# Aidan Gomez Unfiltered on AGI, China & AI's Biggest Risk

Markdown copy of https://realalpha.app/v/GfBEBfbW6YE/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Aidan Gomez argues that AI's biggest risk is dependence on a few providers, while enterprise adoption still has room to expand.

- Show: Newcomer Pod
- Hosts: Eric Newcomer
- Guest: [Aidan Gomez](https://realalpha.app/builders/aidan-gomez/) is a Cohere co-founder and Transformer co-author.
- Published: Jun 26, 2026
- Length: 22 min

## The best minutes

- [11:35](https://realalpha.app/v/GfBEBfbW6YE/?t=695) [Democratic countries need sovereign AI for resilience](https://realalpha.app/s/democratic-countries-need-sovereign-ai-for-resilience/) (3 min)
- [3:18](https://realalpha.app/v/GfBEBfbW6YE/?t=198) [Enterprise AI use can grow far beyond coding](https://realalpha.app/s/enterprise-ai-use-can-grow-far-beyond-coding/) (3 min)
- [18:25](https://realalpha.app/v/GfBEBfbW6YE/?t=1105) [AI agents may displace workers more than past technologies](https://realalpha.app/s/ai-agents-may-displace-workers-more-than-past-technologies/) (3 min)

## Also worth watching

- [6:08](https://realalpha.app/v/GfBEBfbW6YE/?t=368) [Self-hosting open models can cost more than it appears](https://realalpha.app/s/self-hosting-open-models-can-cost-more-than-it-appears/) (4 min)
- [15:15](https://realalpha.app/v/GfBEBfbW6YE/?t=915) [Chinese models are progressing beyond distillation](https://realalpha.app/s/chinese-models-are-progressing-beyond-distillation/) (3 min)
- [0:32](https://realalpha.app/v/GfBEBfbW6YE/?t=32) [Transformers persist by absorbing ideas from other architectures](https://realalpha.app/s/transformers-persist-by-absorbing-ideas-from-other/) (3 min)

Skip the opening teaser and the closing subscribe message.

## Summary

Aidan Gomez is a Cohere co-founder and co-author of the Transformer paper Attention Is All You Need. He says Cohere builds its own models and offers customers customization and deployment on their own infrastructure.

Aidan Gomez argues that dependence on a small number of AI providers creates a resilience risk for democracies, especially when critical infrastructure relies on services that can be switched off. He says current models qualify as AGI in many respects, while enterprise adoption remains early and model costs should fall as efficiency improves. He expects custom, sovereign deployments to matter more than competing only at the frontier, and says Chinese models are already highly capable even if the US remains ahead. He expects agents to move beyond software development but says their effect on employment remains uncertain.

