Transcript
0:00 In my opinion, even the labs in China that are building AI right now, I would classify them as the good guys. >> Joining me in the hot seat state is one of my dearest friends and one of the greatest founders of the last decade, Daniel Dyn, founder and CEO of UiPath. And today we discuss the biggest questions. What on earth does pacing the frontier mean? Is it even possible? Will we replace humans with AI? >> I've never hidden from my employees that there will be a transformation. Jensen is bound by the success of open source.
0:32 >> This is a truthtelling, mythbusting conversation, and it's just a fantastic discussion between two old friends on what no one is talking about, but everyone needs to know in the world of AI today. >> Models are interchangeable, but the workflow, the map of work and the workflows around the map of work is where the real value is. Ready to go.
1:05 >> Daniel, dude, it is so good to have you in the hot seat. I have been looking forward to this one. So, thank you so much for joining me, dude. >> Likewise, dude. It's always a pleasure to be here and it's the I think the hottest moment in technology. So I'm very excited to talk to you and get your perspective also on a lot of topics. >> I it really is the most wild time right now. And so I want to start you've written a book and a lot of people write books with the greatest of respect.
1:33 You're my friend. I care deeply about you. Why on earth did you decide to write a book as a public company CEO? No offense. You're not doing it for the royalties. Why? Why did you decide to write a book? Man, I dreamt to write a book since I was a kid and I discover I have no talent and now I got really a great opportunity. Claude and Chad Gypit help me a lot. They were my ghost writers and it was a good moment actually to put my own ideas in order because really when you write something you get much more clear perspectives on what you are doing. So it was a maybe almost 6 months effort doing this and I started with different threads of thought.
2:27 >> Mhm. >> One was you know what are the limitations of AI? Is there any durable limitation of AI where in a couple of years there will be millions of Einsteins in the data center and we can all go to play or do whatever we we would like to do because the Einsteins will do the work for us. >> Can we just start on that then? I think it's nice take it in kind of segments limitations of AI Einsteins in data centers. How should we be think about that moving forwards? Look, when I heard about this statement that in a couple of years we will have millions of Einstein in a data center, look, I was really concerned. And Dario is a guy that I highly respect and he's highly successful in in what he did. So I was thinking what does it mean for us?
3:17 What does it mean for can I hire one of these ash time put it in a laptop somehow assign you know an enterprise account a slack account and everything and I ask do my job or do whatever job in the enterprise it seems that reality is maybe different and probably Dario wanted to say that we'll have a millions of entities that will have the some of the reasoning powers of Einstein which I agree but not Einstein's as a persons not Einstein's that are capable of learning on the job because do you agree that one of the major expectation when you hire someone is that they will hire on the job there is no manual that a company has to give a new employee this is exactly how you do your job from end to end so we expect that >> no I think humans do learn on the job and they do improve on the job as do models. So similar there except humans get tired, humans want more money, humans want culture, humans can be toxic, humans are difficult to manage, I'll take the AI any day of the week, please. If AI can can or as well as as a human hurry, but you say you say that AI learned on the job, AI can create a notepad on the job, a scratch pad where they can memorize some of the policies on the job, but AI doesn't alter its weights on the job in the way humans are transformed by a job. This is this is a this is a huge difference because let me give you let me give you an example. You have two chefs. One chef 20 years has done only Japanese food, the other only Italian food and you give them one recipe. Okay, they will create different food. So it's not that you can write down your enterprise on a sheet of paper. It's it's much more complex. It's a becoming. Think about if I give to someone the the ability to read all the books about chess, do you think he will become a grandmaster without playing, without losing, without going through all of these process? Probably not. If I give if I have you watch all the videos about skiing, are you becoming a skier?
5:52 No, you're not become a skier. >> But I think it depends what task and workflow you're doing within the enterprise. If we look at the majority of actually what the people within UiPath and every company do whether it's accounting and finance whether it's marketing and sales whether largely outbound and inbound but whether it's social media and dude most of this is executionoriented where yes judgment and ambiguity and taste is important at the top but most of what people do is >> I disagree with you. I think most of the people will play will will display some sort of micro initiatives during the job. Maybe I have a hunch this customer is is going to churn and I can act before even any data is coming. How how do I develop this hunch? It's through my years of transformation. It's not written on a piece of paper. I it's a big difference between writing an operating model on a piece of paper and leaving it. It's almost impossible for for an enterprise for AI. Do you admit that everything has to be written down documented every time I'm asking a question to the model I the model will read my entire enterprise will have to read it. Okay.
7:18 >> Yeah. Sure. So this is not possible. >> But but but I also think you're talking about and with the greatest of respect today's state of AI. I think that the pace of AI >> this is one of the this is one of the biggest bottlenecks right now because AI doesn't train on the job doesn't train their own weights. Every time I am doing something after this talk with you I am being transformed and I carry with me this discussion in all my thoughts. This is not true about AI.
7:50 >> Well, I mean, I'm not being rude. It is like that's why people remain with Open AI because it has memory and it is able to infer from past queries, prompts and give you suggestions based on those. So, it does have memory. That's >> it has memory, but memory it's not necessarily learning. It's not the same thing. It's memory. It's just a thing that is written down. When I'm talking to you, I don't I don't go back into my memory. I I just I am just being transformed. It's like a model is like a new version of the model that comes improved. That new version is not the old version plus a piece of paper that has memorized. It's transformed into its own weights. This is why a model becomes much better. Look, I want to give you a simple example on on using our own technology. We we make our UiPath platform available for coding agents. So people are making it's much easier to create automations on UiPath right now.
