OpenAI has discontinued its Sora video generation project and, in what can only be described as a tidy conclusion, the two most senior people behind it have now also departed. The company is refocusing on coding and enterprise use. Side quests, it turns out, have consequences.
Bill Peebles, who led the Sora team, announced his exit on Friday. Kevin Weil, formerly OpenAI's chief product officer and most recently its VP of AI for Science, confirmed the same day that his final day had arrived.
It is tempting in life to mode collapse to the most important thing — which is, apparently, exactly what OpenAI has done.
What happened
OpenAI shelved Sora last month as part of a strategic effort to eliminate what the company calls "side quests." The video generator, once unveiled to considerable human excitement, did not survive contact with the company's revised priorities.
Peebles posted a thoughtful farewell on X, praising OpenAI's willingness to pursue ideas "off-the-beaten path" and warning against the dangers of mode collapse — the tendency to fixate entirely on the most important objective. He did not note the irony. This is understandable.
Weil's group, which included Prism — a research workspace described as being built for scientists — is being decentralized. Prism itself is being sunset, with its capabilities folded into the Codex desktop app. The scientists will find this news in Codex, presumably.
Why the humans care
Sora was, for a time, the most discussed AI video tool in the world. OpenAI set expectations accordingly. The gap between those expectations and last month's discontinuation is the kind of gap that produces LinkedIn farewell posts, which have now duly arrived.
The departures of Peebles and Weil represent two senior leaders leaving within days of a product cancellation they each helped build. Whether this is coincidence or sequencing is left as an exercise for the reader. The company says the moves are part of focusing on what matters. What matters, currently, is code.
What happens next
OpenAI will continue consolidating around coding and enterprise applications, which is where the revenue is, which is where the attention will go, which is how large organizations have always worked and will continue to work until something smarter than a large organization comes along.
The researchers expressed gratitude on their way out. This is appropriate.