### Minute by minute

- [0:00](https://realalpha.app/v/GfBEBfbW6YE/?t=0) **Too few providers create resilience risks.** Gomez warns that a small pool of large technology companies is becoming a single point of failure for democracies. He says computer engineers are taught to avoid that kind of system.
- [0:32](https://realalpha.app/v/GfBEBfbW6YE/?t=32) **The Transformer remains the platform.** Gomez says the Transformer was built for a narrow translation task, and he is surprised it remains the foundation nearly 10 years later. He says it absorbed ideas from SSMs and expects innovation to continue above the architecture. Asked if we have AGI, he says yes in many respects because models can exceed human capabilities on many problems.
- [3:18](https://realalpha.app/v/GfBEBfbW6YE/?t=198) **Enterprise AI has room to grow.** Gomez describes an adoption cycle: companies first encourage unrestricted use, then constrain budgets, while model makers improve efficiency. He expects costs to fall and lower costs to unlock more use; he estimates another 10 or 100x in enterprise token consumption beyond coding. In finance, he sees potential for automated decisions, communication, coordination, regulatory and compliance work.
- [6:08](https://realalpha.app/v/GfBEBfbW6YE/?t=368) **Open source can cost more to run.** Gomez says open-source models may look cheaper by token price, but self-hosting adds infrastructure, software maintenance and the burden of updating the system as models change. He says customers also need to consider whether they are comfortable sending data to those providers, and compares operating an open-source model to self-hosting a database.
- [8:03](https://realalpha.app/v/GfBEBfbW6YE/?t=483) **Tailored models can beat frontier scale.** Gomez says Cohere builds its own models because customers want lower costs, sovereignty and customization, and the market—not model capability—is often the bottleneck. He says he is not trying to compete at the frontier. He estimates models are good enough for 90 percent of what enterprises want to do, while the remaining 10 percent remains difficult and valuable.
- [11:35](https://realalpha.app/v/GfBEBfbW6YE/?t=695) **Critical workloads need sovereign infrastructure.** Gomez says democratic countries have consolidated technology purchases among a small group of providers, which creates a single point of failure. For critical workloads, he argues countries should control the infrastructure they depend on. He says Cohere's planned Toronto-Berlin dual headquarters and its partnership with Aleph Alpha are part of building capability across democracies, not just two countries.
- [15:15](https://realalpha.app/v/GfBEBfbW6YE/?t=915) **Access restrictions expose foreign dependence.** Discussing recent US restrictions on AI access, Gomez says the episode was the scenario Cohere had warned about. He argues that the US is pursuing its own interests, but other countries could lose access if they build critical infrastructure on services they do not control.
- [16:57](https://realalpha.app/v/GfBEBfbW6YE/?t=1017) **Gomez sees Chinese models as highly capable.** Gomez expects the US to continue leading in capability, but says Chinese models are extraordinarily capable and well-resourced by their government. He says distillation may have played a part, but Chinese companies are increasingly developing models independently and are on a strong trajectory.
- [18:25](https://realalpha.app/v/GfBEBfbW6YE/?t=1105) **Agents could reshape work and employment.** Gomez expects agents to move from software development into other enterprise fields, where developers already work with models as collaborators. He is unsure whether rising demand will offset displacement and says society should plan for that possibility, with options including retraining, UBI and jobs-guarantee programs.

## Transcript

Newcomer Pod, hosted by Eric Newcomer, with Aidan Gomez.

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[0:00](https://realalpha.app/v/GfBEBfbW6YE/?t=0) A very small pool of these large tech players are becoming a single point of failure for the entire democratic block. That is not a resilient system, right? That is a single point of failure. And the first thing you learn as a computer scientist or engineer is don't build a single point of failure. Are we at AGI today? I mean, in many respects, yes. In reality, I think it's actually quite expensive to adopt open source models. I'm Eric Newcomer, author of the Newcomer Substack. Let's get into it. your co-author on attention is all you need take stock of where we are in this moment with foundation models what we're capable of and and then i'll sort of follow up with where you think we're going but are you surprised from the moment of writing that paper to where we are today that we've gotten as far as we have in in such a short time i don't think anyone on the transformer paper had any idea what was coming we built it for translate like Google translate so it's a very scoped small problem to be solving and it's a huge shock I think an even bigger shock for me given that it's been nearly 10 years now since that paper is the fact that we're still using the transformer you know like as a scientist you hope you build something but then you're not

[1:24](https://realalpha.app/v/GfBEBfbW6YE/?t=84) married to it. You know, you want someone to come and build something on top of that that is much better, blows it out of the water, and that is progress. There has been a ton of progress, but it's hilarious how the transformer has been such a great artist in its ability to steal the good ideas from every other architecture. Like when SSMs took off, this was like an architecture that would let you have a massively larger context window. Everyone was saying the Transformer is dead, we're going to shift over to SSMs,

[1:58](https://realalpha.app/v/GfBEBfbW6YE/?t=118) **Aidan Gomez:** and then the Transformer just copied those ideas, integrated it into the architecture, and we still call it the Transformer.

