The Exchanges

Q Um, what do you mean by that? Can you share examples?

A So, um, every knowledge worker at Turing, either in coding or any other domain, um, I want them to use AI to do the work first. So agents should create. Humans should steer. Humans should create prompts, skill files, give the agent access to the right tools, access to the right knowledge sources. Humans should create the scaffold. An AI should create. For example, if you're creating a board deck. Yeah. Don't, the human should not, the investor, let's say the, the, my head of investor relations, he shouldn't create the board deck. He should give the model access to the right sources of knowledge, access to the right tools, write a skills file to apply consistent formatting, um, give, give the right instructions on how to create the board deck. Uh, use Codex or Cowork or Gemini or Grog to build it. Um, and then when there are mistakes in it, prompt the model, um, to improve it. Sure. And then keep improving the skills file so that next time it's even faster. Basically, um, the agents create V-one of the work. Humans are verifying and iterating. Humans never create V-one. There's an interesting post from OpenAI on, um, how they used, how they used codecs to, uh, to, for one particular software project, which was agent first, human second. Um, humans were not allowed to write code directly. They can only, like, prompt agent.

AI assessment note: “For example, if you're creating a board deck.”

Q the fields, right? Where judgment counts, right? In public markets, the data is more visible. In private markets, uh, you know, you don't discover or data is not visible mostly, right? Especially at the earlier stages. So according to that thesis, the human judgment becomes less and less, uh, um, important because if AI super intelligence can make those judgments where to invest, then everybody has access to that information.

A So I actually disagree. I think human judgment is going to become even more important. I think what's happening is the floor is going up. The, for example, if a, um, If a software engineer is working with a coding agent, uh, that has now written code for a week, um, you do, I mean, reviewing the code that get, that is, that, that got written needs work. You have to know where to look, what to check. Of course, AI will also help with that. I, um, and in the, in your point about investing too, I think, I don't think these models are still at a point where, um, human input is not needed. I think of these as productivity accelerators. Sure. Like today there are things that you're using tools for to accelerate your ability to do, to do work. Um, I think Um, these frontier models are the ultimate knowledge work accelerator, so humans can solve problems at higher and higher levels of abstraction.

AI assessment note: “So I actually disagree. I think human judgment is going to become even more important.”

Q Yeah. And for our audience, if you can share, let's say, in very simple terms, uh, Training and inference has been the key and, and like say the last few years, there was huge investment in training part, but now inference is also coming up well. So probably if you can share what has been happening and where is it going in both these areas?

A Yeah. Yeah. So, um, there is like, um, There is scale up in compute for a few different things. Um, there's pre-training compute. So I would break training into two parts, uh, pre-training and post-training. In the past, it was a lot of the compute went into pre-training. Now a lot of compute goes into reinforcement learning in post-training as well. Um, especially after O-one came out and DeepSeek came out. So there's a lot of compute that's needed for still for training, right? But what we're seeing is as these models have gotten so much more capable, uh, people need so much more. Like I'm compute constrained, like for how much I want to create agents, run agents. So it's, um, Uh, so, so inference compute is skyrocketing because demand is skyrocketing.

AI assessment note: “inference compute is skyrocketing because demand is skyrocketing”

Q That's what you are hinting at, right? Whatever, uh, problem horizon that feeds into model or becomes an outcome of model. But to, to contest you, say the work that you are doing for an enterprise, the model Are coming or will come in that kind of a work. So how, how do you make that defensible and growing?

A Yeah. Yeah. We are, uh, we are building on top of the models. Oftentimes with intelligent context management, memory management, we, for the enterprises, oftentimes they want a system that is relatively model agnostic. Uh, you'd want to make sure that you have the right auditability, governance, traceability, verifiability. You want a human in the loop. You need a workflow that's constructed on top of the models. We would, we would not build our own model. We don't want to compete with our customers, but we are building around the models so that enterprises can unlock the fullest value. And in the real world enterprise, there is still a lot of last mile messiness. Where it is, there is a gap between what the model is capable of and what an enterprise is actually able to extract, uh, value out of. And it's, it's non-trivial. Like the, and even at Turing, it requires work to make sure that You're operating in the right agent first, human second way. Um, yeah.

AI assessment note: “we are building around the models so that enterprises can unlock the fullest value.”

Q And, and how did OpenAI reach out to you? Like what, what was your mode or positioning in the data space?

A So we were not a data company when they reached out. Like we, I think OpenAI at that time was looking for, I mean, I don't know for sure, like, but they were looking for who has the best coders on the planet, and it is steering. We have the, the world's best platform for software engineers, and now we've expanded that to experts in every industry you can think of, financial services, retail, healthcare, life sciences, Every, um, function, software engineering, sales, marketing, finance, every role in the org chart in that function. Um, we have, I mean, Turing's platform has the smartest humans in the world in every domain. I think OpenAI at that time was just looking for, um, somebody who can generate really high quality data for coding models really quickly. And the fact that we used AI to source talent, vet talent, match talent, manage talent meant we could scale up really fast while keeping quality high. That was the key.

AI assessment note: “we were not a data company when they reached out”

Q No, thank you. So Jonathan, I want to understand from you, Turing started in the hiring business when, when you initially started. You became a unicorn in that business, and, and then you, you started the, the business of being the provider to the Frontier AI Labs on the data side, right? How did this transition happen? If you can walk us through the journey.

A Great. Maybe for the audience, I'll give a little bit of an overview of what Turing does, and then I'll walk through the journey for how we got here. Um, so our mission is to accelerate superintelligence to drive real economic progress, and we do that in two specific ways. First, we work with all the Frontier AI Labs to advance the models in SWE, Enterprise, and Frontier STEM by providing high-quality data, evals, and aural environments. In doing that, we discover Uh, the jagged intelligence of these, uh, frontier models and agents. We discover what different models and agents are good at. We leverage those insights as a differentiated deployment partner in the enterprise. At the enterprise, we've gone deep in financial services. We work with some of the largest asset management firms in the world and build end-to-end AI systems for them. And when we build these systems, we discover where these models and agents break. Uh, we leverage those, uh, our knowledge of those capability gaps to create even more realistic data evals, oral environments for the frontier AI labs. So, so we can solve even bigger problems in enterprise, see where things break, generate even more realistic data, solve even bigger problems in enterprise, and we execute this loop. So most data companies don't see deployment, and most deployment companies don't see data.

AI assessment note: “I'll give a little bit of an overview of what Turing does”