Skip to content
The Twenty Minute VCThe Twenty Minute VC

Turing CEO Jonathan Siddharth: Who Wins in Data Labelling & Why 99% of Knowledge Work Will Disappear

Jonathan Siddharth is Founder and CEO of Turing, one of the fastest-growing AI companies advancing frontier models. Jonathan has led the company to an astonishing $300M ARR with just $225M raised and a profitable company. A Stanford-trained AI scientist, Jonathan previously helped pioneer natural language search at Powerset, which was acquired by Microsoft. ----------------------------------------------- Timestamps: 00:00 Intro 00:51 Redefining “Talent Marketplaces” Today 03:46 Data, Compute, Algorithms: What is Most Abundant? 16:59 The Biggest Challenges Enterprises Have with AI Adoption 20:57 Why Will 99% of Knowledge Work Will be Gone in 10 Years 28:53 How Will Data-Driven Feedback Loops Replace Technology as the Moat 34:20 Is Revenue BS in Data Labelling? Are Players Calling GMV Revenue? 43:43 Are We in an AI Bubble? 52:22 Why is SaaS Dead in a World of AI? 01:00:32 Will the Phone be the Primary User Interface to an AI World? 01:07:46 Quick-Fire Round ----------------------------------------------- Subscribe on Spotify: https://open.spotify.com/show/3j2KMcZTtgTNBKwtZBMHvl?si=85bc9196860e4466 Subscribe on Apple Podcasts: https://podcasts.apple.com/us/podcast/the-twenty-minute-vc-20vc-venture-capital-startup/id958230465 Follow Harry Stebbings on X: https://twitter.com/HarryStebbings Follow Jonathan Siddharth on X: https://twitter.com/jonsidd Follow 20VC on Instagram: https://www.instagram.com/20vchq Follow 20VC on TikTok: https://www.tiktok.com/@20vc_tok Visit our Website: https://www.20vc.com Subscribe to our Newsletter: https://www.thetwentyminutevc.com/contact ----------------------------------------------- #20vc #harrystebbings #jonathansiddharth #turing #datalebelling #ai #data #saas

Jonathan SiddharthguestHarry Stebbingshost
Dec 1, 20251h 17mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 0:51

    Intro

    1. JS

      I think the era of data labeling companies is over, and it's now the era of research accelerators.

    2. HS

      Today, I do not pull any punches with Jonathan Siddharth, founder and CEO of Turing, a company that he has scaled to over 350 million in annual recurring revenue.

    3. JS

      All knowledge work is going to be automated. It's only a matter of time. I don't see an AI bubble. These models are incredibly powerful today. SaaS, as we know it, I think is over. I think it's completely over.

    4. HS

      Ready to go? Jonathan, I've been so looking forward to this. Thank you so much for joining me in person. It's such a treat to do it in person while you're in London.

    5. JS

      Thank you for having me, Harry.

    6. HS

      Now, I want to start with a little bit of definitions because

  2. 0:51 – 3:46

    Redefining “Talent Marketplaces” Today

    1. HS

      everyone thinks they're talent marketplaces, and then everyone pushes back on talent marketplaces. How do you describe it, and why are we not dealing with talent marketplaces anymore?

    2. JS

      So, I think of a talent marketplace as something that's basically matching talent to something, maybe it's an opportunity. So, so Turing is not a talent marketplace. The, um... what we do at Turing is we're training superintelligence. We work with seven out of the eight frontier labs. To get to superintelligence, you need research, compute, and data. Research, the labs do in-house with OpenAI, Anthropic, DeepMind, et cetera. For compute, we have, uh, Jensen to thank, and maybe NVIDIA as well. But, uh, on the data side, Turing power is the data pillar. On the data side, there's been a significant shift in the last couple of years. So a few years back, uh, the models weren't quite smart enough. And as the models have gotten increasingly smarter, the data needed to improve them has become harder to generate.

    3. HS

      And this is because it's more sophisticated data that's required to improve the models. It's like vertically specific people in task and workflows that isn't so obvious, like cat pictures.

    4. JS

      That's correct. That's correct. And it's, uh, there's a shift in the data going from simple to complex. I mean, let's take coding, for example. Uh, a few years ago, the kind of dataset a, a contractor who gen- could generate might look like, "Hey, write a Python program to sort some numbers." Today, the data that's generated might be, "Write a, uh, a B2B marketplace app that connects, uh, doctors with patients. And write it on, for Android, uh, with Kotlin/Java, write it for iOS with, uh, Swift, and write it on the web, like with Next.js or something." Right? That's the complexity. So there's a shift in going from simple to complex. So it's no longer the kind of data that low-skilled, medium-skilled contractors c- can generate. You need expert humans in every domain.

    5. HS

      Yeah.

    6. JS

      The second shift, uh, is, uh, we've gone from teaching AI to take tests and pass tests to teaching AI to do real work. It's less about having AI pass the bar. It's more about, can AI do the job of a lawyer? Can it do the job of a privacy lawyer, a compliance lawyer, a paralegal? So having AI be good at doing economically valuable work. So that's, that's a shift. Uh, the third, uh, shift is we've gone from chatbots to agents, right? Like, the, we started off with ChatGPT where you're asking questions, getting answers, which is great. But now it's about the models becoming agentic, where they can execute complex, multi-step workflows in a real-world business setting. And the type of data you need for that is totally

  3. 3:46 – 16:59

    Data, Compute, Algorithms: What is Most Abundant?

    1. JS

      different.

    2. HS

      How is that different? That's so interesting. So in the transition from chatbots to agents, how does the data required change with that transition?

    3. JS

      When you are training a chatbot, uh, you'd usually do a lot of, uh, SFT and RLHF. Uh, SFT, where you're g- giving the model input prompts and output completions. You kind of teach the model to imitate experts. Uh, with RLHF, like, you're basically, uh, teaching the model to produce responses that a human would tend to prefer. Uh, RLHF is used to train what's called a reward model, and then the model is trying to produce completions that, uh, give it a high reward, right? With agents, and let's define an agent, I mean, different people define agents in different ways.

    4. HS

      Sure.

    5. JS

      Um, I would define an agent as, um, something that's capable of taking action, um, in the real world or in the physical world. Uh, something that's ex- executing a multi-step workflow, calling different functions. Um, uh, the agent could be, um, o- operating a computer, uh, or making backend API calls to actually do stuff, right? Uh, you might have an agent to file your taxes. You might have, uh, an agent to prepare your monthly financials. Um, so to train an agent, um, you would also want to teach the model how to do tool use. So you teach the model how to call other functions, how to use other applications-

    6. HS

      Yeah.

    7. JS

      ... to be more leveraged. Today, the dominant paradigm is reinforcement learning. Um, and oftentimes, these agents are trained through reinforcement learning where you'd build what's called an RL environment, which is like a mini world model for business.

    8. HS

      Yeah.

    9. JS

      Um, and in that RL environment, you have input prompts, output verifiers, and you'd have the full system state, uh, tracked along with the data model. Let me give you an example.

