has spent all this time building all this infrastructure for doing decentralized training, doing, they've trained a bunch of models of different sizes. They understood like the difficulties in that on the data side, especially, right? Because like, that's the hardest. [SPEAKER_01] There are models of tens of millions of downloads. Yes, no joke. [SPEAKER_00] But they have problems with like the, how do they get more data, right? Like decentralized data collection has always been much harder than centralized. And like, so for training, that's, you know, one problem that inference doesn't have. So then they were doing inference. In the process, they learned a lot about the infrastructure for scaffolding around an inference engine, right? So it's like, I have an inference engine. How do I pass context between it? How do I share context? [SPEAKER_00] And then I think, you know, around the time OpenClaw came out was also in like Codex and CloudCode were really coming out. And, you know, I think the thing people really got used to when those came out was this idea that I could have the same context shared between multiple models. And I think in the early days, everyone thought that was going to end up being a form of lock-in for OpenAI and Anthropic. Because they were like, oh, our models are going to have like a certain format for context. [SPEAKER_00] And you can move between our models, but like, we're going to lock you into that. If you want to move to an open model, it's going to be hard. And as of yesterday, that's completely obliterated, right? Like OpenAI now supports open models in the same context. So context sharing, all the stuff that like news, OpenClaw to some extent really focused on of like having the same context, be able to move freely between models across multiple agents, across multiple models. [SPEAKER_00] That stuff, I think is just like interoperability is here, right? Like to put in blockchain-y language. And so I think they were just like in the right position at the right time to be like, we've built all these things. We know why they're hard. We know why training's hard data. We know why inference is hard. Orchestration and figuring out how to like do context management. [SPEAKER_00] And I think technically was like, the problem with OpenClaw is they don't realize that. Because like they don't do that much of their own inference. They're not like, like it's like built around giving the user all the options to figure out, to go figure it out yourself. It's like, here's all the parts of this car. Go build the car. [SPEAKER_01] True. Could you, I mean, I'm glad you're getting into it. Because I think people might on the surface say, hey, Hermes agent is just, you know, a similar kind of better version of OpenClaw. But what you're describing to me is like, there's real architectural kind of genius behind it that makes it a much better experience without like an easy trick or two. Is that sort of the case? [SPEAKER_00] I think it's just like an Apple versus Google thing, but like in the inference era, right? Because, you know, a lot of the most successful, you know, the interesting thing is like all the crypto ideas that didn't work have analogs in no token land, like no crypto token land, that are huge companies and have worked a lot, right? So like the rented GPU stuff, all these inference providers have been raising at, [SPEAKER_00] you know, Decacorn valuations plus because there's getting a ton of demand. But the problem is what all, so like the Maslow's hierarchy of compute needs, as I'll call it, is, you know, everywhere. And then the beginning was like, okay, rent a GPU. Okay, well now I can rent the GPU, but then I have to go figure out how to set up the environment. I have to go set up the models. I have to go make sure I like manage memory. Oh, I wanted to run inference really fast on ultra fast mode. I have to go set up a ton of DevOps. [SPEAKER_00] And so now it's like, okay, I have to be a developer to start the GPU. And then there are companies like Together, Modal, Base 10, Fireworks, who are like, we'll give you a served endpoint for a particular model. We'll do all the optimization. We'll like optimize the inference engine, all the sampling. And so it's almost like they make an RPC for you. So the crypto version of this is like, one, I run my own node. That's run a GPU. [SPEAKER_00] Two is I use a node that someone gives me an RPC endpoint to, like an alchemy, right? That's what Together and Base 10, Modal, you know, you can view them as doing that. And then, but the problem is that's still not that easy. So like compare the experience in crypto between using like FOMO or Phantom to like going to Etherscan and submitting a transaction on your own, right? Like that's more like what the RPC, right? [SPEAKER_01] It's a good comparison. [SPEAKER_00] And so like clearly that like the TAM of the number of people who can use that while bigger than the one that's like run your own node is like still not that big. And the harnesses are truly like these like kind of wallets that manage a lot more state for you. They manage your history. They manage your preferences. They manage private keys and stuff. They like try to give you a ton of defaults so you don't shoot yourself