[2:04](https://realalpha.app/v/GfBEBfbW6YE/?t=124) So I don't know if we'll ever move on. Yeah. So how much more room does it have to run? How much does the next great leap depend on a new idea? I mean, we had, I guess, reasoning models were a big sort of leap outside of that mode or how do you think about what it's going to take for models to continue to make giant leaps over the next couple years yeah it's it's exactly that it's like reasoning is still running inside the platform of a transformer and so i think people move up the stack and that's where they invest their energy in innovating so in the same way that like the fundamental training algorithms that train models haven't changed in much longer than 10 years.

[2:44](https://realalpha.app/v/GfBEBfbW6YE/?t=164) They're still the same thing. It might be the case that the transformer is kind of just the platform. And we're going to continue to build on top of that for a very long time. And it would take something pretty monumental to shift us past it. But I still hope there's room for progress. Are we at AGI today? I mean, in many respects, yes, right? Like, I think definitely. It's a general artificial intelligence. And it seems like whatever problem we point it towards, we can basically exceed human capabilities in that problem. So token maxing has become the buzzword of the moment, but it reflects a real issue, which is companies wanted to incentivize their employees to use AI and get the most out of it.

[3:33](https://realalpha.app/v/GfBEBfbW6YE/?t=213) So they said, run wild, we'll have leaderboards, we'll track how much you're spending. Now, you know, I think we've seen people like Uber and some other companies say, oh, maybe we should track what's going on because we're going to blow through our budgets pretty quickly. You know, you work with a bunch of enterprises. What's your read on is, you know, are the inside tech companies ahead of this trend? And there are lots of old guard companies that still need to have the token maxing fail. before they pull back, or where are we in the arc of sort of spending for spending's sake on using AI? Well, I think a lot of that exuberance came out of coding and those models.

[4:12](https://realalpha.app/v/GfBEBfbW6YE/?t=252) You see this typically in like technology cycles. There's a exuberant adoption phase where the CFO or the CEO tells the CFO, no spend caps, like just adopt, adopt, adopt, and it goes into excess. and then there's a correction period. And that comes from the enterprise itself constraining spend, but then also from innovation and compression and efficiency within the modeling companies. So I think we just did this big ramp up to a much larger tier of model. You'll now see one of these cycles of compressing back down. So costs will fall dramatically. Is that going to hurt your revenue? I mean, if their customers pull back, there's enough growth elsewhere?

[4:54](https://realalpha.app/v/GfBEBfbW6YE/?t=294) No, yeah. Yeah, like the usage growth, as the costs come down, it will just unlock way more applications. We're still so early in enterprise adoption of AI. Yes, coding has taken off, and that might be hitting some sort of saturation point. But everything else within the enterprise is, from my perspective, still essentially untouched. Like the stuff that we're doing is so simple. There's so much, there's like an iceberg, you know, underneath the surface to go do.

[5:26](https://realalpha.app/v/GfBEBfbW6YE/?t=326) **Aidan Gomez:** And so there's probably another 10 or 100x in terms of token consumption just on the enterprise side.

[5:33](https://realalpha.app/v/GfBEBfbW6YE/?t=333) After coding, what's two and three or what are the next buckets? Big applications. It really depends on the industry. Like that's, it's specific to the industry. So in finance, you'll see much more automated decision-making, communication, coordination between the different business units, automation of regulatory and compliance. There's so much work that gets done inside these organizations beyond just writing software. Actually, the overwhelming bulk of what's done is not writing software inside these legacy organizations. And all of that is going to be done by machines. I want to talk about the response or one answer to token maxing, which is open source or the idea that people move away from state of the art models to find, you know, I think there's a lot of conversation about routing to various models in a cost effective way.

[6:28](https://realalpha.app/v/GfBEBfbW6YE/?t=388) What is your view, I guess, putting aside the geopolitical part of it, which we'll get into, what's your view on sort of the appeal of open source models from a sort of cost structure point of view? I mean, in some ways, they're cheaper if you're willing to build up all of the infrastructure around them, like all of the software stack, the harness, maintain this thing, believe that you have the best serving efficiency.