    10. HS

      Yeah.

    11. JS

      Imagine a workflow for a salesperson that an SDR would go through, uh, where before a sales call, let's say the salesperson has to research the prospect, uh, look up Salesforce to see whether somebody from the team has already spoken with this human, uh, and maybe, if needed-... uh, looking up this person's contact information, maybe using ZoomInfo or something like that to reach, reach out to them, right? This required this human to use three different tools, LinkedIn, Salesforce, and ZoomInfo, right? In an RL environment setup, you'd create a mini world model with clones of these applications that are created with a fake database, uh, with, like, synthetic data. And the prompt might be, "Hey, um, prepare, uh, for a call with this person and then after the call is done, update Salesforce," right? Like, let's say that's the prompt. Uh, and you have what's called a verifier to check whether the agent completed the task, right? Um, and in this case, and this is where I think it's, AI is kind of beautiful and somewhat magical, you set up the agents in this environment, and the agent is gonna try different trajectories, try different tool calls to try to complete the task.

    12. HS

      Uh-huh.

    13. JS

      And you would set this up so that, uh, the curriculum is optimally defined. Uh, if you create the c-... The curriculum is the set of tasks that you has, have this agent do. If it's too easy and the agent completes everything, the model doesn't learn much.

    14. HS

      Yeah.

    15. JS

      If it's too difficult, the model doesn't learn much. You'd ideally want, like, the, the right mix where the model is kinda getting positive and negative feedback. Um, it's very similar to, um, the technique that AlphaZero used in mastering Go when the model played against itself. So this is, like, another... It's kind of like a form of synthetic data because the agent is trying different approaches by itself, but it's humans. I mean, in this case, at Turing we create these RL environments at massive scale for every workflow you can think of, across every function, across every industry.

    16. HS

      And so you create the RL environments to create the data that then allow the models to train to have further use cases like that.

    17. JS

      Correct.

    18. HS

      Got you.

    19. JS

      Correct. And we create RL environments for every industry you can think of. Retail, healthcare, life sciences. Imagine, like, this four-dimensional matrix where the first dimension is every industry, like financial services, retail, healthcare. Podcasting maybe is one of the dimensions.

    20. HS

      Please.

    21. JS

      (laughs) The second dimension could be every function, software engineering, marketing, sales, finance, et cetera. The third dimension could be every role in that org chart. Let's say in sales there was SDR as, like, a role. The fourth dimension is a workflow that a human goes through in that role. You can think of every role a human has as, like, a composite of workflows, right? And we are creating RL environments for every workflow, for every role, in every function, in every industry. That's like $30 trillion of knowledge work.

    22. HS

      Is that possible to have that breadth and quality?

    23. JS

      Yes.

    24. HS

      How? (laughs)

    25. JS

      Time and lots of money.

    26. HS

      (laughs) Because I, I was speaking to, Candidly-

    27. JS

      Yeah.

    28. HS

      ... one of your competitor's board members the other, the other day in prep for this. And he said, "The big thing that we all got wrong was we are so in innings one of the acquisition of verticalized data." Like, there is so much room to run in the data acquisition for dental, for SDRs, for product managers, you, you name whatever function you want. Do you see us very much in innings one of the data acquisition for these very specific vertically focused workflows?

    29. JS

      Absolutely. It's innings one, and I believe in slow takeoff. I'm sorry to pour cold water on all the AI doomers that might be listening to this, but we are not in a rapid takeoff scenario. I believe in slow, steady takeoff for, uh, AGI and eventually superintelligence. So we're still in innings one. It's gonna take a while before we get all of this data, uh, into the models.

    30. HS

      Uh, so when we think about then the breadth that we go after, um, us specializing in RL environments, just so I understand, like, the marketplace that we sit in, 'cause there is your Macaws, there's your Serges. How do you differ for those, for people who are wondering, "Hang on a minute, I thought they were all one?"

  4. 16:59 – 20:57

    The Biggest Challenges Enterprises Have with AI Adoption

    1. JS

    2. HS

      And, uh, my question to you is, I, uh, I spend a lot of time with very large companies, mostly where I speak to them. The thing that I'm astounded by is we hugely underestimate the pace of AI progression in terms of the technological capabilities. But consistently, what I see is the laughable state of internal data, internal processes. I mean, Jonathan, these guys are so far off adopting Slack and Notion, let alone building custom models and embracing the latest AI tools. I, I, I respectfully push back on the all-knowledge were automated. Maybe in 20 years, but not in a 10-year timeframe. Am I wrong?

    3. JS

      Uh, w- what do you think is the biggest constraint or obstacle?

    4. HS

      Uh, the inability for them to try and implement new tooling.

    5. JS

      But what if the cost is too high? If they didn't do that, if the hypothetical insurance company that I told you about, if there was a competitor of theirs that could operate with one-hundredth the headcount-

    6. HS

      Mm-hmm.

    7. JS

      ... while delivering a better experience to their customers by pricing insurance deals better, making more money from, you know, insurance premiums, less claims payouts, they'll get their lunch eaten.

    8. HS

      I think they will. I think you'll see this, like, transfer of value from old incumbent who can't adopt new tools to startup company, hence why we invest, who is eating their lunch, absolutely.

    9. JS

      Ah, I see. So your theory is that the incumbents won't adapt and it'll just be a forest fire and you'll just get-

    10. HS

      100%. We will be on a 10 to 20 year decline of incumbents who are unable and unwilling to adopt new tools because of data, because of permissioning, because of internal buying processes.

    11. JS

      I, uh, that's an interesting point, Harry. Um, I-

    12. HS

      I mean, like, go to a European bank.

    13. JS

      Mm-hmm.

    14. HS

      You will be astounded by how bad it is internally. Respectfully, the poor quality of the processes of buying technology, it's just, it's, it's abhorrent.

    15. JS

      I have a hypothesis. My hypothesis is that companies will be very slow with back-office automation. But in the front office, for example, I speak with financial services clients in New York, like some of the biggest, uh, biggest companies. I speak with the C-suite of these companies and they are extremely interested in applying AI to help them make better investment decisions because it directly translates into helping them make more money.

    16. HS

      Mm-hmm.

    17. JS

      And I've found that it's a lot easier to, like, convince people to use a piece of technology to make more money than to save money.

    18. HS

      Yeah, I hear that.

    19. JS

      And in financial services, it's pretty brutal, right? Like it's kind of an efficient market. If you don't, if there is alpha to be found in, like, how you can allocate capital better, make investment decisions better, figure out what opportunities to invest in, price deals better, you'll get killed if you don't, if you're not at the bleeding edge. And I've heard, uh, Mark Chen, the head of research at OpenAI say this, about how, um, financial services is usually at the bleeding edge among all the other industries in the S&P 500. Uh, but even they are, like, about two years behind usually, like relative to the state-of-the-art. Um, so I agree with you that in back-office automation, it'll probably be very slow and it'll, it, it'll probably be the upstarts that'll do things well. Uh, I think the change management will be too slow. Um, but I'm optimistic about front office, uh, especially in financial services, life sciences, pharma, where if you can, like, accelerate the time to discover a drug or to get to a molecule, like there are like these, um... If you can help somebody win in, like, the main thing that they care about in their, in their industry, uh, I think they'll adopt it

  5. 20:57 – 28:53

    Why Will 99% of Knowledge Work Will be Gone in 10 Years

    1. JS

      faster.