in the foot. [SPEAKER_00] And so I sort of think of like OpenClaw as a little more like MetaMask or MyEtherWallet. Like it clearly has the right stuff, but it's like it's annoying to set up and it doesn't give you the right defaults. Like it still forces you to do a ton of stuff on your own. And I think the Hermes model, Claude Code, et cetera, is all about like giving you the right defaults, learning your behavior, right? [SPEAKER_00] Like generating skills based on how you're using it and learning about you as a way of like having some memory and like preferences. But then also figuring out how to manage all the infrastructure of like which model should I use? How much inference should I spend on? How much should I spend on this model for tokens? How much should I spend this model for tokens? And I think it sort of was built with the idea that a model router would exist. [SPEAKER_00] So I think one problem for OpenClaw is it's built under the assumption that the user will do their own routing. The user will choose their own model. The user will choose like what context they want to share or not share. [SPEAKER_01] Most people don't know the difference though, especially at scale. People don't know the difference between reasoning, non-reasoning at the time. [SPEAKER_00] Yeah. But they do know the difference when the answer they get looks shitty. For sure. [SPEAKER_01] But then they just think OpenClaw is broken, I think, for like mass retail. [SPEAKER_00] Exactly, exactly. But I think that's why, you know, hey, the MetaMask thing is the same thing. Realistically, right? Like the MetaMask thing is like the same, you know, it's like, is it broken for retail? There's a ton of FOMO users who could never use MetaMask, right? That's like, that's just a fact. And so I think that's sort of the way you should be viewing these. [SPEAKER_00] And so I think Noose was in the right place at the right time because they understood all of these problems one level below. Whereas OpenClaw was like, we're going to shove it to the user. That's their problem. And that difference in philosophy, while simple sounding and obvious, changes the architecture of how you design the harness, right? It changes like what decisions you unearth to the user, which things need to be their own sort of like models that learn something about the user, which things need to be asked to the user. [SPEAKER_00] You know, like when you use any of these bots, whether they're the consumer ones, like the Muse or GrokBot or the productivity ones, like Hermes, CloudCode, et cetera. They all ask you questions, right? They're like, hey, do you think this is the correct interpretation of this, whatever, right? And it's like learning something about a preference of yours. OpenClaw didn't do that. It would just be like, you should have written that in the config file. [SPEAKER_00] If you didn't, fuck you, right? And I think that kind of thing is like the difference between product engineering and like backend engineer, right? And like the problem is OpenClaw like treats its users as backend engineers. [SPEAKER_01] That's interesting. Okay. And then it is really cool to talk to you because you're so deep technically and you understand like how things are actually built. And then a lot of the design decisions. I think a lot of people try Hermes and they say, hey, this just works. But from what you're describing, the, hey, it just works is because of kind of a hundred correct decisions on the product side and the backend side. [SPEAKER_00] Yeah. I think like the interesting thing on the news, right? Is like they have this mix of like these technical people, but also this kind of like design product side stuff. And like, you know, I think a lot of people on the internet made fun of them in the beginning for like, oh, is this company? Yeah, exactly. Like, oh, is this company just like a very expensive merch company? Yeah. But I think that also teaches you a lot about like aesthetics about your user base. Like you care about the aesthetics of your user base and like what they're, what they have to do. [SPEAKER_00] And I think like that's something OpenClaw doesn't care about, right? Like, and not just OpenClaw. I think like a lot of the harnesses that were built at that era and still a lot of the ones you see now, right? Like there's this whole DIY harness thing, right? Of like people trying to build their own, like optimize it for, you know, my particular, it reminds me of people who are like, you know, spend two years customizing their text editor because like they need like every shortcut so that like when they're coding, it's really fast. Like 10 years ago, obviously now. [SPEAKER_00] Yeah. And there's like a DIY community for this harness stuff, right? And Pi has clearly taken over that community. Everyone forks Pi and starts their own. But I think what Hermes went for is like, there's going to be people who want DIY. There's going to be people who want enterprise. But then there's people in the middle who are like prosumers and like over time more consumer. And like that philosophy is what I really liked about what Nuse was doing is like they started with prosumers, but they made it easy to use. [SPEAKER_00] And then they kind of widened the