Mar 24, 2025 · 16m · tbpn
Brendan Foody on using AI to predict job performance and scaling from $1M to $100M in 11 months
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Mercor co-founder and CEO Brendan Foody discusses scaling the AI vetting platform from $1 million to $100 million in annual revenue, pivoting toward providing elite subject-matter experts for foundation model evaluations, and why long-term commercial value will accrue at the application layer.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 46.3% of the talking time here. How this is scored →
The hosts as informed peer 4.7 Guest teaching 5.0 Guest disagreement 0.8 The hosts pushing back 2.0
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 5:52 Dismissing legacy vendor approaches
Foody firmly counters the premise of customer churn risk, asserting that companies fail only when ignoring leading indicators and relying on legacy systems.
Hardest push from the hosts ▶ 5:02 Coogan probes cyclical revenue drop-offsCoogan directly questions the longevity of Mercor's exponential growth curve, arguing that specialized model training runs create massive revenue oscillations.
Biggest teaching moment ▶ 3:01 Evolution of the human data marketFoody details the structural transition in AI training from low-cost crowdsourcing of grammar to hiring top-tier specialists for frontier evals.
The host holds their own ▶ 4:09 Coogan cites Google arm farm data bottleneckCoogan demonstrates sharp technical grasp by citing Google's reinforcement learning arm farms to ask about physical motion data collection.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Mercor's Founding Story and Hyper-Growth Culture | 3 | 5 | 0 | 1 | The hosts set up open-ended prompts about Mercor's fast trajectory and company culture. Foody outlines the founding journey and how they keep young college-dropout employees grounded on long-term data flywheels rather than vanity metrics. | |
| Expanding Placements from Tech Roles to AI Training Data | 6 | 5 | 1 | 2 | Coogan demonstrates industry knowledge referencing Google's robotic arm grasping farm and the potential robotics data wall. Foody clarifies the shift in AI human data from low-skilled crowdsourcing to high-skill domain experts across medicine, finance, and law. | |
| Navigating Revenue Volatility and Long-Term Demand for RL Evaluations | 6 | 6 | 2 | 5 | Coogan challenges the sustainability of Mercor's revenue curve, asking if cyclical training runs will cause severe revenue oscillations. Foody reframes the dynamic by explaining that reinforcement learning makes creating human-in-the-loop evals the main bottleneck across the entire economy. | |
| Enterprise Customer Adoption and Customization at the Application Layer | 5 | 6 | 1 | 2 | Coogan questions why AI agents still cannot reliably book flights despite heavy math benchmark progress. Foody explains researchers historically prioritized PhD-level benchmarks like GPQA and IMO math over executive assistant evaluations and tool use. | |
| Evaluating and Training AI Models on Subjective Humor | 6 | 5 | 1 | 2 | Hays and Coogan probe model commoditization, humor evals, and why ex-OpenAI employees consistently launch competing foundation model companies. Foody explains low API switching costs and the ideological pursuit of AGI driving founder decisions. | |
| High-Stakes Venture Fundraising Perks and Conclusion | 2 | 3 | 0 | 0 | A relaxed, comedic wrap-up where Hays asks about wild VC perks during competitive funding rounds. Foody shares anecdotes about private jet offers and racing Ferraris in Vegas without presenting pitch decks. |
Statements from this episode (10)
Mercor scaled from $1M to $100M in revenue in two years
“And then fast forward two years, we scaled from one to a hundred million in revenue, and so we're only 21 and running this really exciting company that's working with most of the most prominent companies in Silicon Valley.”
Foody: AI data training shifted from low-skilled crowdsourcing to high-caliber expert vetting
“What we realized was that there was this really large shift happening in the human data market where large AI labs are hiring thousands of people to train the next generation of LLMs. And it used to be this crowdsourcing problem that was super low skilled, rig…”
Foody: Top AI labs are prioritizing data far beyond frontier model capabilities
“If you look at the most sophisticated labs and what they're really investing in, it's a super high caliber expert data that is like far beyond the model frontier of capabilities.”
Foody: Human data market will grow dramatically as RL requires ubiquitous evals
“My broader take on the human data market is that It's going to grow dramatically because, so we're getting to the point where RL is so effective that you can create almost any eval and it will be able to solve that eval. And so the barrier to applying AI throu…”
Foody: RL environments make model customization far more data-efficient
“And I think a big reason for this is that it's now much more data efficient to customize models with RL environments. And a lot of this, like New kind of data versus fine tuning data that people would do historically.”
Foody: AI agents will reliably handle flight booking and mundane tasks in 2025
“And so I think as that starts to happen, I would really expect that this year you're able to use operator or whatever the equivalent agent is to start booking flights and doing a lot of these more mundane tasks.”
Foody: Mercor hires Harvard Lampoon writers to train AI models on humor
“A lot of our customers at the frontier as you've seen in recent releases are starting to think a lot more about Humor and these exciting things. We have been hiring a bunch of people out of like the Harvard Lampoon and equivalent places that have these comedic…”
Foody: Single-line API switching makes pure model businesses hard to defend
“A lot of investors don't quite realize when they're not like in the code of building these products is how low the switching costs are for API, right? Like it's a line of code to switch back and forth and like see how new models are doing. And so it's hard to …”
Foody: Ex-OpenAI founders prioritize AGI ideology over business unit economics
“I think a lot of it ties to just like the ideology around AGI, right? This is the most important problem in the world. And so how do we all race as fast as possible to get there rather than a lot of the, you know, unit economics and competitive dynamics that i…”