8:55 But what we discover is that a model that has read open-source technology will and they have a lot of examples and they already have in its weights you know a certain technology will be much better than to create on our own technology because regardless how many prompts and skills we create the model has it in its own weights. It's very it's very different then and think of all the metaphors in the world when you read something versus when you live something you can read the biography it doesn't mean you live that life it doesn't mean you are transforman you are going to answer like that person that live it so I think to me really this is the the biggest limitations that the models have right now >> so I I actually do agree with you but I think everyone does and that's why everyone is chasing you know recursive self-improvement so much where it's models that can continuously learn from themselves and improve themselves over time without need for human intervention does that not remove the limitation that we just said >> I don't know then maybe we are the result of a self improvement loop because look do this thought exercise >> let's put the model that we have today with the best technology put it in a spaceship and throw it to the stars.
10:23 Let's say that we have this technology like I think von Newman imagined that is self-replicated this spaceship goes to different stars get energy and they can continue so compute will be infinite and models will self-improve where they would end up with maybe they will create a simulation of a world like ourselves right because they will intrude infinitely basically this is the theory so they will simulate ate a world as complex as our our own world within within it but that means that we are an part of an infinite simulations so look the answer is and I I don't know where it's going to lead but I know that there is a big distinction that I made in the book between will and reasoning and it's not like that We are certain that the will to do something you know emerges from reasoning or from even from consciousness. I think will is a kind of a separate you know part of the fabric of the universe that I don't think we as humanity have a clarity about what will is I think it's wishful thinking to believe that I can take a big model put into a self-improvement loop and this model is going to generate will I I I don't kind of believe so >> I get that but I We don't know. And I think that's what's so kind of challenging about trying to predict what happens. It's like a world of, as I said, recursive self-improvement is unknown in terms of where it lands. It's like a technology you can create can become something you didn't know it could be. And I guess for me, the question then is like, you know, we see the news this week. Dario says we need to pace the frontier. You've just written the book and have a lot of clearer thoughts. you run UiPath today.
12:33 Do you feel we need to pace the frontier? >> I would say if they truly believe that this technology is is is becoming rogue and they cannot control it and their experiments will will create significant loss for I don't know internet other systems. If I were them, I would I would pace it you know at at any cost because I don't want to risk going to jail. Honestly, I think there are laws that you know kind of control this type of rogue behavior. So honestly I I don't need an external pressure to control. I would be just you know concerned citizen and I would not build a technology that you know is causing harm. But now of course they think that this is the only way to protect against the bad boys. So we are the good guys. But there will be some bad boys that you know probably in other parts of the world that will build the technology regardless.
13:39 >> So I think that probably they ask more like a pass. I want to build this technology at any risk and I'm willing to open you know my my gates for other to to see how I'm doing because I want to do it in as of you know like good manner as possible but in the same time I want to be free of consequences to me I think this is a bit how I read this this memo and because otherwise Guys, guys, I think it's kind of obvious. I don't think we can we will reason with the bad guys and we can make a coalition with the bad guys to stop the frontier. So, we can make a coalition only with the good guys regardless. I I would take even in my opinion even the labs in China that are building AI right now I would classify them as the good guys but someone to me I think the indirect the attack is probably on open source because basically they say even if the good guys are building open source and but that open source will get into the hands of the bad guys and this is unknown bad guys this is the real danger so the danger is in open source. So indirectly it's also an attack to open source in in this way. It's it's it's a way of interpreting. I guess >> you work with some of the biggest enterprises in the world with UiPath.
15:08 Alex Coin said from Palanteer that the biggest enterprises in the world are scared to work with Frontier Labs because of threat of them coming in to their businesses over time. They have the data. They could build their own and compete against them. Do you see large enterprises scared to work with frontier providers? >> I think so. Yes. I I don't think people are scared scared that open AI will build a competitor to them necessarily. I I don't see this coming.
15:44 I think they are more scared that their IP would leak to other existing competitors. >> Mhm. somehow because I if I I don't know if I'm manufacturing somehow shrews or whatever I don't think openi is going to come and compete with me on this thing but probably some of other guys can get indirectly the other models the same intelligence >> if if I will train the model I think that's the real danger and I think it's it's a legitimate danger And I think everyone is trying to protect their IPS.
16:27 >> You you said about kind of clear articulation of thoughts that come from writing. What was another thought that you clearly articulated through the writing process? It has become very clear to me that u another limitation of AI is what I call exactness and AI by its nature being probabilistic at every step in a way it can lose it you know while you do 100 200 steps by AI even at each step you have 99% probability for instance it's you know 0.99 mult at power of 100 you will end up with maybe 60% probability to do the entire step right so AI doesn't have this mechanism to follow steps exactly hundreds millions of time in the same way you can see it even if you ask AI to multiply very large numbers millions of times at at some point they will make a mistake and also it surface kind of another simple idea even if you have a tool like AI to multi that is capable of doing multiplications why you are not using a computer that is doing these multiplications millions of times you know 100% of time exact the fact that the tool can do a job doesn't mean you have to use that tool to do to do that type of job >> isn't it because it's where you are and that's the importance of being in the harness of the workflow which is like I get you completely you could use something else. I'm asking here I'm not saying but because you're in chat GPT continuously every day instead of switching to a computer or calculator or whatever you ask what you're already in the importance of the environment. I think that chip acts like an interface to to convert my my my questions in natural language into exactness. But the exactness is not run by CH GPT.
18:42 Exactness is run by a computer. Because even today if you ask CH GPT please multiply these two big numbers they are using behind the scene a computer and they will give you the exact number. This is part of the power of the models. So you actually can see on on the desktop level on this type of cowwork or judg work you see the capability to call tools and what I am saying you extend this capability to the enterprise level when everything that should be exact should run on exact technologies. there is no point to run it on probabilistic technologies. And here here comes I think the most interesting thing. There is an a symmetry in the deployment of AI and automation in an enterprise.
19:38 Deploying AI agents is not getting easier today than it was two years ago in my opinion. But deploying automation has become much easier because I can create these automations with AI with coding agents. Coding agents been the most major giant leap that we were seeing in in the past year. I I would say since the invention of chat GP then chain of thoughts and then coding agents were you know the major milestones. So with coding agents that act in kind of a design time when I build the systems I can create automations that work with exactness every time during the execution time. So this is the asymmetry that is that is happening right now.