    2. HS

      I'm always told by a dear friend, Rory O'Driscoll at Scale, I don't know if you know Rory, but fantastic investor, and he always says to me, "Listen, value generation from AI is fundamentally dependent on one simple question: Will we see the transfer of budget from human labor to AI technology? And if we see that transfer of budget, oh my God, the $30 trillion that you said," (clicks tongue) it does. "And if we don't, we operate in maybe a slightly larger software technology budget world, but by no means a world that we can have the valuations and the money that we have going in." When you look at that, are there any areas truly today where you're like, "We have seen the full transition from human labor budgets to AI technology budgets"?

    3. JS

      I think the transfer is pretty high in areas like, uh, customer support, um, copywriting, um, SEO, like some of these marketing-related, uh, areas. As you would expect, the transfer is faster in these low-risk-to-fail areas.

    4. HS

      Yeah.

    5. JS

      Like when it's, where it's relatively easy. Uh, but I'd en- encourage your listeners to look up, uh, GDPVal, which is this, uh, paper by OpenAI where they measured the impact of today's AI models in automating all types of economically valuable work. It's a lovely piece of research, I'd encourage everybody to read it, where they did this study where they looked at, I think, like 44 different, uh, uh, nine verticals and 44 occupations. They took, like a very diverse sampling of different types of knowledge work, everything from financial services to real estate to, um, uh, healthcare to, to law, and they took very specific occupations and in those specific occupations, they took real tasks where like a real deliverable has to be produced. Imagine an engineer, like a civil engineer, creating a blueprint for a building they're about to build, or somebody who's at the set of a movie studio coming up with a schedule for how you'd organize your crews. Like real work, right? And for coding, you can imagine, like a real-world software engineering, uh, project. Um, and they saw that today's models were quite good at achieving parity with, uh, the best human experts in that field. Uh, we were roughly at like, um... I say we, I don't know which side am I on? (laughs) I'm on the side of the AIs or the, or the humans?

    6. HS

      I think you're on the side of the AIs.

    7. JS

      Oh, really?

    8. HS

      Yeah, yeah.

    9. JS

      I mean-

    10. HS

      From that positioning, that would infer so. Yeah.

    11. JS

      Yes, yes, but Turing is like a blurry line, right? Like there may be, may be, I mean, as the CEO of Turing, like passing the Turing test is about not being able to tell the difference.

    12. HS

      (laughs)

    13. JS

      (laughs) So, um, so what I've s- noticed was, um, about 50% of the time in, in GDPVal, like the best models were producing work, uh, that was indistinguishable from a human expert, which is remarkable. And kudos to OpenAI where they also flagged that the number one model was Claude 4 Opus, um, although GPT-5 was quite good also. Um...And, uh, and, and this was for relatively simple tasks like imagine just, uh, a task requiring one single step. Whereas in the real world, if I give you a task to do a certain project, you won't just go off and do it. You might ask for clarifying information, you might do other things to acquire more context. You might go brainstorm with other, other humans to do that task and you would do it in a sequence of steps. So there's more room to go, um, but I think we are well on our way to AI eventually automating all types of knowledge work.

    14. HS

      What happens in that world? If AI automates all types of knowledge work, what, what happens then?

    15. JS

      Three things that will happen. Uh, first, I think we are all going to be... we'll all have the potential to be 100X more productive. Today, I'm able to run one company. Elon can run maybe, like, five companies. But in a world where I'm a 100X more productive, maybe I'm able to run a hundred companies. Elon maybe runs 600 companies. Um, I think every human will just be so much more leveraged. Um, the nature of a job itself could change, where there could be... Today, we are accustomed to the idea of, like, one person doing one job, um, but people could be doing multiple jobs at the same time. People could be running different companies at the same time. The second implication, I think, is it's gonna be wonderful for entrepreneurship. So you, Harry, I think are gonna be very happy because today, if I think of, like... For a lot of ideas, founders are intelligence constrained, um, I think of being capital constrained as a form of being intelligence constrained. For example, if you pick a, a therapist who wants to start a mental health startup, today that founder would have to raise at least a few 100K if not a few million to recruit some software engineers, maybe a, uh, a marketing person or a growth person, maybe a product manager. But in the future where AGI exists, this person will recruit, uh, a marketing GPT, a software engineer GPT, uh, a PMGPT and get off the ground for a lot less capital. A million flowers will bloom. Lots and lots of non-technical founders will start companies. We'll see a broader distribution of founders than just those who live in London or Palo Alto who are kind of connected to these pools of capital who might start companies, which I think is wonderful for the world.

    16. HS

      Do you think we will? What I mean by that is there's six and a half million people today in the UK as a working population who actively do not work because of an inability to work. Uh, I'm not gonna get into the analysis around that because I'll get in trouble for it. Um, I think we grossly overestimate the intelligence of the general population. And I know that sounds incredibly arrogant, but most people say actually, a lot of people just don't want to work and are not at the level of recruiting GPTs assistance. Do you not worry that it will widen the chasm between those that have and those that haven't?

    17. JS

      I'm an optimist and I think the opposite will happen because what we're really doing when we're training super intelligence is you'll be basically training intelligence as an API, and what's the alternative to that? It's like hiring a human to provide you with that intelligence. That human is quite expensive, right? And if anything, that creates an even broader gap between the haves and the have-nots. Whereas for $20 a month, if you ha- if you had access to the smartest experts in coding, in STEM, in sales, in marketing, I feel like more people will be able to start companies and produce active, uh, like valuable work. I also think, uh, I believe that when we have access to super intelligence, I firmly believe this, we are not all going to chill out on a beach somewhere and contemplate what do we do next. I think we're gonna solve... I mean we humans, like, we are tool builders, we are problem solvers. We'll solve problems at higher and higher levels of obstruction and I feel like in a world where we have AGI, we'll just solve much more exciting problems. Maybe we'll cure diseases, reverse aging, maybe go to the stars. There's all sorts of fun things we'll do. I don't think we'll be bored.

    18. HS

      (laughs) I'm glad. I, I, I often hear the UBI and we're gonna sit and write poetry and I'm like, "I think that might be a little bit challenging."

  6. 28:53 – 34:20

    How Will Data-Driven Feedback Loops Replace Technology as the Moat

    1. HS

      Um, when technology is not the moat, what is the moat? I had the founder of Base44 on, um, and he said, "99% of code in the next year will be written by AI. Technology is no longer the moat." What is the moat in that world?

    2. JS

      I think one moat will be data-driven feedback loops.

    3. HS

      Mm.

