OpenAI presented GPT-2 and initially limited release of the largest model over disinformation risks, foreshadowing open vs closed dilemmas.
In February 2019, OpenAI announced GPT-2, a 1.5B-parameter decoder transformer capable of coherent prose. Staged release of the large checkpoint ignited misuse debate.
The model showed informal few-shot behavior and surprising narrative continuity for its size. Journalists and academics debated hype versus real risk.
GPT-2 prepared GPT-3’s cultural and technical ground: more parameters, more data, more emergent abilities—and more governance.
Brown et al. showed a large enough LLM performs tasks with few-shot prompts without task-specific fine-tuning.
OpenAI launched ChatGPT and reached 100 million users in months; conversational AI became a global habit.
OpenAI launched GPT-4 with notable gains on professional exams, robustness, and later visual ability.