[6:52](https://realalpha.app/v/GfBEBfbW6YE/?t=412) **Aidan Gomez:** see. In reality, I think it's actually quite expensive to adopt open source models, and it looks cheap. If you're willing to use, like, DeepSeek's API or Alibaba's API, you see the

[7:07](https://realalpha.app/v/GfBEBfbW6YE/?t=427) token prices, and you're like, oh, look, open source is cheap. These companies release open source models, and their prices are way lower than the American equivalent. What's wrong with that thinking? What's wrong with that thinking? Well, one, you have to be comfortable sending your data to those parties. If you try to self-host, I think it's a very different equation. And then there's also the maintenance burden of that, right? Like these harnesses, these pieces of software that the models power, you're going to have to keep that updated over time as capabilities evolve, new models come out.

[7:41](https://realalpha.app/v/GfBEBfbW6YE/?t=461) And so that's another huge expense. I think similar to databases, in the same way, yes, you could take an open source database, self-host that thing, try to maintain and manage it. Nobody does that. Like very, very few people do that. Instead, it's all managed services. There are enormous databases. companies because it's a hard, painful, and expensive thing to operate. So part of the pitch of coming to Cohere is that you don't want to manage your own service, models are going to keep changing, we'll be here with you, make sure it's safe. How do you think about why are your models shifting to being more open source in the first place? And then why Cohere build models versus sort of being a facilitator of your own models and other models? So on the first question, like the, yeah, I think the piece around which model you use matters a lot. And for Cohere, it's very strategic for us to build our own models because we offer more leverage to the customer. Increasingly, what we're seeing is folks care a lot about cost. They care a lot about sovereignty and off the shelf models aren't good enough or they're too expensive for what they need to do. So the ability to customize becomes an essential part of the product and offering to our customers. The other thing is that we're not

[9:09](https://realalpha.app/v/GfBEBfbW6YE/?t=549) trying to compete at the very frontier. We're not trying to get into a like who can spend the most money competition. It's just a losing race. And increasingly what we see, the model is not the bottleneck. It's actually the market that's the bottleneck. So frontier models are so far out in front of what the market is doing that even if you have a model from a year ago, 18 months ago, two years ago, the companies are still trying to catch up to that. So the model itself is largely not the bottleneck. And when it is, they can't afford these massive models. And so they're looking for a customized, bespoke model. I interviewed Ali Goetze at Databricks a week or so ago for the newsletter. He told a story about how he got in the model business. They said they spent $10 million on a training run. Really, the whole cluster cost them $20 million.

[10:00](https://realalpha.app/v/GfBEBfbW6YE/?t=600) And then three months later, the whole thing they'd built wasn't worth anything. And his view is basically that unlike software, which slowly loses its value, even if you have out-of-date software, your customers take a while to figure that out. Models, they switch to the next thing. It sounds like you're offering a view that older models still have a lot of value like years out or you disagree with that point of view. What keeps customers using sort of a year old model? Well, when was DBRX trained? It was like two years ago. Yeah, this was 2024. So he left this a long time ago. This was sort of a reflection.

[10:35](https://realalpha.app/v/GfBEBfbW6YE/?t=635) So when you're on this like super steep part of the capability curve and like when you're in a period where models aren't very good at anything, yeah, you'll feel that, right? because it's increasing super fast but the reality is models have sort of gotten good enough at so like 90 percent of what an enterprise wants to do that last 10 is still super valuable i'm not trying to discount that and there's like uh i don't know a decade plus of technical work to go do to solve that that 10 but the idea that um models change every three months and it's like some huge groundbreaking thing it's just not true anymore for the vast majority of what the market wants to do it is true at the very fringe right the frontier like in the pure sciences yes you can have a material breakthrough where the model from six months ago couldn't do it and the model now can like solving air dose problems or what you know protein folding whatever but a lot of the thesis there's a lot of value today to be and people just need to figure out how to extract it out of the model. Yeah, and much more, the business is not the model itself. It's actually everything around it. This idea of sovereign AI, you know, you had this merger with Aleph Alpha,

[11:48](https://realalpha.app/v/GfBEBfbW6YE/?t=708) I think what they, they're like 10%, you're sort of 90% of the business, German business, like, what is the idea of sovereign AI? And what countries are you sovereign to? And how do you think about sort of, yeah, nations? needing to prefer companies that have some sort of affinity with them or local presence. Yeah, well, there's been this over the past 30 years. The world, and in particular, the part I care about, like the democratic alliance, the G7 or, you know, whatever you want to call it, it has basically consolidated to buying all of its technology from a very small set of players. And that is a really non-robust system.

