Jonathan Siddharth, CEO of Turing, predicts that evaluating and training AI models will become one of the most common jobs in the world — a part-time role for millions of people across every industry. In this conversation from Davos, we discuss why coding AI took off first, what comes next for enterprise automation, why he thinks current AI adoption across the S&P 500 "rounds down to zero," and what it'll take to close the trust gap. Siddharth also shares his perspective on what makes Claude Code work, why Tesla's self-driving data won't help Optimus generalize, and why almost all enterprise AI systems will be human-in-the-loop.
Turing's data powers the top five models on SWE-bench. The company works with all of the major frontier AI labs, has a network of four million coders, and raised $111 million at a $2.2 billion valuation.