    4. JS

      Um, for example, uh, one reason Google had such a great lead in search for a while, um, e- w- was the, these data-driven feedback loops that come from people using your product and generating data that gives you, the algorithm developer, a high quality gradient for which direction to step in, right? So page rank, the importance of page rank was known. The recipe for ranking search results were, was well known among Google, Yahoo, Microsoft, and a few others. Obviously, people move around these companies all the time, but the advantage Google had was because everybody preferred Google and liked, liked working, u- liked that search engine, um, the... Y- you saw a much more representative set of queries. Uh, you had data from clickstream, to, from the clickstream of what results people were clicking on. That helps your algorithms improve at a much faster rate. I think data-driven feedback loops will be key for all types of enterprise applications also. Um-Today, OpenAI and ChatGPT has a good data-driven feedback loop. In enterprises, again, I think it's wide open. Like, whoever is deploying the right custom fine-tuned models and agents for specific workflows or roles or functions or companies, if you get in first and solve a customer's problem really well, you start getting that flywheel going where you will discover first where the models don't work well, and you will use that data to work with a company like Turing to generate additional data to plug that gap so you will improve. And this is what I mean by it's important for the models to touch reality. I feel like the models have touched reality in consumer, we haven't yet touched reality in enterprise. And the only way we'll improve is by deployment.

    5. HS

      And that deployment is fundamentally predicated on handholding, correct?

    6. JS

      Yes, handholding, and I feel like there's still a lot of first-mile schlep and last-mile schlep that needs to be handled.

    7. HS

      What does that mean, first-mile schlep and last-mile schlep?

    8. JS

      When I say first-mile schlep, I mean, um, for that underwriting copilot example that I gave you for that insurance company, I painted a pretty rosy picture of, like, how you take this model, you fine-tune it on your proprietary underwriting data. In the real world, it doesn't work that way.

    9. HS

      (laughs)

    10. JS

      First-mile schlep is, let's say I'm talking to the CEO of this insurance company or the CTO of this insurance company, they'll say, "Our data is a mess. It's in silos. It's super fragmented. Some of the data is in spreadsheets. Some of the data is in a file that Bob has, and Bob doesn't work here anymore," right? The data kind of is all over the place. You first have to, um, acquire the data, convert the unstructured data into structured data, into a format to fine-tune LLMs. Um, you might wanna set up good infrastructure for evals, you'd wanna, uh, cre- create good evals for the models or agents. Um, you might wanna build a workflow design for partial autonomy. So this human underwriter that's about to, um, use this model to evaluate, um, these medical histories, you might wanna build a cursor-like interface for them so that they can work alongside the AI to do their job. Also, training the humans in these new workflows, um, you, uh, wanna make sure you're collecting data the right way. Um, you'd, uh, the way, for example, we do deployments is, uh, like a tandem system where you'd have a human, uh, and an AI doing the same job for a period of time, where a manager can see the output of both. If the agent is right and the human is wrong, you train the human. If the human is right and the agent is wrong, you've created a data point to fine-tune the next iteration of the agent so the agent is steadily improving over time.

    11. HS

      If the agent is right and the human is wrong, why don't you just fire the human?

    12. JS

      It depends on at, like, what, uh, frequency, right? Like, you, you track things like precision and recall, like, um, uh, like, you, you'd, you'd, you'd wanna analyze this over a period of time. Like, you wouldn't fire them over, like, a single mistake.

    13. HS

      Bit harsh.

    14. JS

      Yes.

    15. HS

      What's the margin on that business?

    16. JS

      Uh, it varies. Like, we're also in the early innings of, uh, figuring out how to, um, how to price that. Uh, today, like, we do it in a relatively simple way where we're just billing these things for time.

    17. HS

      Yeah.

    18. JS

      I don't think that's the right way to do it. Like, we'll switch to a more value-oriented pricing model at some point. Um, right now, we're just laser focused on, um, the Frontier8, um, labs. So en- enterprises for us is, is a longer term play.

  7. 34:20 – 43:43

    Is Revenue BS in Data Labelling? Are Players Calling GMV Revenue?

    1. JS

    2. HS

      When you look at revenue numbers in this space, a lot of people shout back, "They're not revenue numbers, they're GMV." And given our understanding now that buttoning there's no talent acquisition from your business, it's all an RL environment creation business, when you look at the other announcements of alternative providers, can you help me understand, are they revenue or are they GMV? And is there mislabeling being done here?

    3. JS

      I don't wanna comment on other companies, but, uh, we think about revenues differently. Like, the way, um, we think about revenues is, is more in terms of, um, uh, gap revenues, in terms of...

    4. HS

      It's like traditional revenue numbers. I'm an investor, Jonathan. Essentially, I'm trying to understand and learn from you of how I should weight revenue in today's AI world versus a previous historical world. Like, should I be impressed by these revenue numbers or should I not?

    5. JS

      Yeah, I think it depends on the type of revenue. I think the, um... Obviously these are not SaaS ARR numbers, right? They're, these are not those types of revenues. Um, th- this is a different beast. I think this requires thinking from first principles. Um, there is, um, the revenue here is, um, reoccurring in the sense that oftentimes when you're working, uh, with a lab on helping, um, helping the models improve in some area, and I'll speak to Turing, I don't wanna speak to other companies. Um, when we are helping a lab, let's say improve their models for coding or multimodality or tool use or are working on RL environments for automating all types of professional knowledge work, it's usually a reoccurring project.

    6. HS

      Mm-hmm.

    7. JS

      Where there is, uh, projects will start, projects will end, um, and as long as you're doing a good job, uh, there is lots and lots of demand. Um, but you have to consistently keep doing a good job.And we also take, it's also important to be a trustworthy partner to the labs. Like, we take secrecy very seriously.

    8. HS

      Hm.

    9. JS

      Like, uh, we have, we make sure that, um, our projects are all firewalled, uh, between labs. Oftentimes, even with teams within the labs, like sometimes, like, that's the level of secrecy that you'd need. I'm reminded a little bit about, um, how I've been told Foxconn operates. I mean, I don't know anything about that. But I've been told that, like, they have different floors where, you know, maybe they're, on one floor, the iPhone is getting made, and on another phone, another floor, maybe a Pixel phone is getting made, and obviously you have to like firewall all of that.

    10. HS

      Of the eight largest providers, do they not spend with all of you?

    11. JS

      They spend with a handful of companies. Uh, they do that for, uh, to have some level of, um, uh, resilience, um, and I imagine there's some price benefits to having more than one person that they could work with. But I think the resilience piece is important. I mean, we know what happened with the, you, you know, when the Scale investment happened, again, like the, it, it was, um, uh, the labs did benefit from having other partners that they could work with. I would say it's a small handful. It's a small handful that are trusted, and of course there's a, there's probably a giant pool of smaller startups, but it's a small handful of big companies in this space.

    12. HS

      Which one do you worry about most?

    13. JS

      So this is just a big, big market that's growing super fast. I'm excited for, like, all the companies in the space. I feel like different companies come into this world with, uh, a different DNA.

    14. HS

      Which leader do you most respect?