space. And around the time they were growing, sorry, was when OpenRather started really growing. Like the two of them kind of grew synergistically. I sort of think of it as like the way Binance and Tron grew synergistically in some ways. But basically, the beauty of it was now there was public data, right? You got a leaderboard. [SPEAKER_00] You could see which apps were generating the most tokens. You could see which models were being used. You had like almost like blockchain crypto data for the stuff that you did not have prior to Nuse and OpenRouter growing. And that was when it was like, okay, this company is doing something really special. And like, yes, I've invested before. But like clearly this is like, you know, probably going to be the most important, you know, competitor to the centralized companies. [SPEAKER_00] And in a sort of crypto-like way of like being the most credible competitor to normal finance. [SPEAKER_01] Yeah, no, I couldn't agree more. I mean, Tron, I want to like double click on where you're taking this, right? Like you mentioned like the consumer agents. There's a lot of vendor-owned consumer agents. Like Meta owns Muse, OpenAI owns OpenClaw. Instinct, I guess, is its own. There's, you know, SpaceX, XAI owns Crock-Bot. Nuse is the only one that's open source and not lab-owned. [SPEAKER_01] Do you think that that has any implications in terms of like what people will use it for or the restrictions or like what they'll build or like second order effects of being open source? Because it's the only one on that spectrum that is open source, not vendor-owned. [SPEAKER_00] Yeah, I mean, there are, like I said, there is this whole DIY community. But again, it's like it's a little more like OpenClaw-ish where it's like you have to pick up, pick everything yourself. They're the only one that's like open source but package. So I think of it a little more like Linux distributions. You know, like there's the Linux kernel. There's Red Hat. And then there's people who build distributions on that. There's Red Hat, Ubuntu, Gentoo, whatever, right? [SPEAKER_00] And realistically, I'm probably going to get flamed for saying this by old heads, but Ubuntu is probably the only real Linux distribution that ever got like non-developed, like technical user usage, right? Like realistically, because it's like it is. [SPEAKER_00] And it's because they did all this defaults packaging, making it easy to update stuff, figuring out how to deal with like non-open source code interacting with open source code. So like Hermes is like is a good example of that, right? Where it's like I can use Claude and OpenAI with an API key within Hermes, whereas a lot of the pure open source stuff is like extremely like militant of like no, no, no, we don't want to touch any of the close. You know, Ubuntu had the same philosophy. [SPEAKER_00] So I sort of view it a little bit like Linux distributions. However, I think one really cool thing, another thing prior to us doing this investment that we're doing when we're doing diligence, obviously the open router thing made it be like, oh, I've turned this into a blockchain data thing. And so like for me, it's my background. That's like that made everything make sense. [SPEAKER_00] But the second thing was like I started just looking at GitHub of people who fork termies, like if it's open source, right? One way of measuring that community strength or people like the product is forks, right? And to some of the other competitors credits, like I think Pi is probably the most successful from a pure DIY standpoint. [SPEAKER_00] Like it completely eclipses OpenClaw, like from the DIY community because it's very lightweight. I don't think a lot of people know that, by the way. Yeah, yeah, yeah. I mean Pi, you look at the usage forks to Mintz, it's like surpassed. I think OpenClaw, the whole like OpenAI sale, pseudo sale thing, like that was like the death note. [SPEAKER_01] Kind of feels like TBPN. Like I don't listen to TBPN since OpenAI bought them. [SPEAKER_00] But yeah, it's somehow like they like sold themselves, sold too early, you know. But Pi is really community based. And I think the interesting thing about Pi is Pi probably has the most forks because it's like, you know, in fact, my brother is a CS professor. And he was saying like in one of the like intro CS classes now, they like fork Pi and modify it. And it's like, that's like how they teach you how the harness model works because Pi is like very simple. [SPEAKER_00] It, it, it, but it, it has this problem of like, again, you have to build all the adapters to everything that you might use. Um, so I was like, okay, it's Hermes open source stuff still sufficiently competitive for people who are DIY builders. You know, like, cause like, I think that's actually still important. Like their ethos should still be that it shouldn't just be, um, you know, Hey, we're an open source harness. [SPEAKER_00] We have a lot of users, but then we're actually like, no one uses our open source code because that sort of means you, something is wrong usually. So I started indexing a bunch of forks, looking through forks and I started finding some really crazy and interesting forks, um, which may, and these are not things news or Hermes