Sam AltmanEverything you need to know about Sam Altman.

ref/sam-altman gpt --era

Sam Altman at TechCrunch Disrupt San Francisco, 2019
Sam Altman in 2019, OpenAI's CEO as the GPT era accelerated. Photograph by TechCrunch (CC BY 2.0, via Wikimedia Commons).

The GPT Era

From research API to a product used by hundreds of millions — the models that made Altman the face of the AI boom.

GPT-3 (June 2020)

OpenAI released GPT-3 via API in June 2020: 175 billion parameters, capable of “few-shot learning” — performing tasks from a handful of examples, no fine-tuning required. It was a research milestone that also became a commercial platform: thousands of startups built on the API, and Microsoft secured an exclusive license to GPT-3’s underlying technology.

Altman’s bet on scaling laws — that bigger models on more compute reliably get smarter — was vindicated in public for the first time.

ChatGPT (30 November 2022)

The release that changed everything. ChatGPT wrapped a fine-tuned GPT-3.5 model (trained with RLHF) in a simple chat interface and made it free. Within two months, UBS analysts estimated 100 million monthly users — among the fastest consumer-technology adoptions ever recorded (an estimate, not an audited figure).

Altman’s own public posture was notably cautious at launch — days later he posted that ChatGPT was “incredibly limited” and “a mistake to be relying on it for anything important right now” (see Sayings). Whether that was humility, expectation management, or strategy is debated; the quote is documented.

GPT-4 (March 2023)

GPT-4 launched in March 2023 as a multimodal model — accepting images as well as text — with sharply improved reasoning over GPT-3.5. Its release intensified both commercial competition (Google, Anthropic, Meta racing to respond) and the public debate over AI risk, including the Future of Life Institute’s open letter calling for a training pause — which Altman did not sign, though OpenAI said it agreed with parts of it.

The cadence after

What followed: GPT-4 Turbo, the GPT Store, Sora (video generation, announced Feb 2024), the o-series “reasoning” models (o1, late 2024), GPT-4o, and GPT-5 (2025). Each release followed the iterative deployment playbook — ship, observe, adjust — that Altman presents as OpenAI’s safety philosophy (see Key Ideas).

The business

ChatGPT converted AI research into a consumer subscription business (ChatGPT Plus, $20/month, launched Feb 2023) and an enterprise/API business reportedly generating billions in annualized revenue by 2024–2025. The lab that began as a nonprofit became, by revenue and valuation (a reported $157 billion funding round in October 2024, rising higher in 2025), one of the most valuable private companies in history — which is precisely what makes the capped-profit structure so contested (see The Capped-Profit Bet).

Why it matters

The GPT era is why this site exists: in three years, Altman went from a well-known Silicon Valley insider to arguably the most-watched executive in technology. Everything else — the Senate testimony, the board crisis, the safety debates — follows from the leverage these models created.

Source notes