Brandon Foodie

CEO & Co-Founder of Mercor, a company specializing in expert human data and evaluations for AI.

Claims

As speaker

Mentioned in

Key positions and views

Expertise over Crowdsourcing: The AI data market has irrevocably shifted from needing massive volumes of low-skill annotations to requiring high-quality, nuanced evaluations from vetted domain experts [20, 30].

The Future of Work is Evaluation: A significant portion of knowledge work will transition from performing tasks to designing and creating 'evals' to train and test AI systems that will perform the tasks instead [16].

Academic Benchmarks are Flawed: Current AI research is inefficiently guided by academic benchmarks (e.g., math Olympiads) that are disconnected from the practical outcomes enterprises and consumers value [4].

RL Environments are the Future: The ultimate form of AI training data is Reinforcement Learning (RL) environments, which he predicts will eventually encompass the entire economy [10, 11].

Legacy Data is Overhyped: Companies are wasting billions on traditional SFT and RLHF data, which are diminishing in utility for cutting-edge models [22, 40].

Podcast consensus on Foodie

Points of consensus

▶Mercor has experienced unprecedented and accelerating revenue growth, scaling from a $1M to a $500M run rate in 17 months, and is reportedly profitable and capital efficient [2, 26, 29, 36, 45].Apr 2026

▶The AI data market has fundamentally shifted from a low-skill crowdsourcing model to a high-skill talent sourcing problem, as advanced models require expert human evaluators to identify complex errors [14, 20, 30].Apr 2026

▶Traditional data types like Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) are diminishing in utility and are over-invested in by the industry [22, 40].Apr 2026

▶A significant portion of future knowledge work will involve humans creating evaluations ('evals') for AI systems, which is seen as a more efficient use of human intellect than repetitive task execution [16].Apr 2026

Points of debate

▶Regarding financial strategy, Foodie asserts Mercor is highly profitable and doesn't need financing for operations [37], yet simultaneously states it is 'likely' the company will raise a new round soon for the strategic benefit of market signaling [38].

▶On growth potential, Foodie claims Mercor has enough customer demand to 'double overnight' [39], but also acknowledges this growth is constrained by the company's capacity to onboard a sufficient number of qualified experts [39].

▶Concerning AI timelines, he predicts AI will be 'significantly better than humans at their jobs' within 2-3 years [32] and that superhuman AI software engineers will arrive by H1 2025 [42], while also stating that AI superintelligence is a 'longer road' and will not be achieved in the next three years [27].Apr 2026

▶On market focus, Foodie criticizes the AI research community for focusing on academic benchmarks disconnected from real-world outcomes [4], while his company's strategy is to first dominate the market serving these same AI labs before expanding to enterprise hiring [49].Apr 2026

Key themes

▶Hyper-Growth and Market DisruptionApr 2026

Foodie consistently emphasizes Mercor's record-breaking growth, scaling from a bootstrapped $1M run rate to a profitable $500M run rate in just 17 months [36, 45]. He positions this growth as a direct result of disrupting the legacy AI data market, which he claims was focused on an obsolete model of low-skill crowdsourcing [20, 30].

This narrative of unprecedented speed and profitability is designed to establish Mercor as the definitive market leader, making it a high-risk, high-reward proposition for investors betting on a category-defining company in the AI data space.

▶The 'Eval' EconomyApr 2026

A core thesis for Foodie is that knowledge work is shifting from task execution to task evaluation. He predicts that humans will increasingly be employed to create sophisticated tests ('evals') to train and validate AI systems, a more efficient use of human expertise than repetitive work [16].

This vision reframes the 'AI taking jobs' narrative into an 'AI changing jobs' narrative, positioning Mercor at the center of a new, higher-value labor market that it aims to define and dominate.

▶The Primacy of Expert Human DataApr 2026

Foodie argues that as AI models become more advanced, the need for generic, low-quality data diminishes rapidly. He claims the new bottleneck and key differentiator is access to high-quality, expert-generated data and evaluations, which is why Mercor pays experts significantly more than competitors and focuses on sourcing and vetting top talent [20, 30, 46].