will monetize. Right. But these are things of like brand recognition. People really care about how it's been built and they like how it's been built. This is a way of measuring that, right? [SPEAKER_00] It's almost like measuring doing sentiment analysis about a brand via people who forked it versus like, that's super interesting by the way, versus revenue. Right. Like I'm not, I'm not trying to assess like monetization on this. And I would find these people who are like, I'm some random Chinese hardware robot manufacturer. And the interface to my like little robotics embedded hardware thing was like fork termies. Right. So it's like, I, I, it wasn't like a unitree humanoid. [SPEAKER_00] They were like small robots, but like they don't have money. Right. Like they're trying to have to be cheap. Like the robotic stuff is like that to spend all the money on the hardware. Like they, that they need to lower their inference costs to zero. That's why they love open source models. But for robotics, it's, you know, especially, uh, you've seen a lot in a lot of the, um, uh, new evals. It's actually quite hard and competitive to have a single model be really good. [SPEAKER_00] In fact, one of the reasons I think the open source world is better is that I think the routers are much and like the ability to move between multiple different models based on like specific tasks for robots. It seems very true that it's like, okay, I need to use one model for planning. Like I have a robot that has a task, walk to the laundry room, open the laundry machine and like fold the clothes, open the dryer and fold some clothes. Right. Like the folding clothes thing is always used as like the hard tests for a humanoid robot. [SPEAKER_00] Cause like, no, we, we still don't really have one that can do it generically as crazy as that sounds. Um, so you might have one model that's actually really good for the, like moving my feet, walking stuff. You might have one model that's really good for planning of like, okay, what's the layout of the space? Uh, which direction do I need to go? What path do I need to take? They might have another one that's like for dexterity for like, okay, to open the door, I need like a specific amount of force. [SPEAKER_00] And interesting thing is like, these have different amounts of reasoning versus like amounts of like tool calls or like external actions needed. And a lot of the evals have shown that like segmenting to different models for different subtasks and then putting it together is like oftentimes a lot, giving you much better performance than a single model. And so the interesting thing is like, how do they manage that? [SPEAKER_00] How does, how does the average user who's going to use this robot, you know, how do they figure out, do it? And like these guys basically, a couple of these Chinese companies, like fork termies basically had some modifications for like their, you know, they, they obviously modified it to like connect to their device. They made tool calls for like move hand, move head, whatever, you know, but that to me is a sign of product brand market, product market fit of your brand. Right. Right. [SPEAKER_00] It's like people want to use it. [SPEAKER_01] So for that type of stuff, I'm going to sell this robot company. They're not doing like door to door sales or marketing. [SPEAKER_00] I don't need, I remember I told them and they're like, we didn't even tell this, this person for, uh, so like my point though is like for open source projects, this like community sentiment community. It's something like a brand, like how much people use your stuff is actually very important. Like, you know, I think oftentimes, especially in the current AI market where first of all, half the AR numbers are fake. [SPEAKER_00] Everyone is just so focused on like rushing to tell you monetization stories about companies, but there's a sense in which retention and stickiness is much more, will be much more important. Say if there's like a crash or downturn. And this is, this is one of the things that as we're looking at news is like, do they have this type of like really good brand with their use? And that, that was something we, we found a lot of. [SPEAKER_00] And I think that kind of cemented the, the deal for us of like, you know, it's like, if not now, then when, you know, like, it feels like this is like the rocket ship. [SPEAKER_01] But no, I'm, I'm with you. I mean, true. And maybe just to recap, like you've kind of walked us through why you like Termi so much more than, than the competitors. And then like it's design and the people using it and these really interesting companies, like basically proving PMF and social buy-in and like brand longevity. Right. But I guess maybe my more like pointed question for you is like looking to the next year or two, like where does it go? Right. [SPEAKER_01] Like you have Muse, you have Instinct, you have a lot of these consumer agents. I personally don't feel Nooses is really competing with them or vice versa. I view Nooses or I view Hermes as like kind of what you described, like for pro builders building applications for the long term and really wanting to own it. And maybe like my second order thought is I really feel like people want to use