AI Labs Will Trap You or Replace You - Dillon Rolnick, Nous Research

AI Engineer · 15 min · 152 sentences · from YouTube's caption track

Each timecode opens YouTube at the start of that sentence. Line anchors (#s42) are the cue ids in the WebVTT, and every line carries its start and end seconds. All transcripts has every talk, and the whole corpus as one file.

00:01[music]
00:12I'm Dylan Rolnick.
00:13I'm the CEO of News Research.
00:16And today we're going to talk about building on foundations and platforms that you can actually own.
00:29Whoop, there we go.
00:30All right.
00:32How do you make back a cent a billion-dollar infrastructure and research bet?
00:37Labs individually have spent hundreds of billions of dollars training and building ever more sophisticated intelligence system and the industry as a whole has spent trillions, perhaps.
00:56The problem is though that the models have begun to commoditize.
01:02You guys in this room who are hopefully practitioners of AI have known this intuitively.
01:07You switch among models readily.
01:09You switch between Astra, Fable, Opus, Kimmy, whatever it may be without too much regard for who's on the back end or what it may be.
01:19You want to try the new model.
01:21We've seen this in data, too.
01:22There was some charts going around Twitter this week showing the average cost of a token going down while the total spent on tokens surges and the infrastructure costs associated
01:35with serving those tokens has stayed flat.
01:38Meaning where there used to be margin in a very, very performant model, it's now starting to narrow as we see competition and you know, a baseline level of intelligence be available to people.
01:53Okay, so if the models are starting to commoditize and you've spent hundreds of billions, trillions of dollars on them, you need to make that back.
02:03And you need to find a different way to do it.
02:06And in my opinion, there are two pretty obvious ways to do it.
02:11One, you find a different layer of the stack to compete at.
02:14You go down to the infrastructure, or you go up to the user.
02:19Or two, you build God, and you build God faster than anyone else, and you monetize God.
02:27If you're in front of the user, you have an opportunity for stickiness.
02:30Stickiness means margin.
02:32This is why the fight has largely shifted to the agent layer lately.
02:36We started talking, I heard somebody previously talk about 5% of, you know, a workforce engaging in agents.
02:42This is becoming more and more the zeitgeist of where the AI battle is taking place.
02:48Consumers, though,
02:50[sighs]
02:50are fickle, and they're terrible customers.
02:53Businesses, on the other hand, spend tons of money on tokens, and they have economically productive use cases for those tokens.
03:02So, if you're going to be competing at the user layer, and you're going to be bringing this fight somewhere else, you're going to be bringing it to businesses.
03:11The best thing to do is to get ingrained in a business, cuz when you're in a business, it's hard to take you out, right?
03:18And if you think about your business 2 3 years into the future, how much of your workforce, as measured by its output, is agentic?
03:2970% conservatively, probably more.
03:34And if 70% of your workforce is organized by somebody else who may or may not have adverse incentives to you, they can be organized into a strike.
03:43And so, what you would do if you had to make back a trillion-dollar investment, was you would embed, you would use your asymmetric leverage, and you would change the rules of the game.
03:55Or, like I said, you would build God.
04:00If you succeed at building God,
04:03[sighs]
04:04what you would do is you would point God at the most lucrative problems in the world.
04:10The best shot at earning it back is not to sell it for an ever-decreasing margin to others to solve these hard and lucrative problems.
04:17The best shot is to solve them yourselves.
04:28Labs think that markets are solvable problems.
04:33And that's exactly how they view the financial industry as a problem to be solved.
04:41And a tractable problem at that.
04:43Perhaps not even a very hard one.
04:46Right now though, they know, at least the smart ones know, that they don't have enough context, know-how, or experience in the financial industry to be able to do that themselves.
04:56So, they will partner.
04:58They hire investment banking analysts.
05:00They partner with some of the biggest firms on the globe.
05:04And they start to build their foundation here.
05:09My belief is that they will eventually go to the source because, like I said, that's where the money is, and we're trying to make back a trillion dollars.
05:18And so, you might say, "This is nothing new.
05:21We have always been trying to trap people into these games and change the margin, and once we have you and you're stuck, we are able to raise the prices."
05:28But, the difference is that the stakes are much higher.
05:31We're talking about meaningful deca-percentages of the Earth's economy.
05:38And you in this room need AI.
05:44Everybody has seen what they can do, what these models can do, what agentic systems can do, and there's no going back.
05:51People in this room who are um making decisions, vice presidents, and executives, and engineers about what they're building on, are going to have to make decisions now about where they grow roots around these tools and processes that will affect the future of these businesses.
06:10And whatever you pick, you'll start to harden.
06:14You'll gain know-how.
06:15It'll stop up the intelligence of the organiza- organization itself.
06:19You will start to build processes and systems around it.
06:22I implore you, do not build on quicksand found- quicksand foundations.
06:27Build on open foundations.
06:29Build on Hermes Agent.
06:33So, I just spent a fair amount of time convincing you, hopefully, that the highest expected value move in AI right now is to either trap you and jack up the prices,
06:45or replace you directly.
06:48So, why am I standing here?
06:51Uh we view the world very differently.
06:55And the first thing I hope you guys will learn about me and about News Research is that we built this company primarily for moral and philosophical reasons.
07:05And a lot of times in business, morality and incentives don't always align, and when that happens, it's probably easiest to change your morality.
07:13But, we think the world is changing, and we think it's different.
07:16And for that reason, we think that the best course of action is one in which we are both aligned.
07:25When we built the company originally, we built it to combat a seemingly inexorable trend towards centralization and centralized control of perhaps the world's most most important technology and most important invention.