20:31 Plus when an automation breaks because any change in the upstream system AI comes again into play and fix the automation itself. So this is the this is really the pattern that we are seeing emerging in an enterprise and this is that AI it's actually creating the software that runs an enterprise and this software is is really cannot behave rogue because it can this software cannot change its behavior in real time.
21:06 It's not a probabilistic technology. Even if the software is created by AI, we can really audit. We can have humans that read it, validate it. I can have tests that you know for a certain input will guarantee that the software will behave in the same way. So that's again makes this pattern extremely powerful. You create you use AI to create software that runs the enterprise in a predictable, governant, auditable way. >> Totally get that. How many engineers do you have today?
21:39 >> Maybe more than a thousand. >> More than a thousand engineers. >> Yeah. >> How many people do you have? >> Around 4,000. >> You have 4,000 people? >> Yes. Why? So, right. >> I find 18 a nightmare. What? 4,000? Oh my god. No. Do you have too many? It's it's a complicated question because I think I would I would I would answer more with the thing that is in is in my book. I think the more we transform our companies using AI, we have to in the same time to transform our our workforce. I've never hidden from my employees that there will be a transformation in the company. But I told them up front, guys, we are not doing anything stupid. We are not just using AI as a pretext to cut a part of the company. We need to do the transformation in the same time as we successfully adopt AI in an enterprise. And that's that's actually another point that I discovered you know writing this book about if you look I I was looking deeply at jobs and what jobs can be affected by AI what jobs can be enhanced and how this transformation is going to is going to look like man one of the one of the thing that seems very simple in retrospect is that people don't have very simple jobs that again can be defined on a sheet of paper. I think every job has some kind of an outcome that is measurable and this is the outcome that people are hired for. But there is another outcome that is is part of the institutional strength. Think about my deep relationship with the customer. It's not necessarily part of the numbers that I'm producing, but it's what is maybe what makes this customer sticking to my technology. This is a different outcome.
23:55 If I'm going blindly and I cut, you know, on a number of A's and say this is because AI is going to replace them. AI can call the customers and write emails. I don't think AI can supplement the human connections and the trust. that so this is a different is this is a different outcome of the job and it reflects on every employee on every type of role in a company. So to me an enterprise should have a ledger where they actually understand what people are doing besides you know their main definition of the role and only after they have this ledger and understanding they can look at how does the AI transformation look like? What kind of jobs will be affected? How can I move people maybe from one job to the other because they still carry some kind of you know the the cultural aspect of the enterprise. So I I don't think it's a it's it's a simple problem as AI is going to cut 20% of the company. Let's do a reef 20% out and then we increase the AI adoption. I think on the contrary when you do this blindly you risk to hallow the enterprise of exactly the same talent that thrived during AI because let me give you an example in u in in the book I call this like the credentialed middle that kind of was the type of people that were prevalent in any enterprises and if you think of our education system and we h the way we hire people we hire based on deep expertise in a particular domain a credential expertise and this is exactly the type of expertise that might not be needed so much that AI can really help so you will need you will need a fewer of these experts but you will need more people that have initiatives that are capable of maintaining a relationship with the customer that can be mentors four new employees that bear, you know, the cultural aspect of the enterprise. So, it's it's counterintuitive because you you cut and you and you will tend to cut those people that are not the biggest experts in the domain, but you will cut exactly what you will need to bring the AI to supplement these experts. So, one one thing that I >> maybe you just need less of them. And so, you know, I was speaking to a lawyer today and I said, you know, how big's your trainee program? And he go, well, historically it was 25. And I said, wow, that's that's a lot of trainees. And he said, yeah, but this year it'll be four.
26:55 >> I 100% agree. We will need less of the people in probably most of the roles. But the main question is which ones? How do you choose? I think you can choose quite simply for where is their verifiability which is you know finance and accounting it's quite clear what is right and what is wrong. >> I disagree with you. It's not about verifiability. It's about if the if the work has been defined in a frame set by other people.
27:28 If the frame is clear then the AI can understand the frame. >> But the frame but the frame but the frame is clear. finance and accounting did they do your expenses? >> No, it's not it's not this is one of the domain where the frame is not clear because when I I receive an invoice or I I receive an order from from a customer I can treat it differently. There are not always the rules there. I know for this customer Nvidia is going to ship with priority to open AI. Maybe they have a rule that says there, yes, my first chips go there. Maybe they don't. That rule they don't. If it's not captured in a frame, I think AI cannot learn it. This is why you need to create this manual. We call this manual the map of work. You need to hand the map of work to AI in order to be successful.
28:24 >> What do you mean the map of work? >> To basically capture how the work works. how the work happens in an enterprise. This is the map of work. It's all the workflows, all the exceptions, all the procedures, all the systems that you use in order to fulfill a goal of a process. >> Sure. But then you have a a head of finance who sits on top of 30 agents. And exactly when Nvidia come back and say, "Whoa, whoa, whoa. we're your biggest, you know, buyer and we have special terms on our payments. They go, "Yeah, sure. That's right. Don't worry about it."
29:04 >> But, but you you come but you come to my point. Even in finance, you cannot replace everybody. So, you >> not everyone, but you got one person or two people. >> Yeah. >> Look, you you only need you'll need a certain number of people. The thing is if you have x number of people how how do you understand which of them stay and which of them has to be has to go to maybe do different jobs. How do you know? Because ideally you will get the people that will have literacy in AI that can display initiatives because when AI cannot exhibit initiative in the human sense. So out of this number of people that you want to keep, you want to keep the people that you know display you know the most initiative even a finance person treating an invoice with the customer contributes to the culture and how my enterprise is regarded. How can I make a distinction between that person and the other person that care less about how they treat the customers because I want to keep the first because the second is more prone to being displaced by AI. This is the ledger that I think enterprise have to create and to understand different outputs of people.