Hermes for two reasons. [SPEAKER_01] And that's because if you own the architecture around Hermes, everything you build, the data, the memory, where you're hosting it, the interactions, everything. That's obviously an asset for you or your business. But also you get like actually unique outcomes, like in terms of like when you ask the model something. So like I feel like there's a lot more to the open source side than people are giving it credit for. But I'm not sure if you agree, given what we've seen. [SPEAKER_00] Yeah. So I'll tell you what, you know, news, you know, has been working towards based on stuff that's said publicly so far at least. [SPEAKER_00] And then I'll tell you what I think where I see the dream kind of going and also where I think the crypto, you know, kind of we can kind of, you know, come back to crypto. You know, so far we've been like crypto people have fucked up. They're on every time on say I stuff and like everything they've done has just like leave crypto behind on this one. [SPEAKER_01] I mean, there is a weekend. [SPEAKER_00] But I think I think there's actually a very real reason to not do it. Interesting. OK, cool. [SPEAKER_01] I'm eager to get to it. [SPEAKER_00] And so. The. You know, I think news has obviously been focusing a lot on. Increasing sort of. Their. Enterprise usage. So like, you know, like the nice thing about her is being open. Open source is like enterprises can. Clone it. Customize it. Use it internally, like add their own privacy layer, add their own compliance layer, add in permissioning. [SPEAKER_00] Connected to their databases. And. News has really been focusing on that. I think they have a bunch of big news coming up on that soon. The. And that's been going really well, I think. You know. [SPEAKER_01] Sorry to interrupt you. But also like. When you have millions of users, you have advocates internally at every company who want to bring this to their whole enterprise. Right. So you already have like the GTM sort of in the bag. [SPEAKER_00] You know. Nothing's in the bag until the contract signed. But yeah. Fair. Fair. [SPEAKER_01] Fair. A true OG. [SPEAKER_00] Yeah. But. But. Like. Yeah. I think. I think that there's certainly that. I mean, they have like. Realistically, they're in the like tens of millions of users. So like they're in terms of like. I think. If I remember correctly, it's like tens of millions of unique GitHub accounts access them daily. So if you use GitHub as your civil resistance outlet. It's not perfect. Like obviously people can have tons of GitHub accounts. But GitHub has made it expensive to have multiple GitHub accounts versus most social media accounts. So. [SPEAKER_00] They have a ton of users coming. From everywhere. I think. Obviously, A. Their subscription model with Portal has been growing. As well as their open router usage. I think they just passed two trillion. As the number one out there. They. Are. Also. You know. Have been working really close to NVIDIA. Like in this round. NVIDIA and Microsoft both participated. I think enterprises are going to access. [SPEAKER_00] News. News is going to be like one of the main open source. Harnesses you can use. At Microsoft. Microsoft. And NVIDIA has installed it on their hardware by default for DGX Sparks machines. So. They've been actually kind of trying to play both parts of the market. Like the consumer. Open source. Closer to the ground model. Which they've obviously excelled at. As well as expanded enterprise. [SPEAKER_00] Now. The. Interesting thing. So like the. You know. The. The. The thesis we had. You know. Kind of. For investing. Beyond the fact that they're growing. And like. They clearly have this kind of like. Developer market. Brand market shit. Is. One thing I. [SPEAKER_00] Got from scraping all this open router data. In the process of doing diligence for like. Three or four months. Was that. I started really analyzing how the open router market itself. Started to work. And it really looked. And I wrote this post about this. Like a month ago. But. It was sort of like. A culmination of work since April or May. And. Basically. I'd analyzed and found. How many? [SPEAKER_00] You know. The way open router works. Is like. There's applications. The applications say. I want access to a model. There are many providers. Almost like. Blockchain. Operators. Or intent. Operators. And blockchains. Who. Who quote a price. That is like. Here's how much I'm going to charge. To run this model. For a million tokens. On input tokens. Cash tokens. And output tokens. And then the router. Matches you to a provider. So. It's like. Node operators. Basically say. Here's how much I'll. Sell you. [SPEAKER_00] Access to. To run this model. And then. You know. Open router does the matching. So you can kind of think of this as like. Open router is a little bit like. A DEX router. Or crypto exchange. The harness or application. Is like a wallet. It stores the user's private data. It stores the user's API keys. In their tool calls. It stores potentially. Private secret keys. Like if they're trading. And then you can view the. Model as a little bit like. A protocol. Like it's like. Open source thing. That has a license. In fact. [SPEAKER_00] A lot of the. Open source models. Now have