This position directly critiques the business models of earlier data annotation companies like Scale AI and frames the market as a talent and quality problem, not a volume problem, justifying Mercor's higher price point and expert-centric marketplace.

▶The Inevitable RL-ification of EverythingApr 2026

Foodie holds a speculative but strong belief that the entire economy will eventually become a Reinforcement Learning (RL) environment for training AI [10, 11]. This long-term vision suggests that all economic activity and human work could ultimately serve as data to improve autonomous agents.

This grand vision provides a philosophical underpinning for Mercor's business, suggesting its current market is just the first step in a total economic transformation, which could attract long-term, vision-oriented investors but may seem overly ambitious to others.

Source episodes

Sentiment over time

Not enough data for timeline

Changes over time

Pre-January 2023

While its founders were still in college, Mercor was bootstrapped to a $1 million revenue run rate and achieved profitability [24, 50].

January 2023

The company's 17-month hyper-growth period begins, marking the start of its scale from $1M to $500M in revenue run rate [45].

March 2023

An initial prototype of an AI interviewer built on the newly released GPT-4 is not functional, highlighting the early state of model capabilities at the time [43].

August 2023

The founders meet with the xAI team at the Tesla office, indicating early traction with major AI labs while they were still college students [50].

Undated (Post-Aug 2023)

Mercor's business quadruples following a competitor acquisition by Scale AI, which Foodie identifies as a market tipping point [47]. At its Series B, the company was valued at $2 billion on a $20 million revenue run rate [33].

Present

Foodie claims Mercor has reached a $500M revenue run rate, is profitable, and is growing faster than at any previous point in its history [2, 37, 45].

Suggested prompts

How does Mercor's reliance on a limited pool of high-end experts present a scalability risk to its claim of having demand to 'double overnight'? &nearr;Given Foodie's critique of academic benchmarks, what specific, proprietary evaluation frameworks has Mercor developed to prove superior AI performance on real-world enterprise tasks? &nearr;What are the potential conflicts of interest or data privacy concerns when the entire economy is conceptualized as an 'RL environment' for training AI models? &nearr;How does Foodie reconcile his view that general AI API capabilities will be commoditized with the massive valuations of the foundation model companies that are also Mercor's key clients? &nearr;

Key concepts

Mercor's Growth 3 ep AI Data Market Shift 3 ep Evals (Evaluations) 2 ep Reinforcement Learning (RL) Environments 2 ep Future of Work 2 ep RLHF / SFT Data 2 ep Hiring & Talent Sourcing 2 ep AI Model Capabilities 3 ep Market Competition (Scale/Surge) 2 ep Business Strategy & Vision 3 ep

Notable quotes

“what we're sharing imminently is that we scaled the business from one to 500 million in revenue run rate in the last 17 months”

Brandon Foodie · Mercor CEO & Co-Founder, Brendan Foody: How They Grew from $1M to $500M in 17 Months

“I actually think it highly probable that a huge chunk of knowledge work just trends towards creating evals”

Brandon Foodi · Mercor CEO: Evals Will Replace Knowledge Work, AI x Hiring Today & the Future of Data Labeling

“one of the largest inefficiencies in all of AI research is that the evals that people have been going on of humanities last exam and PhD level reasoning or Olympiad math are wholly disconnected from the outcomes that consumers and enterprises actually care about”

Brandon Foodie · Mercor CEO & Co-Founder, Brendan Foody: How They Grew from $1M to $500M in 17 Months

“RL environments will subsume the entire economy because it doesn't make sense that humans would be doing monotonous, redundant work”

Brandon Foodie · Mercor CEO & Co-Founder, Brendan Foody: How They Grew from $1M to $500M in 17 Months

Report last updated: Apr 21, 2026

Get started free

Back to Entities Intelligence Report