Can make it working like you want or not?
Yeah, and I think we put a lot of work right now into helping our automation developers build better prompts, suggesting them prompts, helping them build evaluation sets. It's much easier to test a script that works as intended than a prompt, because the input data can make such a huge difference and it's very hard to test.
Will we see the size of companies reduce in the future? You mentioned there will move to a more verification and approvals based function for humans. That could also mean that you need less jobs. Are jobs changing over time?
We are seeing in agriculture, if you think about 100 years ago, 50% of UK or US population worked in agriculture. Nowadays it's only 2%, and they are mostly people that supervise machines doing their work. Yes, many of the jobs today will change, but there will be new jobs. Society evolves. You don't want to be stuck in the existing society. This thing will create abundance, and the only way to keep up with the economical growth is to increase productivity. With population aging and actually on the verge of starting to reduce, like in many first world countries, it's only increasing the productivity. This tool is essential to increase productivity.
Why are we focusing on the doom day scenario when actually we live in one of the best worlds possible, where the rate of unemployment is the lowest in I don't know ever maybe? So I could pose a different argument back to you, which is that the speed of progression through this technology cycle will likely be much faster than any other technology adoption cycle. You mentioned the farming industry, that was over multiple decades that you saw the adoption of farm machinery. People were able to transition. This transition, you underestimate the inertia of corporations honestly.
Even with RPA, our RPA technology is pretty good and it's not fully penetrated, not at all. It's about when, you're not fully penetrated, like 10%, 20%. It's kind of hard to understand really, but I would say it's probably less than 10 to 20% penetrated. But the rest is, it's not an easy technology to deploy. You need to put an entire program behind it. It's kind of the same with Gen AI. It's not that suddenly you will have an agent that sits next to you and you show it what to do and then it will run the job. No. It's going to take the next 5 to 10 years with the current state-of-the-art LLMs to get to very wide scale deployment of agentic plus automation. If LLM of course will go to AGI...
That is bluntly what many people are thinking, is that LLMs will continue. We are seeing more compute lead to an equal level of performance increase and that actually there is a lot more room to run in LLMs. Sam Altman says that we will see AGI in 2025, which you have about 15 days.
Yeah, look, my definition of AGI is a bit different than their definition of AGI. So AGI for enterprises would be when I have an LLM that has the capabilities of a guy with an average IQ like 120 points, but predictably, not 180 in some fancy math jobs and 60 in other types of jobs. Predictably. This is AGI. When we will get there, I think the jobs landscape will change completely, and it's not about RPA or automation. I think every industry will be subject to a big change. But I also believe that we need a new giant leap in order to get there. I don't believe that actual LLMs reason in the sense that I expect a guy with 120 IQ to reason. It's still a stochastic engine in the end. There is a big argument: is intelligence purely stochastic or is it something else? I believe there is something else, because otherwise, look, I am stupid compared to LLMs in many tasks, but I am much better than LLMs in so many others. Why is it that LLMs make logical mistakes that I cannot do, while they can solve some problems like math Olympiad problems that I don't have any chance to do? This question has to be answered because it shows me that actually the essence is different. It's a different type of intelligence, but that intelligence is not equipped to work in the context of business operations where you need reliability.
Yeah, it totally makes sense.
I'm just actually thinking, though, you know, when we're talking about the transition of labor forces and kind of what happens to people. Masayoshi Son said that there's n trillion of CapEx that will need to be invested to see AI agents fully deployed, and that would then create 9 trillion a year in GDP gains.
It's very small to invest n trillion to create n a year. Look, I think it would be an easy investment if there is a predictable outcome. But do you really think that just adding GPUs and with the existing algorithm to train, they will suddenly become godlike intelligent? I don't understand. And it looks like there are many signals that the training is kind of plateauing. I think I even heard Sundar saying recently that they are not seeing much more gains by training alone, so there should be something else.
Do you think Nvidia will be able to retain that chip monopoly? You've seen Amazon, Meta, Google all trying to build out their own chip capabilities, owning the whole kind of supply chain. Do you think Nvidia will be able to retain it?
It's hard to see it at existing levels. I think half of Nvidia's revenue today is from these five hyperscalers, and if they are all in the business of building, it's kind of hard. But obviously Nvidia has such an important moat in both hardware and software. Who would have thought in 2007 that Apple could maintain such an incredible moat?
I completely agree. When we think about business models of the future, everyone says, hey, we're going to kill the seat model, we've had enough of seats, we're going to replace it with units of work or units of cognition, and that's what you're going to pay for. Do you agree with that transition, and with your statement there of you underestimate this inertia and speed of enterprises, is that wrong?