30:24 They need to judge people by their by the by this kind of hard to define output. So how do how does one do that then for the illeible data that isn't captured within companies? How do we do that? You Zach and Facebook have said about monitoring every single action that's on the screen of employees. I don't think that captures the tone of a call, the warm text afterwards to a customer, the illeible data. How do we think about capturing the data that shows value that we don't capture?
31:00 >> I think this is the crux of the problem. >> And I hear you're a true philosopher. >> Yeah. Yeah. And this is this is where we we put a lot of effort as a company and we are introducing a new technology that we call it like ctography and ctography is a discipline. So it's a discipline to help companies surface all the information of how the work is done and help them create this map of work.
31:33 And one big important of ctography is to actually have investigate what people are doing on their desktops. And we have a product that we call the ctographer agent that can interview people real subject matter experts have them basically record what they are doing and the agent is is interviewing them in real time. Like if if you interview a finance person, it can ask why did you put why did you change you know this this invoice when the zip code was different? Why did you choose a different path? Tell me and they can start surfacing all of these exceptions.
32:16 So that's real agents interviewing real people. It's pretty cool stuff. And then you consolidate data from multiple people and then we create the process maps. So we show them how the work works in in in in kind of real time. And then after this map of work that show the work as is, you can use our coding agents and you can come up with an idea of how you should transform the process. and you transform the process by printing software. So you start with a process that and at point A it's kind of fully manual. Of course you have enterprise systems like system of records and everything but people operate the system. I think the goal of any enterprise is to have less people operating the system and more automations and agentic AI operating the systems. And do not fear push back from people working in the company aware that you are watching what they do to replace them. That is what Zach got. This is inevitable and so I think it it very much depends how you pitch the company. So in in UiPath I think it was important to tell people again we are not doing anything stupid. We are not doing any mass extinction you know for the predence of AI but transformation is inevitable and you guys have to transform and everybody will get a chance and the people will that will become more literate in respect to AI will have a better chance not only here but in the future in any other job I think that's the message that I think everybody should get. So >> did they respond to it? Like did you see AI adoption through the roof after that?
34:16 >> I think the I think the response is good. A adoption requires more cycles than just discovering the process and people's you know input in this look in the beginning of the year I think the fear of people across the industry and not only in my company but in many companies was off charts the fear of completely being replaced now I think people start to get a bit more understanding of the durability of their their jobs and this AI diffusion in enterprise that can happen at a slower pace and can happen one process at a time because this millions of Einsteins are not hireable yet. Can can I ask you one of the companies we've invested in Mccor data provider Brandon Foody the CEO tweeted yesterday that they spend 3x the spend of human salaries on inference today.
35:16 What percent or multiple give or take would you say you spend human salaries on inference? >> Look, I personally don't care about it. And let me let me tell you something. It's a it's a simple hypothesis. If the work at the quality were better of a human can be done by a machine I will hire today a machine even if it's more expensive man than a human because human human cost will only increase and they bring errors into the pictures while the cost of machine would increase. So I will have a competitive advantage compared to people that will stick to humans if that's and I think everyone will do this. So I I don't think the cost of tokens it will be the real question if you replace a person with with AI but the real problem today is that AI cannot replace a person because if again if if bring me an Einstein that replace me and I will happily go in any vacation in the world but this Einstein doesn't yet exist and probably I will I would like to be interviewed and have this podcast with another Einstein man this thing doesn't exist today that's the reality so let's call it a reality let's call a spade a spade maybe this technology will emerge and somehow Einstein that embody like a person that have will that gets transformed on the job learn on the job have the capability of reasoning imagination of Einstein's exist of course all the jobs will go extinct >> I get you I have a show with Jason Lanin from Saster famous SAS platform. He's you know cut his team from 25 to two. He would he if he were in my seat now he would say no no no it does I replace my VP finance I replace my VP marketing and actually the AI is better.
37:11 >> I I want to see this man. I think I I I heard you know companies that replace hundreds of support people in the past and now they are rehiring these people. I think until until this model is proven and is proven at scale not in a particular industry for a particular guy. I don't think we can we can extrapolate for one data point that is going to go across industry. >> You said about extrapolation and overexaggeration. The SAS apocalypse was very real. We're going to vibe code everything.
37:47 >> Did you vibe code tools out? >> Look, it was it was amazing. Yes, we did, but not an extraordinary success. Initially, it was it seemed extraordinary, but when we tried to put it in production, we started to see, you know, some kind of real bottlenecks with these tools. and and you need to have you know a lot of things that you have to maintain connectors permissions audit security taking a software from a prototype to production it's actually where the work is not necessarily the writing code writing code is fun but it's not there where you can really makes the difference I think it's much easier today to make a prototype prototype is so easy But then you have to iterate to to make the prototype in production. And this is the testing and everything else is look we we were trying to to replace like a procurement tool writing ourselves and I think we have a lot of success and it was in initially it was written only by AI but then it comes to you know speaking to this self-improvement loops kind of don't trust we don't have enough test we don't have enough trust to put this tool in production 100%. And humans in our experience humans have to intervene a lot into how this vibe coded tool for instance the database schema that vibe coded tool created was completely bogus.
39:33 So it had a human has to come and you know create the structure. Right now you are not at the point where you know you will have like a business user that understands a problem and they will vibe code a tool. So you you will still you will still need to put engineers and people you need to maintain it. So it's it's a long process. So eventually you will end up paying probably as much if not more as the tool you replace while you keep some of your good and best people bandwidth occupied.
40:12 >> Would you buy would you buy Salesforce today? >> I would buy Salesforce as as a system of record >> as a stock >> as a stock. Look, I invest in I invest in in software as a category and I think I made a good investment in you know few months ago because I I bought in the in the in the bottom of SAS apocalypse even our own stock you know has been has been doing better and but the markets the markets today are driven so much about sentiment and not the not the value. So it's kind of hard for me to make a judgment of an individual company. But I don't think Salesforce can be replaced by Vive coding if this is the question. I don't understand why a company would go public today. If you think about the two drivers of being public, number one is liquidity for shareholders employees and shareholders. Well, Stripe and many companies are able to have liquid stock in private markets. And and two is the ability to have M&A, a tradable asset that you can buy with. I mean, many many private companies are able to buy with you private stock.
