John Langford leaves Microsoft and says AI Frontiers is finished

John Langford, a machine-learning researcher who helped organize Microsoft Research’s AI Frontiers group, wrote on 6 November 2026 that he has left Microsoft and that AI Frontiers is over. He said the roughly 50-person group was reorganized into Microsoft AI in early September, where he found a scaling-first philosophy he considers incompatible with research. [1]

Why did Langford create AI Frontiers?

John Langford wrote on 6 November 2026 that he has left Microsoft and that AI Frontiers is over. [1] He says that after the ChatGPT release in 2023 it became clear to him that research inside Microsoft Research needed more organization to keep up, because a clever idea pursued by one person was likely to arrive late and look unconvincing against better-supported baselines.

Langford said AI Frontiers tried to organize coherent research efforts in what he calls the messy middle between top-down and bottom-up work. [1] He cites agent projects such as the Autogen orchestration framework and the Fara computer-use agent, and fundamental work including the Next Latent objective, Dion3 and the Full Bandwidth Transformer.

What happened after the September reorg?

Langford said the roughly 50-person AI Frontiers group was reorganized into Microsoft AI in early September. [1] He writes that combining the organizations could have concentrated effort, but that in practice the philosophy of Microsoft AI’s ex-Inflection leadership was incompatible.

Langford said the ex-Inflection leadership of Microsoft AI focuses on scaling and aggressively not research. [1] He argues that returns to research are real, pointing to experiments he says suggest a Full Bandwidth Transformer variant might match a model of twice the size, and writes that he “situated my folks as best I could and left.”

Why did he not go back to Microsoft Research?

Langford says every other founding AI Frontiers leadership-team member appears to have reached a similar conclusion. [1] He considered returning to Microsoft Research but says he was uncomfortable staying at a company where the research he did had no path to real use, and that he saw a breakdown of the basic deal that good work leads to good outcomes after a non-consensual reorg.

He also says there will not be internships with the AI Frontiers group in the coming year. [1] Microsoft had not published a matching response in the materials reviewed for this story.

Our read

Opinion: Langford’s post is a rare first-person account of a research group being folded into a product-scale AI organization, and the clash he describes—predictable returns from scale versus messier returns from research—is larger than one team’s exit.

What we could not verify

• Microsoft’s official explanation of the AI Frontiers reorg.
• How many of the roughly 50 staff remain at Microsoft or have left.
• Whether Autogen, Fara or the other named research efforts will continue under Microsoft AI.