like. BSL style licenses. Where like. If you make more than. X revenue. You have to. Like give them. A licensing fee. Kimmy has that. I think. Which is. Yeah. No. I think all that. GLM. They all do. GLM also added something similar. Not the same amount. Kimmy's is much lower. Like it's like a very low cap. Of how much you make. But this is very much. Like the crypto. Licenses of like. If you fork. If you fork me. You have to pay some fee. [SPEAKER_00] And then the. Inference provider. Is a little bit like. A node operator. A validator. Or a liquidity provider. They kind of are doing. Both functions. And if we take this analogy. A little more seriously. And you know. The data kind of shows. A lot of things. That look very similar. What phenomena. Can you expect? Well one phenomena. You can expect is. MEV. In crypto. Like people trying to front run. People trying to manipulate. The price. And. Most of my post is about. Finding some examples of that. [SPEAKER_00] And. Just to show that like. Hey. A lot of this. In any open marketplace. Whether it's decentralized or not. It just has to be open. Like open entry. [SPEAKER_01] True. And sorry to interrupt. But what is MEV with. When you're asking for AI inference. Because like in crypto. You're competing to. To make money. Solve a block faster. Fill a transaction. But in AI. Like if I beat you. Yeah sorry. Just want to understand that. No no. [SPEAKER_00] Yeah for sure. So basically. Remember I said there are. Three prices. That an interest writer. Quotes you. There's a. Input token price. There's an output token price. And there's a cash token price. Let's just pretend. The input and output are the same. They oftentimes differ. But like. There. It's a marginal difference usually. But the cash token is usually. Hundreds or thousands of times cheaper. Why? Because like. The inference writer. Is using way less compute. Because like. Cash tokens are like saved. [SPEAKER_00] So they don't have to redo. A whole calculation. And so. An interesting question is like. How does open router. Allocate you to a node provider. They don't just give you the lowest price one. What they do is they take. They give you a. They. They. They give a probability. Of each provider being chosen. Based on the price. That they offer. And then they assign you. Based on that probability. Now. That probability also includes. Some type of like. A quality score. [SPEAKER_00] Like have they had good uptime. Have they. You know delivered. More than X tokens per second. Like. So that part is like. Open router grades them. That's almost like. A credit score for each. Entrance writer. But then. They have a price. So it's like. And so one. Question you might. One thing I was kind of. Looking for. Was like. Are there people who are lying. About how much they're. Giving you in cash tokens. So they. They tell you. There's a really low price. And then they get a bunch. So that increases their probability. [SPEAKER_00] But then they say. They never cash tokens. Because there's no. Cryptographic verification style thing. I can't verify. Like they just say. This is how many tokens. I cached. [SPEAKER_01] And true. And sorry. If they're not providing you. The amount of cash tokens. They're quoting you. Your level of service. Goes down in terms of. Of how. Like the memory. Or your response. [SPEAKER_00] Your level of service. Goes down in the price phase. So I'll give you just a similar. Suppose I have an agentic workflow. That like. Involves. The age. You ask a prompt. Agent asks another prompt. Agent asks another prompt. Agent asks another prompt. So there's five prompts. So. As an example. In GLM 5.2. There was. ZAI. Let's say that they were charging. One dollar. Per million tokens. And the cash tokens. Are like .001 cent. There was a. A provider. And that's done by ZAI. Like the model creator. [SPEAKER_00] Usually runs an inference writer. And. And you can treat the model creators. One is sort of the gold standard. Like they're not going to lie about their caching. Because then everyone is going to be like. Your model shitty. Like it's a brand thing for them. Um. Um. Um. Basically. Uh. You say. Okay. The true optimal. Or like. Close to it. Caching rate. Is like whatever the model provider tells you. Okay. Now suppose there's someone who's saying. I'm going to offer you. [SPEAKER_00] Five cents per million tokens. So 95% discount. Right. Okay. [SPEAKER_00] Uh. And actually imagine. There's now like. 21 queries. So like. You know. I make a prompt. Agent asks another prompt. Agent asks another prompt. That happens 21 times. On the ZAI model. It'll charge me. Imagine that the. The caching is perfect. It'll charge me $1. For the first query. And then the agent's repeated queries. Are like .001 cent. Very small. So it'll cost me basically a dollar. Like under a dollar and one cent. On the other hand. If the person who is. Charging five cents. [SPEAKER_00] Says five cents. But then they lie about caching. Then I get charged five cents. On each query. So now I get charged five cents. 