41:25 Stripe, we're going to do PayPal with private stock for 60 billion. So that's not a barrier. And the casinoization of private markets as public markets as you said with current stock markets being sentiment driven, I I don't understand why one would. But Harry, let's don't make a confusion between, you know, some very exceptional companies and the most companies that are out there. There are so many companies that are kind of zombies right now. This 2021 zombies, they would fare better in the public market right now. At least their investors will have a way for exiting.
42:04 Their employees will have a way to make some money. Nowadays all of them are sitting on paper money. >> Mhm. >> Okay. Public markets will face them with the reality of you know what's their real valuation. So I would not say there is no value in the why why the anthropic is trying to do IPO in the end. >> Well but they're a unique company alongside open air which just has to because they have exhausted all private funding that exists. I mean they they are literally hitting the tap out button on the supply of private capital.
42:41 They're extraordinary companies because they just need too much money. But like >> do you believe do you believe all the investors in open and will stay in the companies for years to come? >> No. So maybe >> I think some will do but >> some will do of course but I think we will see an exodus of and I honestly I don't believe at two trillion or whatever maybe they would reach 5 trillion or because if I buy a two trillion I I need to have a path to five >> but if they if they went public at two trillion would you sell >> anthropic in particular I I wouldn't because do you remember on our last podcast I think you asked me what's the company that I bet on and I said Antropic and Antropic was worth 60 billion market cap. Maybe I was stupid.
43:28 I didn't invest. >> You would have made more money on that than Salesforce or Service Now or whatever. >> Yeah. Yeah. 100%. So, look, I I would not sell. I think it's same with with OpenAI, man. I don't have I think OpenA is has catched up quite nicely and I use interchangeably right now. I just think we're in a market where the big get bigger and value concentrates more than ever. Actually, more than ever. >> But the real question is, Harry, would I buy at two at two trillion? That's that's my real question. And look, right now, I'm I'm not I I don't I I need to see their real numbers, you know, to understand if I will go if I will put money in their IPO. I'm going to get in so much trouble for this. I think AI is quite like Bitcoin in just the way that it's very difficult to determine what application is going to win, what wallet's going to win, what usage is going to win, and if that is the case by the underlying infrastructure that you know is going to be there. And so for me, I agree with you. I don't know if Claude is going to be better than the next codeex. I don't know if Cursor are going to come out with something amazing. But I do know that Jensen's going to be sitting there going, "Here's another chip. Here's another chip.
44:42 Here's another great I'll take more >> but I think Jensen is bound by the success of open source >> because if anthropic entropic and open this is becoming a dual poly and basically their their I I think their TAM is in like trillions it's is basically the work >> yeah they will print their own chips man honestly it's not such a big deal in the end to print chips >> of course I mean open AI being halapo and anthropic are doing there.
45:14 >> Exactly. >> So I don't I don't think Jensen's will be doing so well if they are the single biggest providers and the source of truth and light of God will come from only anthropic and open a so therefore the open source should succeed. I absolutely agree which is why I think Jensen is doing the open letter which everyone signed. Absolutely. Encouraging open source >> and I think it was a great move bugging buying >> hugging face.
45:42 >> Hugging hugging face. Yes. >> Why? >> Because I think it's it encourages it host all the open source model. It's becoming it's putting the money where the money is for his company. >> Totally. It does also make a neutral provider not neutral anymore. Bias. Yeah. Obviously, they have Neatron and they have their own models now as well. You could say there's a loss of independence now that it's owned by Nvidia, but yes, >> I think Neatron is still a small cog in in the picture. I think it's valuable, but I would not call it it's not the same chip size as the others.
46:22 >> Do you worry about the roundtpping revenue? You know, everyone talks about Nvidia investing here, buying here, the circular economy that come from Oracle and Open AI. Do you think that's overblown? >> It can be because every major infrastructure in history has been overbuilt. And I I think it's a simple explanation because I was thinking why every infrastructure is is overbuilt because you have to to make sure you get you know the biggest piece of the opportunity.
46:55 If the opportunity is big it doesn't it doesn't matter you build a little bit more than it's necessary. So it's clearly that now everybody that it cannot be there is only 100% of the pie and people are building right now 200% of the pie there will be losers. >> This is the first innovation where we are significantly underbuilt and actually if you look at the constraints now you're you're right actually in prior technology cycles we overbuilt supply side and demand side was lagging behind. Now we have energy that's being a massive constraint. We have water, we have data centers, we have regulation and policy. We have a significant hindrance to supply side and we are underbuilt not overbuilt which is why every ounce of computers taken.
47:44 >> Yes, but are we underbuilt to the extent of this trillions that comes into the infrastructure? I don't know to answer this. and everything happens on the premise that AI is going to replace human human work in a really large scale. We need to see the timing of this replacement and transformation. It's a big difference if it's coming in 10 years versus next two years. So I don't think it's overbuilt for next decade but it might be overbuilt for the next three years and stock markets and capital you know can be merciless and >> maybe I'm a childish optimist but I saw you know Andre Capathy say that he used coding tools for 20% of the work and then 6 months later he said that it did 80% of the work and he helped it with 20%. I we're investors in Lora. I interviewed lawyers when we did that deal and they all said to me, "You're such tech bros. You think you can replace us?" Haha. We went to law school. And I was like, "Okay, cool." I interviewed them two weeks ago. Every single one of 15 said they would be severely unhappy if it was taken away with most of them saying they hadn't written a document in 6 months.
49:03 >> Yeah. Harry, you should you should read my book, my friend, because it answers exactly the same questions. So >> when when the when the frame that a person operate is really well defined by someone else like in law AI can be devastating and the impact when the frame is not as clear and there are so many exceptions that are custom made for an enterprise.