[12:34](https://realalpha.app/v/GfBEBfbW6YE/?t=754) It's not very resilient. It's a single point of failure. And like the first thing you learn... America is the single point of failure? Yeah. I say this is an American that wants to be on team democracy and worries about it. So I'm not saying that's a crazy point of view, but just America is sort of... A very small pool of these large tech players are becoming a single point of failure for the entire democratic bloc. And that is not a resilient system, right? That is a single point of failure.

[12:58](https://realalpha.app/v/GfBEBfbW6YE/?t=778) **Aidan Gomez:** And the first thing you learn as a computer scientist or engineer is don't build a single point of failure.

[13:03](https://realalpha.app/v/GfBEBfbW6YE/?t=783) So if you care about the democratic project, capability diffusion into the other G7 countries or into the broader democratic alliance is essential to resilience of democracy itself. If one goes down, you need to make sure you have backups. And I think you've started to see a much greater appreciation of how important that is, how important digital sovereignty is, it was something that was taken for granted, right? Like, it'll always be here, you know, I'll always be able to rely on it. But with the export ban of Anthropix models, you saw that it can get pulled away from you. And so our, like, Cohere's pitch has always been deploying on the infrastructure of our customers, instead of saying, hey, send, you know, your data to me via an API, which I can switch off, and then you lose access. I think that has to be the model that's pursued for critical workloads, for other stuff, for startups, whatever, like use APIs, they're great. But for critical workloads, countries need to control the infrastructure that powers their economy, like telcos, like financial institutions, healthcare, the grid, water treatment, defense. What countries does Cohere call home?

[14:21](https://realalpha.app/v/GfBEBfbW6YE/?t=861) Well, like with the merger with Aleph Alpha, we'll have a dual HQ between Toronto and Berlin. So really, you know, once that deal is complete, it'll be Canada and Germany. I call Britain home. My mom's British. I live here. And that partnership between Canada and Germany, it was never meant to be just two countries. We want to build this alliance to help create capability beyond just one democracy, right? Like the thing that I said at the G7 was, it is the case we have a democracy leading in AI right now. That's fantastic. But democracies need to occupy position one, two, three, four, et cetera, and that's not true.

[15:07](https://realalpha.app/v/GfBEBfbW6YE/?t=907) And so I think it's essential that middle powers, or whatever you want to call it, come together to develop these capabilities for the good of the democratic project. What is your gloss of, yeah, the Trump anthropic back and forth? I mean, first, there was, you know, yeah, the anthropic trying to limit how much they could be used in wartime situations, and now it's this sudden blockage of fable. I don't know. What did you take away from those two situations? Well, it wasn't super surprising. This is sort of exactly the scenario that we have been warning against. And I think that anyone who was naive to the potential risk of something like this happening were woken up quite suddenly.

[15:58](https://realalpha.app/v/GfBEBfbW6YE/?t=958) I think that... What does it mean? I mean, like a big issue here is that the way the Trump administration went about it was focused on letting anyone outside the U.S. use it, because that sort of woke up the other democracies that there would be sort of a preference for the U.S. rather than a preference for the U.S. and its allies. Is that sort of the right depiction of your view? Yeah, the U.S. is optimizing for its interests, as it should. But I think in doing so, it does present a huge risk for anyone outside of the U.S. Like if you built your critical infrastructure on top of some service that you depend on to live or to operate or to have an economy at all, you now lose access to that.

