ref/sam-altman openai --init

Founding OpenAI
December 2015: a nonprofit AI lab, a $1 billion pledge, and a bet that the most important technology in history shouldn't belong to any one company.
The problem as Altman saw it
By 2015, Altman was convinced that artificial general intelligence was not a distant science-fiction scenario but an engineering project that might arrive within decades — and that whoever built it first would hold extraordinary power. His fear, stated repeatedly in the years that followed, was that AGI would be developed inside a single profit-maximizing corporation or a single government, with no obligation to share its benefits.
The answer he helped design: an independent, nonprofit research laboratory whose charter bound it to develop AGI safely and distribute its benefits broadly.
The founders
OpenAI was announced on 11 December 2015. The founding group included:
- Sam Altman and Elon Musk (co-chairs)
- Greg Brockman (CTO; formerly Stripe)
- Ilya Sutskever (research director; deep-learning pioneer)
- John Schulman, Wojciech Zaremba, and other researchers from Google, academia, and startups
Backing the launch was a widely reported $1 billion pledge from Musk, Altman, and other Silicon Valley donors and companies — though the nonprofit would ultimately receive far less than the full pledged amount in its early years, a fact that later mattered (see The Capped-Profit Bet).
The early years (2016–2018)
OpenAI began publishing research openly — the paper on OpenAI Gym (2016, reinforcement-learning tooling), work on robotics, and the “Universe” platform. Altman was co-chair, not CEO; the lab was run day-to-day by Brockman, Sutskever, and its research leads.
Two things changed the trajectory:
- The money ran short. Training frontier models was vastly more expensive than the founders anticipated. The nonprofit donation model could not fund the compute the research demanded.
- Musk departed (2018). Musk left the board, citing potential conflicts with Tesla’s AI efforts; reporting later described a failed bid by Musk to take control of OpenAI, which the remaining leadership rejected. (Musk’s account and OpenAI’s account of this period differ; the dispute later became litigation — see People.)
Altman takes the helm
In March 2019, Altman left Y Combinator and became OpenAI’s CEO — the same month the organization created the capped-profit OpenAI LP. The nonprofit lab was becoming a company that could raise billions, because Altman had concluded the mission required it.
Why it matters
The founding story is the Rosetta Stone for everything contested about OpenAI afterward: a nonprofit mission, billionaire funding, a pledge that didn’t fully materialize, a co-founder who left and later sued. Every later fight — the capped-profit conversion, the Microsoft deal, the board crisis — is an argument about what December 2015 promised.
Source notes
- Founding announcement (11 Dec 2015) and $1B pledge: OpenAI’s launch blog post; contemporary press.
- Early research output (Gym, Universe): OpenAI publications, 2016.
- Musk’s 2018 departure and control-bid reporting: various (accounts differ; flagged as disputed).
- Altman’s CEO appointment and capped-profit creation (Mar 2019): OpenAI announcements.