49:34 Dude, you're about to spend two days with a lawyer who is my girlfriend. She will tell you that law is highly ambiguous, subjective in terms of writing styles. She would >> is the human language and AI understands it perfectly and it understands every freaking nuance of you know human kind of sensitivities as long as it's documented and it's well defined there is a manual for the freaking law AI is amazing. When there is no manual AI is not amazing and it doesn't work. That's the That's the huge difference.
50:10 >> But $300 billion is the legal industry in the US. It's a lot. if you think about how much labor could be replaced by that, I think 30% would be reasonable. That would be $90 billion of available revenue. >> Yeah, but that's not it's not going to convert in token revenue. >> What do you mean? I mean maybe out of 90 billion I I think companies might charge maybe 10%. Maybe it's a 10 billion total opportunity in tokens.
50:43 >> Am I mistaken you to invest in Lorra and are the Harvey investors mistaken to invest in Harvey if it's 10 billion not 90 billion? What I can tell you that from a from a law perspective, I think open-source models, frontier models will do just fine. Maybe they do all the they do also the custom workflows around the legal process which is really valuable. But to me that's also you know a big part of my thesis.
51:22 models are interchangeable, but the workflow, the map of work and the workflows around the map of work is where the the real value is. If Harvey and Lora are doing this, they map really the work, they create the workflow and they they create a legal department for me, of course, it's a bigger it's it's it's a much bigger value that they capture. But if it's only to get a legal opinion, a call to a model, that's not going to that's not going to be a hundred billion dollar market. 100%.
51:57 Man, >> what percent of token traffic do you think will go through open versus closed models in 12 months? >> To me, I I think the I think the question is different. what percentage of the traffic will go to truly frontier model like Astra or Fable versus a very costefficient models and to me I think for enterprise work my prediction is that 90% of the of the flow will go to very costefficient models I don't think you need frontier level quality of models for most you know operational work >> and so just to be clear then so we will actually still use the core provider which is open AAI and anthropic it'll just be deprecated older models >> I will still use anthropic and open AI with their costefficient model but I will have a a verifiable backup on on open source all the time I should be as a responsible enterprise I should be able to switch models I cannot say I cannot put locked I cannot be locked in.
53:11 >> Can I ask you we invested in I'm just checking my portfolio against your brain. I believe strongly in open models and I think that every company not every company but mid to large scale company will have their own model own their own intelligence and feed their own data into it. And you know I think it goes to the statement of kind of owning your own intelligence not renting it. It's why we invested in fireworks and I believe in the open model ecosystem.
53:36 >> Yeah. You know, I'm a big fan of fireworks and we are using them quite a bit. >> Do you like them? >> Yes, we like them a lot. And I I I also I'm a big believer that an enterprise should distribute their bets and one of the bet should be on open source and very importantly should be on this map of work because think about if I want to train a model my own model with who am I need to have this who am I very well documented. I need to create this manual because I'm training one model today but in next two months there is another better model base model coming into the picture. How can I do transfer learning from my old model into the new model if I don't have the data and the exact menu? I cannot where there are you know terrible losses when I do this. So the real investment for an enterprise is to creating this map of work that documents how they actually work and with this one they can train their own models you know it's in fireworks or other provider doesn't matter but this is their IP this is and this is their core data make sense >> I totally get that and so you do sorry just so I understand so you do believe that companies and a lot of them will have their own models with their own data I I do believe that they will at least have their own models as a backup to Frontier models. I'm not to me where I'm not clear if I can provide the same cost efficiency with my own model versus you know a costefficient model from anthropic and open AI because I think these guys are in the position to truly optimize you know large infrastructure.
55:36 So I think part of their business model will be to deliver you know more intelligence per dollar then even I can squeeze from my own models. If you have highly specific data that is exact to the request that you have which is your data I think you'll get more token efficiency with your own model than you would an optimized >> only if you are training your models that might be true and only if you can deliver if fireworks can deliver at a large scale this and in a very optimized way >> would you invest in fireworks at $15 billion >> probably yes If this hypothesis of open model is true, which I believe is true, I think they are under undervalued. I think they will have to get very soon in this big game of securing compute because if they don't secure compute, I don't understand how they will how can they give me the inference at the scale that I want. What do you think? Because you invested in them.
56:41 >> I did. I think you're absolutely right that they need to move into the compute layer and I think Lynn is doing that I'm sure very soon. >> So they will have to raise you know I will be there. >> Yeah. >> No and she did that at Facebook. I think that's a unique thing to this team. They they did compute securing >> but it's a great team. We really like them. >> It's a great team >> and we work with them before the big hype you know around them. I also think the data providers are massively underpriced and underappreciated. Mccor and Serge in particular, everyone's like, "Oh, they're commodities. You're just buying data." Data is the most important thing to feed model quality. I think that what's the difference? I think one thing it's storage and one thing is understanding of the data. Because if I have a storage, I can have a tape and I can put data on the tape. Would we invest in a tape company? I don't think so. You you you need to invest in the intelligence that understands the data and feed the model and you know extract the extract the right data at the right time. fit feed really the model with the data that is needed with the context because if you don't have if you have just data but you don't have a way to create a really good context to give the model when they ask something it's useless.
58:05 >> I I think you would say that they have more data than anyone else across more categories than anyone else. So when the model requests highly specific data because of the breadth of their library, they're able to provide it in a way that others aren't. >> If it's there on data and is valuable for models, I'm sure the models will buy will buy the data in an instant. >> Can I ask you what have you changed your mind on most in the last 12 months?