    15. JS

      Sam Altman, Elon Musk-

    16. HS

      Of the data providers, Jonathan.

    17. JS

      Of the data provid-

    18. HS

      I'm pushing you here.

    19. JS

      (laughs)

    20. HS

      I, I, I, get, th- this one I, I, I'm gonna get a name.

    21. JS

      I mean, I have a lot of respect for, uh, Alex Wang, um, from Scale AI. I feel like Alex and Scale, uh, were prescient in seeing the, uh, importance of data, um, and I admired how, uh, having started in autonomous labeling, like how they kind of navigated the ups and downs. Um, and, um, yeah, I really like the way he operates as well. I feel like there are, there are certain elements of leadership that, um, that I think, um, I share with him. So, and I think he did, um, he did a great job for, for Scale.

    22. HS

      How did Scale being acquired impact Turing's business?

    23. JS

      We just got flooded with a lot of demand.

    24. HS

      Yeah.

    25. JS

      Uh, a lot of demand, and we've also amped up pretty significantly in multi-modality. Um, multi-modality was something I think, uh, Scale was quite strong in. Mo- multi-modality is, uh, teaching the models to operate well with, uh, not just text but audio, video, image, et cetera. And, uh, outside in, I've heard that, um, b- because of their roots in autonomous labeling, like, they're, they were, uh, quite good in, uh, in multi-modal stuff, and the, uh, uh... yeah. It was, it was good, uh, primarily from just increasing demand, and, um, I think it, I feel like, um, they were the company that had been working in this space the longest.

    26. HS

      Do they have a business left? Again, I mentioned Roy O'Driscoll. I think he said in a show with me recently that there kind of, there's this kind of carcass or husk left behind. But, w- if everyone benefited from their being bought, they, they can't be doing that well.

    27. JS

      I, I don't know enough about, uh, their business. Like, in terms of, um, how they're work-

    28. HS

      Do you pay attention to competitors?

    29. JS

      I pay attention to, uh, competitors in terms of, um, things that they do well and when there are, um, any significant learning opportunities, um, from them that could help us serve our customers better.

    30. HS

      Do you worry about revenue concentration? You said about kind of eight of the biggest labs. Uh, s- say if you look at, like, an OpenAI, they have, I don't know, whatever it is, uh, a hundred million? You, you will know these numbers much better than me, but say a hundred million paying customers, I'm just taking it 10% on a billion people, but give or take 100 million, whatever. Um, and then you look at, say, a business like ours here where there's, like, seven core customers. H- how do we feel about revenue concentration?

  8. 43:43 – 52:22

    Are We in an AI Bubble?

    1. JS

      learning.

    2. HS

      So, I'm worried that we are potentially not gonna see AI deliver the immediate revenues that we've promised and we're gonna go through a kind of cooling period, which everyone kind of suggests and thinks that we're gonna go through in the next six to 18 months, which is, as I said, it doesn't hit the revenues that we said it would and kind of the AI bubble kind of deflates slowly. To what extent do you think that's possible? Or we'll see this continuing gradual increase as we kind of touched on there?

    3. JS

      I don't see an AI bubble. Like, I feel like these models are incredibly powerful today. Like, GPT-5 is, like, fucking awesome. I don't know what people were talking about when they're just talking about... You know, I know there was some chatter. I think we've just gotten used to magic and it's like the, some of the... Uh, I feel like, A, these models are incredibly powerful today and they're the worst they'll ever be. They're only gonna keep, keep improving. Um, and I say that about, um, the Gemini Pro models, the Grok models, the Claude models. Like, these models are amazing.

    4. HS

      Yeah.

    5. JS

      And there's a very significant model capability overhang. By that, uh, what I mean is the models are capable of X but what we are getting out of the models is X minus delta. Um, there is... Uh, with the right agentic scaffold, uh, around these models in terms of the right system prompts, the right user prompts, giving the models access to the right context, teaching the models how to acquire additional context, teaching the models how to use the right internal tools, there is significant amount of capability that can be unlocked with today's models. For example, you, Harry, I imagine when you do an interview with, uh, with somebody, one of the things you probably do is you apply your secret sauce to pull out the right clips from the interviews, so, like, what to highlight, what are the catchphrases, what will sort of, uh, drive more engagement. That can be done by a model with the right agentic scaffold that's fine-tuned on all the work that you've done in the past. And you might not-

    6. HS

      I wish (laughs) . Every weekend, I go through every single show-

    7. JS

      Yes.

    8. HS

      ... and I pick out 15 to 20 clips per show, and then I make notes on each one.

    9. JS

      Yeah. Are you saying, Harry, that you wanna use Turing?

    10. HS

      I, I... If you could fucking make it work, dude, I'll pay you a lot of money (laughs) .

    11. JS

      S- when we... When we... If we have... Maybe we should, we should partner.

    12. HS

      That'd be great.

    13. JS

      Yeah.

    14. HS

      Ser- seriously, every weekend, I spend probably three hours per show through... Uh, definitely 12 hours a weekend doing that.

    15. JS

      So, I think there is this model capability overhang where the full potential of the model has not been unlocked by humans yet. And no, I don't think there's an AI bubble. I think the... I think there are some growing pains, but-

    16. HS

      What are the growing pains?

    17. JS

      I think everybody, like, keeps citing that MIT report about, uh, how 95% of pilots fail. But because we are in the business of deploying AGI in enterprises, I can tell you, like, why I think that happens, which is one of the growing pains. Um, step one is that most enterprises need to do some work to structure their data in the right way. Again, that first-mile schlep has to be done. Second, you should, uh, surround the model with the right agentic scaffold that I just described. Um, the right prompting, the right context engineering, the right internal tool calls that you should teach the model to call. All of those have to be, uh, distilled into the models.

    18. HS

      Mm-hmm.

    19. JS

      You need really good evals. You also need a workflow design for partial autonomy. Uh, Andrej Karpathy, like, articulated this first, um, when talking about why Cursor works so well, because it's not designed for full autonomy. It's designed today for partial autonomy, for humans to collaborate with the AI to do that specific task. So that Cursor for X needs to be built for every role, for every workflow to, like, help humans work more easily with the models.

    20. HS

      Does every role need to go through that pathway of Cursor for X before it goes to full autonomy? Or there're some roles, like customer service, where it just goes to full autonomy?

    21. JS

      I think for some roles, uh, where the... where you can see that the models are quite good at matching humans, I think, uh, we don't need that intermediate step. They, uh... There are certain roles where, um, by virtue of how the models are trained, where they're trained with... they're pretrained with tokens on the internet and then of course with, uh, talent from, um, research accelerators like Turing that's fine-tuning the models. But there are certain types of roles where the tokens from the internet give it sufficient intelligence to do the job well. Customer support is an example. But if you pick...... other roles, like for example, if you picked the role of, um, let's say an AI researcher, or you picked the role of a lawyer specializing in, um, venture capital financing-

    22. HS

      (laughs)

    23. JS

      ... uh, it's possible there's not enough of those tokens on the internet, so the models will be relatively weak there out of the box. And also, the way, I don't know, the way a Wilson Sonsini does financing might look different from the way, uh, a Cooley does it. Maybe they have their own way. So you might want to fine-tune them on your own proprietary data, distill the proprietary intelligence of humans working there. Um, so, so, I would... So, so for those things, like you may need to do some fine-tuning. The models may not work very well out of the box.