21 times. So I get charged a dollar and five cents. So. So the idea is like. By lying about how much I cash. I can charge five cents. On each query. Whereas if I. The person who's caching. They charge me a dollar at first. But then basically zero. On the repeated rounds. And so. Okay. The question is like. Interesting. But why. Why would I put five cents as my price. To get a higher probability of being chosen. [SPEAKER_00] So it's. It's sort of like a. It's not. Exactly analogous. To like a front running. Or just. Like making too much of the analogy. To MEV is. Is wrong. But it is a manipulation. Like there. [SPEAKER_01] No. No. There's a way you can game it. There's a way you could attract. For inference. To provide a bad service. Your way of providing bad service. Is like. [SPEAKER_00] You're probably caching. But you're not passing on the price benefits of caching. That's the way of looking at it. Like. Yes. Yes. Yes. And the idea is like. You'll lose money on the first query. But then. If the agent asks a ton of subsequent queries. Then like. You. That's where you make money. And so I found statistical evidence. Like. Like I ran these experiments. And like. Tried to make sure. I went through all the different providers. And basically found there were some providers doing this. [SPEAKER_01] So how does this relate back though. To like. I guess. Using an agent. Or the inference side. So what this made me realize. Back to. [SPEAKER_00] The Hermes side. If this. Analogy to this crypto thing. Of like. Wallet. Front end harness. Dex router. Kind of like. Exchange slash. Dex aggregator. Model. Protocol. Inference. Provider. Liquidity provider. If I look at crypto. From 2021 to now. All the money went to either the top. The wallet. Or the harness. Or to the. Validator slash. LP. Right. Like. Everyone in the middle. Kind of got squeezed. Right. Like. [SPEAKER_00] Dex aggregators. Got squeezed. The models. Can't collect as much fees. Right. It's like. FOMO makes more fees than Uniswap. Right. Like why. Right. Like it's like. It's like. The front end. Is able to capture more than the protocol. Right. And so. My point. With this is like. If you think this looks like crypto. And I think any open market. Where people can like join and leave. Like these infrastructure providers can join. I mean. Open router. Is a quality control mechanism. [SPEAKER_00] Like they can like not let some people in. But they're. So in that sense. It's not decentralized. Like I can't just join with my own node. But. Once I'm on there. It's like an open market competition. Right. And. And that. That aspect is very different than. Sure. The centralized. Like open AI. Or. Claude. Where you don't even know who the provider is. Like they might be using CoreWeave. For some of your stuff. They might be using. [SPEAKER_01] People just wake up. And they say. Models been nerfed this week. Or they don't. Like they just dynamically could turn off. So. Yeah. [SPEAKER_00] I think this opens our stuff. Especially in the routing model. Where there's like. Multiple models used. As well as multiple providers. And potentially multiple apps. Where when I say app. I mean like a Hermes. Or an Instinct. Or a. Whatever. I think. That market structure looks like crypto. All the money goes to the top and the bottom. What is the top? The harnesses. What's the bottom? Inference providers. And like. It's kind of borne out. In the last six months. Where like. They are. I think like. Open Router really sold the top. Because like. [SPEAKER_00] I do think the margins are going to get squeezed there. Because like. The router is like. A very competitive market. Like. Like every SaaS company. Is making a router right now. Right. Vercel has a marketplace. Ramp has a marketplace. So. My point is. It's going to be. The router is like. Hard to kind of like. Like. You can be a successful business. You just like. It's. There's low barriers to entry. Right. So it's like. So. I guess my. My. Longshot point here is. From that model. [SPEAKER_00] Hermes is like. Investing in phantom or FOMO. Right. That's sort of. That's sort of the. The game. [SPEAKER_01] I've never heard someone say it like that. It's interesting. [SPEAKER_00] Yeah. And. It's a good way to think about it. I think. A difference between AI and crypto. Is in. To some extent. In crypto. The wallet is like. The final end state. [SPEAKER_02] But. [SPEAKER_00] Actually. The interesting thing about. Something like Hermes. You could have people build instinct. On top of Hermes. And then inherit a lot of the security properties of Hermes. And. My guess. Is actually. People are going to like. View the. App. Harness layer. As like. A security. And brand. Like context management layer. As a people will start building apps. Like an instinct. You know. Like. Like different types of bots. That exist. [SPEAKER_00] On top of a harness. That. Provide you the consumer experience. But then the. Backend development is done. By these harnesses. And so. In that sense. It's not like the end point. Like a wallet. Right. The wallet is an end point. Because it goes to the final end user. Whereas like. I actually. Um. And so I think that. Is going to be. A very interesting. Game to see. The other thing. That's really interesting. Is the harnesses. The inference providers. Are also. Already doing. Kind of like. Payment for order flow. Type of stuff. [SPEAKER_00] Where it's like. If you guarantee me. X billion tokens a day. I'll give you a rebate. Or