GPT-2 language model generating text about itself

GPT-2: 'Too Dangerous to Release'

Research Breakthroughs The Transformer Era

Impact

What Happened

OpenAI announced GPT-2 (1.5 billion parameters) but initially refused to release the full model, calling it 'too dangerous' due to its ability to generate convincing fake text. The decision was controversial — some praised the caution, others called it a publicity stunt. The full model was eventually released in November 2019.

Why It Mattered

First major AI safety controversy in the LLM era. Brought public attention to the dual-use nature of AI language models. The 'too dangerous to release' framing became a defining moment in AI safety discourse.

Key People

Alec Radford

Organizations

OpenAI

Related Milestones

2018.06Research

GPT-1: Generative Pre-training

OpenAI released GPT-1, demonstrating that a Transformer trained on vast amounts of text using unsupervised pre-training could then be fine-tuned for specific NLP tasks. With 117 million parameters, it showed the potential of scaling language models.

Alec RadfordOpenAI

2020.06Research

GPT-3: The 175 Billion Parameter Leap

OpenAI released GPT-3 with 175 billion parameters — 100x larger than GPT-2. Without any fine-tuning, GPT-3 could write essays, code, poetry, translate languages, and answer questions through 'few-shot learning' (learning from just a few examples in the prompt). The API launched in beta, enabling thousands of applications.

Tom BrownOpenAI

2021.01Research

DALL-E: Text to Image Generation

OpenAI unveiled DALL-E, a model that could generate images from text descriptions — 'an armchair in the shape of an avocado' became iconic. Built on GPT-3's architecture adapted for images, it showed that language models could bridge the gap between text and visual creativity.

OpenAI

2023.03Research

GPT-4: Multimodal Intelligence

OpenAI released GPT-4, a multimodal model that could understand both text and images. It passed the bar exam (90th percentile), scored 1410 on the SAT, and demonstrated remarkably nuanced reasoning. It was a massive leap from GPT-3.5 in accuracy, safety, and capability.

OpenAI

2021.06Product

GitHub Copilot: AI Writes Code

GitHub launched Copilot as a technical preview — an AI pair programmer powered by OpenAI Codex that could autocomplete entire functions, write boilerplate, and suggest code from natural language comments. It was trained on billions of lines of public code.

GitHubOpenAI

Get the latest AI milestones as they happen

Join the newsletter. No spam, just signal.