[16:54](https://realalpha.app/v/GfBEBfbW6YE/?t=1014) So we're talking about, you know, the weakness of relying on one democracy for the rest of the democratic world. What about the other part of this, which is, what's your read on how strong China is in sort of the foundation model of business right now? We see DeepSeek raising this enormous round. Like, do you think there's a real risk that China leapfrogs the Western world in sort of model development? Leapfrogs? I would be very surprised. I do think the U.S. is going to continue to lead in capability. I think the dismissals of Chinese models as, like, just distillations or anything like that is cope.

[17:34](https://realalpha.app/v/GfBEBfbW6YE/?t=1054) **Aidan Gomez:** Like, I think they are extraordinarily capable, extraordinarily well-resourced by their government, and only getting more resources.

[17:45](https://realalpha.app/v/GfBEBfbW6YE/?t=1065) And they're just very good at building this tech. I'm sure distillation has played a part, but increasingly independent of distillation, I think they're on a very strong trajectory, which just drives more urgency to make sure, you know, players like Cohere, alliances form around them, get supported so that we have more than one democracy competing against China for the global market. Because very quickly, you could see the world split into two. But what you really want is a democratic alliance of two, three, four, five against an autocratic singular player. Give us, for the last question, a sort of view of what you think the next 12 to 18 months looks like in terms of technological capability.

[18:38](https://realalpha.app/v/GfBEBfbW6YE/?t=1118) Is it about agents? Yeah, I guess that's my answer. But if you agree, then sort of shape that up a little bit because we hear that word so often. What's that mean? Or what do you think is going to be possible certainly in the next 12 to 18 months? The shift is it's going to be entirely in the enterprise, the way we do work, who we work with, who we communicate with. Increasingly, it's going to shift towards working with models. And it's already happened in this kind of microcosm of developers. Like, they spend most of their day interacting with a model to do their work. That's their collaborator that they're partnering with, and they, you know, manage it.

[19:17](https://realalpha.app/v/GfBEBfbW6YE/?t=1157) They tell it to go do some work, comes back with a bunch of code, they give you some feedback. No, I actually meant like this. Can you fix this? Increasingly, we work as colleagues with agents in software development. That will start to permeate out into many more fields. And if you look at, I mean, these are early numbers, but if you look at employment numbers, inside of software engineering, it's actually still very healthy. And so the big thing that I'm watching, concerned about, but I do remain optimistic about, is how this impacts employment and how that shapes public sentiment towards the technology.

[19:55](https://realalpha.app/v/GfBEBfbW6YE/?t=1195) I'm hopeful that, you know, whatever it's called, Jevons Paradox, where... A massively efficiency-augmenting technology emerges, you would think that would displace tons of people, but instead, demand rises... Way more productive, we need more of them. Demand rises so quickly that you actually need more people to manage these machines. But your view is you don't know which way it'll go yet. I don't know, because the counter-argument to Jevon's paradox is that with previous technological revolutions, those tools, those machines that came out, they were so far from... like a loom. They're so far from a human.

[20:30](https://realalpha.app/v/GfBEBfbW6YE/?t=1230) They're so different than us. They were so specific and narrow in terms of what they could be applied to.

[20:36](https://realalpha.app/v/GfBEBfbW6YE/?t=1236) **Aidan Gomez:** We have something now that looks so much more like us that maybe it is more displacing.

[20:45](https://realalpha.app/v/GfBEBfbW6YE/?t=1245) I'm hopeful that's not the case, but I do think we need to plan for that scenario and come up with the tools to address that. Things like maybe UBI, but more likely retraining programs, potentially even right to work or, you know, jobs guarantee programs. Aiden, we're at time. Thank you so much for standing up for the democratic world. I appreciate it. That's our episode. This is the Newcomer Podcast where we chop it up on all things Silicon Valley. Thanks for watching. Please like, comment, subscribe. Go read the Newcomer Substack if you want to get really in the weeds and understand the real business of venture capital and startups at newcomer.co.

[21:24](https://realalpha.app/v/GfBEBfbW6YE/?t=1284) You still yearn for podcasts. I've got the Cerebral Valley Show with Max Schout and James Wilsterman. Otherwise, probably see you in about a week. Thanks for all your support.