I'll pay you back. Some of the. See. And so in some sense. This token. Being. Able to aggregate. Token order flow. Is like an extremely valuable. Market position. Just like. In the crypto front end market. [SPEAKER_01] For sure. No that is. That is interesting. So. I mean. Hurries though. Like. News isn't directing. That usage though. Right. It's each individual person. Kind of making their own decision. Where they want to send. When it goes in open router. The inference to. But when they're using portal. [SPEAKER_00] Or when they're using another. Like. The point is like. You can use many different. You could just pick your own provider. Right. Like. You could just be like. Have a contract together. Or something. And send directly. Like. The point is that. These open models. Allow you to unbundle. Codex or. Podcode. Right. Like. When you use. Codex. Codex. You're getting all of these features. Like. Implicitly. They're using some router. Type of thing. Or at least. Model selector. Um. Implicitly. They're using. They're using. Different providers. Right. Like. [SPEAKER_00] You don't know. You know. They're doing all these long-term. Data center contracts. You're getting run on one of them. But you don't know which one you're getting run on. And I'm sure you're getting migrated. It's not like. Why would they lock you to one particular provider? It doesn't make sense for them either. [SPEAKER_01] Turin. I know we have a hard stop in two minutes. So I just want to ask you in closing like. Hard question. But like. What do you think the world looks like with Hermes in a year? You mentioned enterprise. You mentioned retail. It's a lot of technical stuff. Like. I don't know. Like. How do you envision like. What the world's like. How people are interacting with it. How they're using it. Kind of the year from now. I think. [SPEAKER_00] There are a lot of people who are building. Who have existing apps. And are adding in agentic functionality. Like. Robinhood. You know. You've tweeted a lot about the Robinhood agentic experience. Not being as good. Maybe it's. Maybe it. Maybe it's changed after yesterday or whatever. Yeah. My belief is that. Every SaaS company. Either has to. Have a harness. Or be a router. Because they need to get into the token order flow. Otherwise they're like. They're kind of like. They're kind of fucked. [SPEAKER_00] If they're not in the token order flow game. Because like. That's. That's the easiest thing to monetize. And. In that world. Hermes is the white label harness for anyone. Right. Like. Every company that. No one wants to start from zero. You have a security. And. And brand risk thing. Like. If your harness has been hacked. No one's going to use it. A lot like a wallet. And so. I think like. Being able to have this really strong brand. It means that other people will want to. [SPEAKER_00] To use you as their back end. Beyond the prosumer. And I think. That's. That's. That's kind of where I. I see that. The growth. And like. The enterprise usage is one. The other might be like. Consumer usage. The other might be this kind of like. Apps that build on top of Hermes. Because like. They don't want to start from scratch. [SPEAKER_01] So true. And if every company becomes a harness. And Heres is the best harness to build on. Why does every company want to become a harness? Or. Yeah. [SPEAKER_00] Why do they all need. Bring users on data. They get a rev share from someone who's managing. The token order flow for them. It's really like. A fintech in a bank. Right. A fintech. Most fintechs don't have banking licenses. So they just do rev shares. With bank underneath. [SPEAKER_01] Love it. It's true. And you took us down such a. A unique. And like. Differentiated. I think this is where it like. A lot of people just kind of. Yeah. And. And another. No it's interesting. Because a lot of people just talk about. Like what they're building. [SPEAKER_00] Which is why I think this crypto thing is true. Muse. Is going to monetize. In a very crypto way. It's charging transaction fees. On stuff you buy in the. App. And it's basically doing that. By doing rev shares. With Stripe. And. Shopify. And whatever. How is that any different. Than this world I just presented. Not really. Right. So. So. I think a lot of the ways. Order flow management. Is done in crypto. In any open system. Where there's competing. Counterparts. Will end up looking very similar. [SPEAKER_00] To what crypto has. And. By trying to take that. The lessons you've learned from there. You can make some pretty. Hopefully useful. Hypotheses. [SPEAKER_01] I love it. Turin. Thank you for. For coming on. I mean. Huge congrats on leading this news round. We're co-investors with you. We're so excited. I'm also just excited for humanity. Because this is like. Open source. And where I want the world. To go. And to be. And your comments. Are just so unique. People come on. And they talk about like. What they're building. Or what they want to do. Like. You brought us down. The eras of crypto. Like. Why this is so interesting. Like. The differentiated build. On the dev side. The app side. Like. The relationship to crypto. And like. [SPEAKER_01] How it could be related. So. This is really unique. Thanks so much. I appreciate your time.