    24. HS

      A lot of people suggest the circular deals between some of the large providers suggests the strains in the ecosystem or the bubble-like tendencies. Do you think that's fair or not?

    25. JS

      I sort of categorized the world into two, two, um, two classes. Um, class one is those that believe in AGI that we... uh, I mean, let's call it the AGI-pilled group that believe that we are on the path to getting to, uh, AGI. And let's define AGI as an AI system capable of at least matching humans in almost all types of intellectual knowledge work.

    26. HS

      Correct.

    27. JS

      Right. Then there is a world, there is a, maybe a, uh, another category of people that don't believe this will happen, we'll hit a wall, right? And in the past, there have been other AI paradigms where we did hit a wall. Um, so for the camp that believes in AGI, and I believe in AGI, um, unsurprisingly, um, but I love AI and, uh, it's been my passion for like the last 20 years. The... I really believe that we will get there. If you believe that, the grand prize is so amazing, right? Like if you've solved intelligence, you've solved all of humanity's grandest problems, from curing diseases to potentially pausing aging to like interstellar travel to energy, like all of our problems are intelligence constrained, right? So if you've, um... So the prize is so large, I mean, whoever wins the superintelligence race will probably win search, will probably win consumer devices, will probably win operating systems, will probably win, uh, cloud, like business productivity software. It's like the prize is so massive that it's worth placing big forward bets in these areas because the cost of not winning is too high. Whoever wins AGI would also probably win social networking. So you can see why the big eight are excited about it, because it's... you're playing for everything. It's like whoever wins this could be responsible for that $30 trillion of knowledge work with full-

    28. HS

      Well if you were Z- if you were Zuck, you spend $100 billion on it. If you lose or fail, likely everyone else will fail, in which case you're behind just like they are and you've lost $100 billion, which isn't a huge amount of your free cash flow, maybe 12 to 18 months of free cash flow. If you don't spend that $100 billion and someone else does and wins, you lose $2 trillion, $3 trillion of market cap.

    29. JS

      Correct.

    30. HS

      You have to play. (laughs)

  9. 52:22 – 1:00:32

    Why is SaaS Dead in a World of AI?

    1. JS

    2. HS

      Do you agree then with the notion that if you believe in AGI, you cannot be investing in SaaS apps?

    3. JS

      SaaS as we know it, I think is over. I feel like quite a few SaaS apps, um, were built at a time when, um, software was relatively hard to build, uh, and complex to build. Imagine if you were building some customer support, um, software, some customer support bot. To build a company like that, you would have had to hire some Stanford PhDs in NLP, you'll collect data for six months, you'll have like, you'll use like a support vector machine or neural network that will kind of sort of work and then you'll deploy it and you'll grind away for a while. There is a significant amount of capital that needs to be invested to get an app like that to work well, so it made sense for many companies to not bother doing that if that's not their core business, let me just use like some third party SaaS app. Uh, now, many of these AI applications are incredibly easy to build on top of these LLMs, so I feel like most companies will start building custom software super easily. I mean, we help companies build some of these custom apps, and the bar to create many of these apps significantly comes down. So that's one risk. Risk number one, companies do it themselves. Risk number two is that the... you, you get sonic boomed by the foundation model companies.

    4. HS

      Sonic boom by meaning...

    5. JS

      So today, like-

    6. HS

      They move into the apps layer and just create it themselves.

    7. JS

      Yes, like the... it, it, I mean, you, it could happen, right? Like the, you could, to... the models are becoming agentic. I mean, you, you've seen many of these agents, right? It's, it's fundamentally, it's about computer use agents. If the models get better and better there, it's possible like the, the model is all you need, like the... I mean, imagine if you wanted the model to, um, hey, update my, um, update my, um, l- let's say I'm doing some HR thing. Hey, update my, um, medical, uh, benefits information to add, uh, we have, we've just had, uh, a, a new daughter, we wanna, uh, we wanna update my medical information.If the model is agentic and it is sufficiently integrated into the, um, knowledge, into the database of the organization, you could, you don't need anything else in the middle. So, so that's the second risk, the models becoming more agentic. The third, and I worry about this a lot, I feel like a lot of our software was designed to be used by humans. Uh, humans navigating a GUI and clicking around and doing things. I think that's gonna go away. Like, with a multimodal... Again, this is why I, I think of four pillars to superintelligence. It's multimodality, reasoning, tool use, and coding. Multimodality is important because we humans interact in natural language, we talk, there's video, all of that. Um, I think the future might look like some type of ambient AI that you talk to, um, that will just go and do things. Um, and maybe it'll use the GUI of the current SaaS application as, like, an intermediate step, or it'll use MCP and use tool calls and get what it needs. The GUI was, like, designed f- for a world where humans were using a keyboard and a mouse and clicking around and doing things. I think humans are, can do better things with their time than click around and-

    8. HS

      Actually, one big change that I have is I never actually type emails anymore. I use Whisperflow and it's, it's so good in transcription that it, I don't ever type emails, ever.

    9. JS

      Yeah.

    10. HS

      Now, the only trouble is everyone knows what I'm saying in my emails. (laughs)

    11. JS

      Yeah. The Whisperflow founders interned at Turing back in the day.

    12. HS

      No way.

    13. JS

      Yeah.

    14. HS

      Oh, wow. Uh, so how do I feel about it? Um, no, I, I don't agree. Um, why? Uh, because the average company today has between 80 and 100 different SaaS products that they engage with, um, to, one, just the multitude of how many they'd have to create, number one. Number two, at maintaining them, you think they're gonna maintain 80 to 100? Oh my god, you're gonna have, like, just teams and teams of people doing maintenance and updates and debugging? I don't think so either. And that is for the technology-savvy. Let's talk about every plumbing provider, uh, law firm, accounting firm, restaurant that can barely use Wix and Squarespace, let alone build out their own CRM system and POS system. Not a fricking chance. And then we move to foundation model companies moving into very vertically specific elements. You know, we're in a business that does AI for patent creation, updates, and collaboration. Uh, SAM is not going there. SAM has health, you know, solving cancer, um, you know, energy utilization. I don't think SAM's touching patent creation and updating. And so I think for, the more verticalized you go, the more defense ability you have. And so I think for those reasons, SaaS has life. Mine is a very biased perspective because it, it's my job.

    15. JS

      Mm-hmm.

    16. HS

      Is that all wrong again? Like, you're the master here, Jitendra, so, like, VC is a literally middleman, you know? (laughs)

    17. JS

      I'd say, Harry, like, you have an interesting dataset because you invest in a ton of startups.

    18. HS

      Yeah.

