Mira Murati leaves OpenAI
OpenAI’s chief technology officer has announced her departure after six and a half years, prompting intense speculation about internal power struggles following last year’s failed attempt to remove the CEO. Commenters debate whether she was quietly pushed out as part of a broader consolidation of control and a shift toward a fully for‑profit structure, especially as multiple senior research leaders leave in the same window. The departures also fuel broader skepticism about OpenAI’s AGI timelines, its safety posture, and whether the current large‑model approach is hitting diminishing returns.
Context of the discussion
- Thread reacts to OpenAI’s CTO announcing her departure with a polished internal letter about “doing my own exploration.”
- Comes amid a string of senior exits (research leads, head of research, co-founder on sabbatical) and rumors of OpenAI shifting from nonprofit control to a more conventional for‑profit structure granting equity to the CEO.
Why she left (speculated)
- Many believe she was effectively pushed out due to her role in or proximity to the failed attempt to remove the CEO; “you don’t miss the king and then keep your job.”
- Others suggest:
- Burnout from an extreme, high‑stakes role.
- Discomfort with OpenAI’s directions (commercialization, safety posture, governance shifts).
- Cashing out after full vesting, with enough reputation to easily raise her own large startup round.
- Personal reasons or wanting more autonomy; some note that people with “fuck‑you money” often simply move on.
- A minority entertains the official framing at face value but is overshadowed by general cynicism about executive departure PR.
Governance, power, and corporate politics
- Broad view that the CEO decisively “won” the 2023 board showdown; current departures are seen as consolidation of his power and removal of internal opposition.
- Several comments argue what’s framed as a “coup” by the board was actually reversed by an investor‑backed counter‑coup.
- The move away from nonprofit control is read as:
- A prelude to full privatization and massive equity grants.
- A betrayal of the original “not in it for the money” mission.
- Or simply an inevitable shift to fund extremely capital‑intensive models and data centers.
AGI, LLMs, and OpenAI’s trajectory
- Some think departures signal insiders no longer believe AGI is “around the corner” or that OpenAI has a unique lead; others think they’re leaving precisely to pursue AGI elsewhere.
- Debate over whether current LLMs already qualify as a form of “AGI,” are just “fancy autocomplete,” or are hitting diminishing returns.
- Mixed assessments of OpenAI’s recent technical progress:
- o1 and multimodal work impress some, especially in reasoning benchmarks.
- Others see mostly incremental gains, heavy hype, and a looming “LLM winter.”
Perception of OpenAI and the farewell note
- Many see the farewell letter as boilerplate “corpspeak” with near‑zero information content, likely lawyer/PR‑filtered.
- Growing sense that OpenAI is transitioning from idealistic research lab to messy, highly politicized, hyper‑capitalist enterprise—and that the talent exodus is a major red flag, though its real cause remains unclear.