Mechanize, a new AI startup co-founded by Matthew Barnett, Tamay Besiroglu, and Ege Erdil, has announced its ambitious goal of developing the building blocks for full economic automation. According to their announcement on X, Mechanize aims to automate tasks typically performed by workers using digital simulations, evaluations, and reinforcement learning environments that mimic real-world labor tasks such as coordination, re-prioritization, and handling interruptions.

The startup believes the lion’s share of AI value lies in automating common labor—not elite data science jobs—pointing out that current AI models still struggle with long-context reasoning, planning, and multimodality. Mechanize intends to solve these shortcomings by offering benchmark tasks that allow AI agents to learn practical long-horizon skills needed to perform everyday labor reliably.

Backing Mechanize are a slew of notable investors including Nat Friedman, Daniel Gross, Patrick Collison, Dwarkesh Patel, Jeff Dean, Sholto Douglas, and Marcus Abramovitch. TechCrunch also covered the launch in detail — read the article here.

The startup estimates the market potential of automating human labor to be around $60 trillion globally, suggesting that successful AI-based automation could transform economic productivity and living standards worldwide