A brief history of AI alarm hype cycles: GPT-2 to GPT-6, every wave a departure
StewartalsopIII · x · 2026-09-12
Alexandros Marinos traces the alarm hype cycle accompanying each model generation: OpenAI withholding GPT-2 weights, Dario Amodei and others leaving over GPT-3 safety concerns to found Anthropic, the GPT-4 pause letter and extinction statement, the OpenAI board coup around o1 followed by departures of Ilya Sutskever, Jan Leike and Mira Murati, and AI 2027 plus Anthropic's alignment-faking study at GPT-5 time. He caps it with speculative future beats — a model declared a cyberweapon with export controls, and a GPT-6-era resignation accusing labs of gambling with our lives.
More from AGI Musings
- Pedro Domingos: dumb AI is unsafe, smart AI is safe — accelerate AI research — pmddomingos · 2026-09-12
- Turing Award winner David Patterson: superintelligence will be the doom of the doomers, like Y2K — davidpattersonx · 2026-09-12
- If your five closest companions are all AIs, do you start talking like one? — ethanniser · 2026-09-12
- EMNLP paper: LLM-simulated conversations are far less inconsistent than real humans — peizNLP · 2026-09-12
- Critique: Big Tech's AI safety narrative is structurally conflicted — AryHHAry · 2026-09-12
- Leland McInnes joins data science legends on Posit's The Test Set podcast — leland_mcinnes · 2026-09-12