    19. JS

      So I would be curious looking at your sample of startups that you've invested in and just tally how many SaaS apps they use today at every stage and see if that has changed, like, post ChatGPT. Like, see, uh, uh, like, I'd be qu- my hypothesis is that today's companies use fewer SaaS apps and have fewer people.

    20. HS

      Do you think we have more or less software engineers in 10 years?

    21. JS

      More.

    22. HS

      Help me understand that, then.

    23. JS

      I think the definition of a software engineer will change. Um, a Stanford doctor who's in oncology, um, who has an idea for, like, some, you know, cancer detection-type app, now that person will be able to create, like, a very simple version of an app that somebody could check by themselves and send, um, as, as, uh, uh, do, like, a home diagnosis. Um, I think it'll, I think there'll be more software engineers because if you define a software engineer as somebody who's capable of building a software product to solve a real problem, that pool of builders is gonna expand way beyond people who've graduated with, like, a four-year computer science degree.

    24. HS

      So we have more software engineers creating more software, and the problem then becomes discovery, no? How do we sol- solve the discovery problem in, in a world of infinite software?

    25. JS

      You might have an agent for yourself that's talking to other agents on the internet, like the... Have you seen Her, the movie?

    26. HS

      Yes. Yeah.

    27. JS

      Yeah. I really think, like, we'll, we'll have, you might have, like, there's probably a, um, an, uh, an agent for Harry that's just tasked with discovering what pieces of, uh, software Harry should be using, and that might be talking to... I mean, I might have an agent that is, um, maybe talking about, uh, what types of, um, that's marketing things that I've built or that I have. And it'll be then, uh, I think we'll just be in a world where we have digital surrogates of ourselves that are communicating with other digital surrogates. Um, I, you might have an agent of yourself that's, like, discovering startups to invest in. And while you and I are chatting, there's probably, I don't know, you might be having, like, a million conversations with entrepreneurs from all over the world, who, today you're, you're kind of constrained by space and time, but, uh, in the future you'll only be compute and data constrained.

  10. 1:00:32 – 1:07:46

    Will the Phone be the Primary User Interface to an AI World?

    1. JS

    2. HS

      Do we lose the phone as the interface to this world? We obviously see SAM and Johnny.I, uh, there's rumors of pendants and some hardware devices. I'm not asking you to comment on that. I'm just saying, does the phone become, still remain the primary interface and design device?

    3. JS

      We'll have some type of a device, um, that is, that we'll carry that's, um, always on and processing multimodal tokens. For example, as I'm talking to you-If I were to envision my perfect device, it would be something that has, it should have cameras, so maybe it's a wearable as a glass, uh, or something that I'm w- uh, having on me that's processing visual input, because I wanna be able to read your body language. I might have, like, um, AirPod-like thing in my ear that's whispering to me that maybe says, "Jonathan, as you were talking about multimodality, Harry seemed less interested. His o- his body cues suggest that he was losing interest, but when we were talking about AR, he perked up." So those types of feedback and cues I think would be good. So I envision a device that, um, I think of it in terms of sensors and effectors. In terms of sensors, like, obviously it has to be listening to stuff, it has to be seeing stuff. But in terms of effectors, it'll probably also be speaking in my ear. Um, i- ideally it should be something that you can talk to and have it do things later. For example, I might say, "Remind me to follow up with Harry on that idea for, um, uh, using Turing to automate clip generation," right? Like, the, uh, so, so, so it has to, like, remember that and come back later. So I do think there'll be all sorts of new devices and glasses, hearing, like, these AirPod-type devices seem obvious. Um, there could be, like, have you, do you remember this device that was called the Meeting Owl?

    4. HS

      No.

    5. JS

      Like, during, like, the COVID era, like, the, one of the tools that sort of spiked was, like, this, uh, it was basically, like, a speakerphone for, like, having conference, um, uh, having better distributed team Zoom meetings.

    6. HS

      Okay.

    7. JS

      When somebody's talking it would focus on them, like, with, like, a camera and it would, it was also a decent speaker. Um, so I can imagine devices like that that people have, like, when, um, yeah. So, it, it's hard, it's hard to predict, but the thing that I almost feel confident about is that the phone will look so different. I mean, when we think of our smartphone, it's basically a computer with, like, a phone app in it, right? Like, the phone app is, like, the least interesting part of the phone. Uh, and I think even for, for an AI device, the, it'll probably have some phone app in it, but everything else I, I feel like will be magical. Like, I feel like I would definitely benefit from a device that's constantly listening to everything that I'm listening to, constantly processing all the video/audio input that I'm processing, and, um, something that's paging things to memory. Like, maybe it'll, like, write things down and be able to look it up later. I see it almost like an extension of my brain.

    8. HS

      Before we move into the quickfire round, I, I do just have to ask, what does your market and the data provisioning market look like in 10 years? I always try and think about, like, market composition and dynamics. Is it a winner take all? Is it a very fragmented? Is it a three or four? What does that look like?

    9. JS

      The market will reward players with research depth because, um, the pace of AI research is so rapid, like, all these RL environments, like, this, uh, spiked in, like, the last 12 months after O1 came out in December and DeepSeq came out in Jan. So now it's like, in addition to imitation learning, we are in this reinforcement learning regime. One year ago, one year later it could be something totally different. So I think the market will reward a company with research DNA, um, and it'll reward a company that can move fast and, uh, adapt very quickly, uh, because this is, I mean, when you're serving-

    10. HS

      Do you think this is a monopoly market or do you think there will be many winners?

    11. JS

      I think there'll be a few winners. Uh, I think there'll be a few winners. Um, a few because I do think for the labs it helps them to have, um, a few partners for resiliency. Um, and I imagine also for price competitiveness. I think there'll be a few, there'll be a few winners. In the realm of robotics and embodied AI we are still very early. At Turing we are scaling up on the robotics side as well in terms of data that we generate, but there's so much data that's missing that the models need to see that they haven't seen yet. I can totally imagine some newer companies also coming up that don't exist today.

    12. HS

      If you were to invest in companies in your space, where would you invest?

    13. JS

      Probably in robotics or embodied AI. The vertical stuff, like, I mean, we are scaling up pretty massively in generating data for different verticals, so I don't, uh, I don't see that as, like, a big white space. But I think everybody's relatively early with robotics.

    14. HS

      Mm-hmm.

