OpenAI was founded in 2015 as a research lab with an unusually grand mission: to ensure that artificial general intelligence — AI as capable as a human across most tasks — benefits everyone. For its first few years it was known mainly to researchers, publishing work on reinforcement learning and robotics.
Betting on scale
Its defining bet was on the Transformer and on scale. Where others tuned clever architectures, OpenAI pursued a simple, radical idea: take the Transformer, make it enormous, and train it on staggering amounts of text. The GPT series — Generative Pre-trained Transformer — proved the bet. GPT-3 (2020), with 175 billion parameters, could write essays, answer questions and generate code from a plain-language prompt, purely from having learned to predict the next word.
The ChatGPT moment
In November 2022, OpenAI wrapped that capability in a simple chat interface and released ChatGPT. It reached an estimated 100 million users in two months, one of the fastest technology adoptions in history, and made "generative AI" a household phrase almost overnight.
The launch reshaped the industry: it triggered an AI race among the largest technology companies, drew a multibillion-dollar partnership with Microsoft, and turned the once-obscure question of AI safety into a mainstream debate. Whatever comes next, OpenAI is the company that took the ideas traced through this Library — the neuron, deep learning, the Transformer — and put them in front of the world.