I don't think it's wrong as a North Star of evolution. Why would I pay for seats when I need to pay for transactions basically? But I think there will be, again it's not binary, it's not zero and one. I think the pricing models in the future will be both seat-based and some kind of consumption-based mechanism combined.
What do you think is the biggest challenge that you as CEO of UiPath have to embrace, approach, chat like wrestle with in the next 12 to 24 months?
My biggest challenge is transforming the company to be an AI-first company and re-energizing our people.
Do they need re-energizing?
Yes. We had a rocky ride into the public markets and probably in 2021 we were...
To what extent was that avoidable versus unavoidable for a company that IPOs?
Maybe I think if I do a second IPO, I would be better equipped.
What would you do differently if you did a second IPO?
I think I would look into finance and go to market a bit differently.
What does that mean? Saying you're going to finance, go to market differently?
I think you can plan maybe a bit less aggressively but more consistently. I think it's better to plan and execute a growth of 30% year over year rather than 80, 60, 30, 20, 10. Because in a way, doing such a huge aggressive growth makes you maybe sometimes steal from the future not knowingly, but you discover in time. I think organic growth in a public market is better rewarded.
Paul Graham wrote about founder mode and the incredible benefits of having founder-like companies. You hire a CEO and then you come back as CEO. What are your reflections on founder mode and the importance of founder-led companies?
I think there are stages where founder mode works better. When I hired Rob, he's actually a great guy, but when I hired him I didn't realize that we were still at a stage where founder mode is essential. I thought we were at a stage where we are more established and we need an experience that is stage-appropriate for something bigger than us. So I think for a $1 billion revenue company, it's actually not that established. You need to get to a critical mass.
Do you think that's it? I think the revenue is actually irrelevant. I think what's relevant is the technology cycle. When you see Larry and Sergey, they are back at Google working, Bezos is at Amazon every day again. Everyone is coming back home because...
But Larry and Sergey didn't get back to the CEO job.
No, but they've come back.
I actually never left the company. I ran product and engineering directly while Rob was co-CEO and SEO. But it's actually, I always think that's harder with respect to you because you're kind of like in and watching someone else. In many ways, it's harder. It's a bit of a dance. People maybe I didn't speak enough to him, he didn't speak enough to me, and it could have been probably better. But I agree, in a time when there is such a huge change in technology, I think the CEO role is so important because you connect instantly product and go to market and marketing. This is a powerful flywheel that has to work. But most of us are not very good at fluff. I hate empowerment, I hate alignment, I hate all the corporate words that people use which mean little.
How do you reenergize a team? How do you say, hey, like oh, you know, kind of Wall Street style?
My way is to speak really very transparently to the teams and say, 'This is where we are, guys, and it's a lot of work ahead of us. We made these mistakes. We have to fix these things. It's not easy. It's going to require a serious, serious work.' And maybe that's not a way to energize them, but it's a way to make them hurt. I think one of the things as you grow big is that people feel smaller and smaller. I don't want that. I want to work in a company where people have joy, people feel that they are empowered to do things, not in the alignment, but it's very important that if you want to make a big change, go and drive a big change and have the voice to fight for it. If you are shut down, then it affects the morale more than the stock price.
But how do you actually do that though? You say, 'Hey, go take big bets, I support you.'
We are trying to reduce bureaucracy. We are trying to empower the regions to have more control. We are trying to get closer to the customer. This is how we build a lot of stuff inside, taking cues from the customer up to the product. One of the biggest rules of management that most people put too much importance on is being disciplined and having regular one-on-ones. I don't believe in one-on-ones with my directs. We have to call each other. If you cannot have an honest, candid discussion, it's not going to work in a one-on-one.
How many direct reports do you have? You know, Jensen, Nick, who we had on the show the other day, they both talked about having 45, 50. I think that's a great model to have as many as you can handle.
I think I have around 12.
What do you think about the role of CEO you are not handling well enough today?
I have two ways of working with people. One is very direct and sometimes might be construed as rude and tough, and one which is indirect where I avoid to really tell them what's wrong because I don't want to necessarily take the bull by the horns. But actually, what I'm telling people I work with is that a very direct style is a sign of appreciation. The moment you see I'm working around and finding my words, that's not a good sign.
Have you changed a lot as a leader?
In the essence, no. In understanding better what it means to run a company of course, but in the essence, I have not changed since I was 17. Who I am, my essence, I have not changed. I would have led this company as a 17-year-old without the experience to run it, but in the same way as a person.
When you look back at the last few years, what did you do that you wish you hadn't done?
I cannot point to something really major. Maybe I hired for experience and I should have never compromised on chemistry for experience.