07:37News Research was built by and staffed by nerds.
07:41Nerds who love science fiction.
07:43And we've all seen what happens when a centralized entity gets control of god-like intelligence and uses that for exploitation.
07:50Skynet was, after all, a closed-source company.
07:57The next thing that you need to know is that there's no longer really a strong moat in proprietary software.
08:04The marginal cost to reproduce a piece of software has plummeted.
08:08A few months ago, uh xAI released Grok bots and we saw that and we thought, "Wow, that's an intuitive user interface.
08:16That's really nice."
08:17So, we pointed Hermes agent at it and we reproduced it in a weekend and released it as a feature within Hermes agent, open source, of course.
08:27And the game has changed.
08:29It used to be that software was difficult to produce and it took coordinated teams of engineers long hours to pilfer open source and take libraries and combine them together into proprietary pieces of software and then sell that software as SaaS.
08:44But, now the opposite is true.
08:47Anything can be reverse engineered and there are many, many motivated open source builders pointing their agents at these different softwares.
08:56If the close competitors do manage to make interesting features, open source will have it not long afterwards.
09:03The motes of yesterday were in proprietary IP.
09:06The motes of tomorrow are in network effects, reputation, and reliability.
09:12What's funny is that in a lot of ways the labs have always seemed to underestimate their own tools or perhaps they are underestimating the users of those tools and their genius and their motivation.
09:25This happened and is happening on the model side and we'll talk about why that's important, but it turns out that models can fairly easily teach another model to be, you know,
09:37nearly as smart as the original model.
09:40And the advantage in a model is very rarely in the architecture or some novel system that was created in a proprietary way for that model.
09:48In fact, it's usually in the quality of the data and the scale.
09:54And sorry.
10:05The uh
10:13The The model started to The model company started to realize that the you know, scale was catching up and that you could create these things with supplementary quality data
10:26by teaching one model from another.
10:29So, they labeled the teaching process distillation and they labeled illicit distillation theft.
10:37You cannot call that copyright infringement because copyright infringement would imply that you somehow took every living human being's written or creative work and then trained on top of it.
10:48They also labeled the process of They also trained the models to not assist in the process of training other models and they labeled that safety.
11:06The next thing that you need to know about agents and about the new world is that agents are like operating systems and they're like frameworks and they're a lot more like those than they are like chatbots.
11:19And so, if you continue to think about your agent as merely a personal assistant you will forever live on the surface of what it means to engage in agentic systems.
11:28And so, what do I mean by that?
11:31If you guys were here a little bit earlier, there was a presentation about a financial firm and how they think about AI and these different inputs that go into
11:39ultimately what becomes an investment decision.
11:42Agents are really at their heart a system that combines determinism and non-determinism.
11:48If you remember from that talk, there are data sources, there are routine code, there are people's opinions and there's judgment.
11:56And all these things come together to create what is the new application.
12:01The new application is no longer necessary that it is deterministic and solid and cold hard code.
12:07It can inject intelligence, reasoning, judgment, all of these things together to combine to create the new application.
12:17And so, as you at your company think about building these new applications, you should be building these systems on foundations that you own.
12:29You should be able to swap in and out the models that you want.
12:32You should be creating ever more sophisticated towers of functionality on top of these core primitives of determinism, non-determinism, code, data sources, document, reasoning, intellect.
12:50But, you should not be spending your entire lives being intelligence companies.
12:56What has made you successful this far?
12:59Most of you are most likely not agent companies.
13:01Most of you are not likely AI companies.
13:04You should be focused on the things that you do well.
13:07And when the state of the art changes and it evolves and what it means to be an agentic system starts to shift, you should inherit that, not rebuild it and maintain it.
13:17Remember, these are tools that should make you better.
13:19They are not drop-ins to replace you.
13:21Do you know how much energy it takes to create a human?
13:27And the beauty of the Hermes Agent proposition is that you can run it however you want, wherever you want, modify it as you see fit.
13:35If you're not having fun anymore, you can take your ball and you can go home.
13:39The moat has changed.
13:41The moat is no longer entrapping you.
13:44The moat is in reputation.
13:47And that's what Hermes Agent offers you.
13:49And so, what do I want out of all of this?
13:54Well, I want to build a standard.
13:56Hermes Agent today is the most used and downloaded open source agent in the world.
14:01It is order of of magnitude more downloaded than some of its close competitors.
14:07And I want everybody to be able to build on these open foundations.
14:11Because as long as one of us us is willing to open source these tools and these agentic systems, water will continue to flow downhill.
14:20And I want to be downhill so that I can work with you to provide support, safety, scalability to go along with the world's best open agent.
14:28I've seen where this goes and we've seen where the motes will lie, network effects, reliability, and reputation.
14:37People are going to continue to solve hard problems with agentic systems.
14:42People are going to solve lucrative problems with agentic systems.
14:46Somebody is going to earn back the trillion-dollar investments that we've made.
14:50Make sure that it's you.
14:52Make sure that you are in control of your own destiny.
14:55You can have your cake and eat it, too.
14:57You can have state-of-the-art features and reliability without sacrificing your autonomy.
15:02Build on something you own.
15:03Build with us.
15:04Thank you.
15:06[applause]