    15. JS

      Um, and robotics is such a vast realm that there could be interesting things to do there. One way I see the space, uh, Harry, is, um, uh, again, th- think of, like, a, like, these three dimensions. Uh, the first dimension being the type of intelligence that you're baking into the models. That could be in coding, in STEM, in functional expertise like sales, marketing, software engineering, or vertical expertise like healthcare, legal, finance, et cetera. So the first dimension is the type of intelligence. I do a cross product of that with, um, the modality, so audio, video, image, computer use. So that's multimodality. That's the second dimension. The third dimension is multilinguality, like, different languages, right? And, um, the fourth dimension is different learning paradigms like imitation learning, reinforcement learning.... pre-training as well, which is unsupervised learning, and all of those may require different platforms to be built. Like, we've had to adapt our platform for imitation learning, for reinforcement learning, for multimode- modality. Um, so I feel like in this matrix, there's, like, all sorts of, um, all sorts of new opportunities that could emerge, and I only listed, like, the, the, um, digital intelligence. I didn't talk about physical intelligence, so I think robotics is, like, wide-open. I mean, the kind of data that a robot that is in someone's home is totally different from, like, a robot that's doing things in a factory, and humanoid versus non-humanoid robots,

  11. 1:07:46 – 1:17:05

    Quick-Fire Round

    1. JS

      and-

    2. HS

      I could talk to you all day. I do want to move into a quickfire round, so I say a short statement, you give me your immediate thoughts. What's one widely held belief about AI that you think is wrong?

    3. JS

      I don't think we'll see rapid takeoff. I think we'll see incremental, continuous improvement in AI. I actually think this is good for the world because if what we believe happens, which is all types of digital knowledge work gets automated, I think humanity needs time to prepare its workforce. I think we could use the extra time to upskill humans, to rethink education, to make sure there isn't massive job displacement. Um, I also think in the steady, continuous improvement in AI models, there'll be value realized every step of the way, unlike self-driving cars. I feel like people have this wrong model for AI that comes from self-driving cars, where you get it 99% of the way accurate and you, if you can't solve the last 1%, they're not useful. AGI is not like that. I think when we automate the job of an underwriter or a claims processor or, or a CEO, there's incremental value that's unlocked for every percentage improvement in the models becoming more reliable. So, I believe in slow and steady takeoff, and that's actually gonna be great for the world.

    4. HS

      You mentioned DeepSeek a couple of times. Do you think we underestimate China?

    5. JS

      It depends on who you ask. The folks that I work closely with, um, don't underestimate China. It's very impressive, like, the progress that they've made in open source, uh, with DeepSeek, uh, Kimi-K2, Qwen. These models are state-of-the-art. Um, so no, I don't think, um, at least among, like, the frontier AI circles that I'm in, I think there's, uh, there's a, uh, clear realization of, like, how close they are.

    6. HS

      The world seems to b- be moving to closed models. Is that good or bad?

    7. JS

      I think it depends on the, uh, application. Um, I feel like in the, f- firstly, in enterprises, it's often a mix of, between proprietary and closed models. I wouldn't, uh, sorry, between, uh, closed and open models. Um, we do see demand from enterprises that want either. Um, the closed models often are easier to, uh, get started with, uh, but there are some cases where enterprises prefer open models to, um, for cost, customizability. Um, and I'm talking about the small language model regime between the half a billion parameters to 10 billion parameters. I worry a little about, um, frontier models. Um, I feel like for frontier models, there is some value in keeping some of the technology closed. Um, the, just because of how powerful they are and, um, I feel like the US labs are extremely responsible and safety-conscious in how they think about training these models, deploying them.

    8. HS

      Do you think Elon is? You mentioned reading his book earlier.

    9. JS

      Yeah.

    10. HS

      Elon is often chastised for his lack of, um, care (laughs) around some of the training elements. Do you think he is, and do you think actually he'll benefit from not having that guardrail?

    11. JS

      So, I think Elon is also, like, he cares a lot about humanity too. Like, the, at least, if you've read, in, in his book, like, one of the things I recall reading is his motivation for starting, uh, for getting into AI was that he wanted an AI that was specieist and loves humanity. Like, the, he, um, that was one of his reasons to, to get into it. And, um, everything I see about, like, the, the Grok team, uh, I feel like their goals are much like any of the frontier labs, like, quite noble in terms of having this powerful AI that can help humanity understand the universe, solve some of our biggest problems.

    12. HS

      What did you believe that you now no longer believe?

    13. JS

      I used to believe that, um, to build a enduring valuable company, you hire a strong, uh, exec team and operate with a lot of leverage, basically hire strong pe- hire great people and get out of the way. I used to believe that. Now I believe you hire great people and work really closely with them and their directs and their directs and their directs, and get as close to the ground as you can.... where ground truth usually exists with the customers. The next step to customers is the engineers writing code and the salespeople talking to your customers. So, now I believe in being, basically I used to like, for lack of a better word, follow the org chart a little bit. And this was also part of, one of my learnings from Elon's, uh, biography is that he was so hands-on. Like he would be, like, walking the factory floor and asking an engineer why this door in the Model 3 has three bolts instead of, uh, instead of maybe two, right? And it is a different way to operate, like where you are in the details of the most important things that matter. Completely working in, like, a flat structure and just operating as close to the, to the ground truth as you can. Um, and generally, like, I feel like in the early days of starting Turing, like, I, now that I think about it, I may have had a, um, subconscious desire to be liked. I think I must have. Now, I don't. Like, the, now I just think about just doing things that would solve our customer's problem the best.

    14. HS

      What was the most unpopular decision you've taken with Turing?

    15. JS

      Turing is, uh, switching from a distributed team to a hub and spoke model. So we are now, uh, working from an office in San Francisco and we've recently opened an office in, uh, Palo Alto. We are gonna be opening an office in London as well.

    16. HS

      Very exciting.

    17. JS

      I mean, for some people, like, that wasn't very popular.

    18. HS

      Until you fired them.

    19. JS

      Some of them, uh, some of them left and yeah.

    20. HS

      We, we like in person. We're, we're big fans of in person here. Um, final one. When you look forward to the next decade, what are you most excited for? So, like, for me, my mother has MS. I think that we'll have some pretty groundbreaking breakthroughs in MS drug discovery that we haven't had for, whatever. That excites me.

    21. JS

      So I'm excited about AI making new discoveries and automating AI research itself, to get to a point where, um, AI is in, is in some self-improvement loop, like, where we, to, so that we could get to super intelligence faster. So automating AI research and getting AI to the point of making new breakthrough discoveries, I'm excited by that. Um, and I've always been fascinated by AI as, like, this exoskeleton that makes you a lot more productive. Have you watched the Iron Man movies, Harry?

    22. HS

      Yeah. (laughs)

    23. JS

      Right. Uh, I, so in the early Iron Man movies, like, he's wearing, like, the suit and the suit is obviously giving him superpowers, right? And then in the later ones, the, the suit is "agentic" where, like, he has these drone suits, right? Like, where there's, like, a army of his suits that go off and do things. I'm excited about a future like that where every human on the planet has access to agentic AIs that help them amplify their fullest potential. Today, Harry might have 100 ideas, but Harry is able to do maybe two of them really well. I like a future where Harry can do the remaining 98. And I like that for, like, the seven billion humans on Earth.

    24. HS

      I like that too for my weekend's sake, to be honest. Um, Jonathan, I, I love conversations which are very natural and free flowing. Uh, you can tell that I, I don't really pay much attention to the schedule. Um, but you've been fantastic. So thank you so much for joining me.

    25. JS

      Thank you, Harry, for having me.

Episode duration: 1:17:06

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode yQLOicn2vPU

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome