2019 Guardian headline resurfaces: GPT-2 was once 'too dangerous to release'
Frail_Waif · reddit · 2026-10-14
A Reddit user dug up The Guardian's 2019 report on OpenAI's GPT-2, where the lab said the model was so capable and misuse risk so high that it broke its usual practice of releasing full research.
The article quotes then-research director Dario Amodei explaining the 40GB training corpus of 10M Reddit-sourced articles — enough to store about 35,000 copies of Moby Dick. The poster quips that 'some people never change,' drawing a line between 2019's staged-release caution and OpenAI's current posture.
More from Companies & People
- Humanities dissertations have collapsed, social sciences are next — with AI as one driver — gsiemens · 2026-10-23
- For consumer AI founders, 'When are you launching in Europe?' is the new 'When Android?'—only 100x worse — omooretweets · 2026-10-23
- Lilian Weng to Discuss Recursive Self-Improvement at SGLang Summit 2026 — BanghuaZ · 2026-10-23
- OpenAI FDE team: make custom work obsolete by feeding customer learnings into products — colintjarvis · 2026-10-23
- Bindu Reddy: record AI security incidents, tech stock slide, rogue agents rattle banks — bindureddy · 2026-10-23
- StepFun shows up at AGNTCon and MCP Con North America with booth #24 — StepFun_ai · 2026-10-23