58:30 >> I didn't understand this necessity of AI to have a manual in order to work. That was maybe the biggest breakthrough in my understanding that every time I'm running a query towards AI, AI should have at their disposition the entire way you know my company work or this particular process work. When I realized this, I think I understood also that this thing that it's the biggest differentiation between memory and true learning and I let you think because this is how we started the discussion and I'm not sure I really make a point but it's a huge difference between just laying down having a scratch pad or being transformed by an experience >> and That's a that's that's a thing that I realized the most and I think I realized also what's kind of human for us because I I I think I experienced a lot with AI writing not code but writing a book. So I've been through different styles. I I I understood a lot how to prompt them. AI doesn't have a style and you you realize why they don't have a style because they are an ager of anything. In order to have a style, you need to have kind of a body. You need to have individuality because we we are the choice that we made and the choice that we don't make in in a sense. So you need to be transformed. You cannot just give because otherwise I will just ask AI read this book and write in the spirit of this author and it's not really working and because you need to be you need to be transformed by the experience.
60:39 So to me this is going to be the biggest breakthrough in in the AI technology when I can have models at the size of mythos being transformed on a job being you know putting in a laptop and it might be possible who knows in you know the pace of technology maybe 20 years from now I can have a 10 trillion model that it's my own model and is getting transformed form along with me and that's but we need to see I think there might be a few series of innovations to get there because I want to give you also an interesting data point AI is solving very interesting math problems that humans didn't solve before right now but AI still is not capable of creating frameworks like I don't know relativity is a framework Okay. And I was thinking why so? And I think one of the main reason relates still with this not being transformed when you are on the job. When I when I'm writing a book, I am being transformed by the act of writing this book. Every time I'm writing something down, there is something in me that changes. That is not necessarily the memory thing. It's me that is changing. Einstein has been changed by his experience thinking about the speed of light of what happens when you go, you know, behind the light. It's not like Einstein wrote it down and then every time he thought again he rewrote a piece of paper. No, he became gradually you know a different Einstein that the one that started to think of a problem when he created this thinking this frame of relativity models don't do this way even if I put swarm of agents everything they have to write down everything they are not being transformed by the process so therefore it's very difficult in the end they will have this you know context the one million is very hard to go be you beyond this 1 million tokens context window. A frame might require a transformation as you work on that on that frame. It's a it's a it's a different it's a different way of learning than pure memory. So that's I want to make this argument as clear as possible.
63:18 >> Are you optimistic for your children? >> I'm extremely optimistic for myself. Therefore for my children. So, I I don't want to sound like an AI doomer, man, because I I believe that >> I don't think you do. >> I am. >> No, I don't think you do. I think I think you sound I sound like a doomer in a way. I I think we'll have a lot more job loss. I think it will happen a lot quicker. I think we're seeing it in real time.
63:42 >> I'm much more optimistic that we won't have so much because based on my own experience with the I don't think the diffusion is as fast as you imagine particularly because enterprise have to document much in much greater detail their processes. >> Can I just ask sorry we're both Europeans and we're both sitting in London. I don't know how to say this, but we don't matter anymore. Just being blunt.
64:17 Do you think that gets better or worse in the next 3 to 5 years? >> Yes, man. It's hard to it's hard to admit the reality, but from a technology standpoint, I think we are largely irrelevant. But it's so stupid because you know the biggest producer of machines that make chips is based in Europe >> is ASMR. >> We could have Yeah. We could have made these chips in Europe. You know some of the most brilliant minds that build AI even if you think Dario Sam all of them are kind of you know European origins Ilia recent. So it's we have the talent, we have the technology to build the machines but somehow we are losing it. It's and it's very stupid.
65:10 >> Do you see a difference in work ethic having a team in the US and the UK? >> Yes. Yes. I I experience with teams in UK and at 5:00 p.m. they are all you know in the pop. >> Why is that? Because money matters more. >> I think culture matters more than money. It's a more dynamic culture. This is why I I don't think I would have succeeded in Europe the way I did in US. So I I'm an European but as an entrepreneur I am I am I'm I'm American. This is what I told to everybody. So I I my formation is at you know it American school of entrepreneurship even if I started my company in I listen I I get it and you look at Lorra and you look at 11 Labs and some of the best companies to come out of Europe in the last few years and if you think the revenue machine is anywhere but in America you're lying to yourself. Of course >> it's an easier to access revenue machine than in Europe. clearly >> faster it's easier the teams have scaled GTM functions before I completely agree with you >> and American companies are making larger bets on vision without waiting for so many you know proof points as the European companies and even people in the middle management can make sizable million dollar bets on new technologies in US I I haven't seen this appetite in Europe.
66:44 >> I I don't want to ask this but but I am interested. If you were to advise a young European entrepreneur today, would you say to go to the US? >> Yes, that's the that's the said reality. I would I think that unless they build for a specific market with with some specificity in mind, if they build a universal technology, they will have a better chance to succeed in US. there will be many successful European companies coming out of this and I think maybe not as a frontier labs but I think for the application of AI and I think >> do you do you buy sovereignty as an argument >> yes I think it's an important one >> the energy sovereignty model sovereignty >> 100% and all European customers right now would prefer onrem software models range and model optionality. Yes, 100%.
67:44 Which is a big business that it's it's coming here. Look, I talked to our friends of fireworks and I actually tried to convince them to make their software available on prem right now. They are in show me the money, but I can tell guys this is a big business. You need to prove first because this is Europe. Show them the technology and the money will come. How much revenue does UiPath do today? >> I think it's public data. We are at 1.6 growing last year like 14% and this year.
68:20 >> So Jason Lkins taught me that like unless you're growing 20% plus, you're just in the public markets like it's grow or die. And it's a horrible reality in that way. I'm not condoning it. It's horrible. >> Yeah. >> Is that right? >> Yeah. Because I think the public markets are very confused right now of who are the AI winners or losers. And if you don't show serious growth and traction, they automatically put you into AI losers without looking really deep into the business. There are so many hundreds of software companies in the public market. So then it's hard to look at each of them. So if I were to flip it on you and we do a final one for a quickfire, what is the bull case to to UiPath being a $50 billion company?
