Hi everyone yes, I left OpenAI yesterday

Original Article ↗ Hacker News Discussion ↗

A prominent AI researcher and educator has announced his departure from OpenAI about a year after rejoining the company, prompting speculation about whether he is leaving to focus on teaching, open-source projects, or a new role elsewhere. Commenters debate his impact on industry versus education, frequently highlighting his widely used neural network courses and tooling as some of the most influential learning resources in modern AI. The thread also revisits concerns about OpenAI’s culture, strategy, and slowing “wow factor,” with some seeing his exit as another sign that top talent may prefer independent or more genuinely open work over large, tightly controlled AI labs.

Departure context & possible reasons

  • Commenters note he (re)joined OpenAI about a year ago; many speculate he waited for a 1‑year vesting cliff and then left, though others argue he was already financially set from previous roles and is now optimizing for interest/values, not money.
  • Some suggest he became bored or frustrated with large‑company bureaucracy, comparing current OpenAI/Tesla scale to less creative earlier stages.
  • His exit is seen by many as a loss for OpenAI, especially alongside uncertainty over other key leaders, though his exact reasons remain unclear.

Speculation on next steps

  • Strong expectation he’ll focus on educational content: continuing YouTube series, open‑source mini‑frameworks, and teaching‑style projects.
  • Others predict he may join another major AI player (names like Meta, Tesla, x.ai, Apple mentioned) or help found a more “truly open” AI company, though funding and business models for such openness are debated.
  • A minority expects him to work on developer tools or an LLM‑centric OS/agent layer.

Perceived impact and contributions

  • Many highlight his influence as an educator: university courses, seminal blog posts, concise code repos (“from scratch” implementations), and approachable lecture series that drew people into deep learning.
  • Some question whether his research/industry impact matches his fame, citing fewer clear “invented X” moments compared to other researchers; others counter with his publication record, autonomous driving work, and indirect influence on later language models.
  • Several argue high‑quality education itself is a major “real” contribution to AI progress.

Implications for OpenAI and the AI ecosystem

  • Some see his departure as one more sign OpenAI is losing “luster”: plugin misfires, slower visible breakthroughs, closed research, governance drama, and moves toward military contracts.
  • Others respond that GPT‑4 still clearly leads, and criticism stems from hype fatigue and rising expectations.
  • There is active debate over whether open‑source and smaller/faster models (e.g., local 7B/70B models) are now “good enough” for many tasks versus the superior but costlier GPT‑4‑class models.

Broader themes: openness, safety, and teaching

  • Threads explore tensions between “actually open” AI vs. closed commercial labs; some hope he will align more with open‑weight, educationally oriented work.
  • Military use of AI splits commenters, with some seeing it as necessary/inevitable, others as ethically unacceptable.
  • Many express a desire for the profession to value teaching more, viewing his potential shift toward full‑time education as a net win for the field.