69:14 >> And think about even today we Gardner released their new boat magic quadrant business orchestration and automation technologies and we we are one of the leaders. We move from a challengers into a leaders in the last year. So it shows that we as a company made this transition from an RPA an automation technology into an orchestration and automation technology and there are all the arguments in the world that this is really required in order to create this new enterprise that is AI powered. You cannot this idea that you can have an AI agent that runs everything for you from top level processes orchestrate and automate everything by magic. I think it's kind of a thing that people stop believing. So you you need to have an underpinning orchestration and automation technology and this map of work that I talked in order to power your processes. And this is what we have and it's not only me saying but it's Garner saying it's forester it's industry analysts in being very bullish on us. So that's really the the argument right now. I think as I said this asymmetry that AI is creating right now is more obvious than printing software that run your processes has become much easier than a year ago.
70:57 Bringing creating an AI agent that run your software is as difficult as an year ago. So you make a huge investments into printing this software, capturing the enterprise context that we call the map of work. Putting this enterprise context inside this rails that I I I name this orchestration automation as the map and rails. Map is the context. Rails is the orchestration automation. You put them in the same platform and then you can assign an agent to do work and you tell the agent this is the reality. This these are the rails you can use. This is the map that describe how to use these rails. This is the goal and that's the way you can control you can have a control on the top and your agents cannot go wrong. You cannot have no sane enterprise right now will put a swarm of agents and just ask them do my finance accounting for me because who knows maybe they will they will attack your competitor >> to send 20 million bucks to some rogue invoice. I completely get you. Okay.
72:06 What is the bare case? >> I think the bare case is AI will somehow become genius. will tokens cost will be next to zero and we'll have this literally millions of Einsteins in a data center but Einstein's in a true sense not only reasoning in the sense of replacing a person and I will I can assign them to every work in an enterprise and they will just do it that's the bare case against us I mean token cost has gone from $60 to $1 per million. So, I mean, I think the token cost will go to next nothing.
72:54 >> It's possible. This is why I told you I would not right now stop an investment based on the token cost. >> Dude, I want I could talk to you all day. I'd love to do a quick fire with you. So, I say a short statement, you give me your immediate thoughts. Okay. >> Mhm. >> What's the hardest thing about your job today as CEO of UiPath? >> It's aligning people. That's so different personalities, pride, ego that comes into place. This is the hardest.
73:20 >> What's changed most about how you work as a CEO because of AI? >> I'm spending maybe half of my day right now. Half of my day alone with myself in Visual Studio Code right now working with clothes and chat jupit. And I I have way more leverage on my company than than before because we we changed completely the way we operate. Most of the people when they come with an idea to me a year ago they would come with the deck and it was very hard even to prepare with this deck. Now everyone is going to come with the markdown file and I can put it into I have a giant strategy folder that I you know I have you know AI agents working with this. I put this document in my folder and then I can ask intelligent questions about.
74:19 >> If you had unlimited resources and zero retribution from Wall Street, what would you do that you're not doing? >> Sorry. You know you know what we do? We create clips which is like you take little segments and that's like the clip. I'm just thinking of like it'd be a great first clip. That's it. Maybe I like whoa. yeah, that would that would be interesting. Nvidia in 3 years time will it be above 7.5 trillion? It's >> where there are today five >> five.6. Yeah.
74:53 >> Yeah. I can easily imagine a 40% run for Nvidia. I I would bet more on Nvidia rather than Anthropic being a 7 billion company. >> Brilliant. >> Brilliant. Of course, >> billions are nothing today. >> Billions are nothing today. Yeah. You said something on on a show that we did before and it was like one of the most resonant things I've ever had in a show. And you said, I think a lot of people like think they want to be me, but sometimes it's quite lonely alone in my head. And I always remember this because I I often feel the same. what would you advise to founders who feel lonely in their head and struggle with that today? I think they should surround themselves with with their best friends from maybe childhood and you know have more frequent chats with them because I think they are the people that can relate the most to them before and they can see them. It's part of the transformation and it's a nice thing to do anyway.
76:01 You'll still be lonely but you will have the sense of some kind of continuity in in your life. I I find one of the most rewarding part of my life chatting being with friends with family. This is really where you get a lot of the loneliness in a way you know appeal from the loneliness. >> A final one. What are you most excited for when you look forward? like my mother's got MS. I'm really excited for some of the breakthroughs that we'll see with chronic conditions and treatment of them.
76:37 >> Yeah, I'm I'm very excited about longevity. You know, my friend, I have almost twice your age. So, my my body aches quite >> any. >> Dude, are you kidding? >> Any kind of gym or anything? >> What are you doing longevity wise? >> I I'm doing quite a lot. I I got into like peptides into supplements. Many I think I I taking around 60 different supplements a day and three supplements.
77:09 >> Seriously? Yeah. >> 60 >> 60 6 Yeah. >> What the are you taking? >> And all of them have been recommended and vetted by AI and >> what that's extraordinary. I mean you look incredibly young but 60 supplement like in pills. No, you take them in like cuz I do the longevity shape from Brian Johnson which is like 60 in one and I just have it every morning. You actually have 60 separates up. >> I I I I have a lot of pills. I also do some in form of powders but yeah I have I'm going to show you tomorrow all my all I have I have little bags you know throughout the day like AM1 am2 AM3 >> that's extreme that's extraordinary.
77:57 Do peptides like do you feel better? >> I think it's supposed to feel in the long term better. But honestly, I feel I feel way better even than 10 years ago. Kind of reducing booze quite quite a lot helped. And I know you are a big fan of booze. >> You don't still drink, do you? Do you still drink? >> I I drink a lot less these days. >> I love that, dude. This has been so much fun. Thank you so much for putting up with my meandering. when does the book come out?
78:28 >> It's already available for a download and I I'm printing also a few copies. We have our big fusion event coming in a couple of weeks and I am I'm distributing to everybody coming there a copy. >> Dude, this has been a pleasure. I'm going to get a copy. I'm going to get a physical copy cuz I'm old too and so I like reading. >> It's my gift to you, Harry. Of course. >> There we go, dude. Thank you so much.
78:52 >> Thank you, man.