Three-Year-Old Essay Resurfaces: To Make AI Safe, Speed It Up
Dan_Jeffries1 · x · 2026-09-10
Daniel Jeffries resurfaced his Feb 2023 essay "Let's Speed Up AI," arguing it predicted the current wave of AI panic. His core claim runs opposite to slow-down calls: problems can only be solved in the real world, so making AI truly safe requires deploying it faster. ChatGPT is his case study — after OpenAI shipped it publicly, the constant jailbreaks, hacks, and mockery acted as a free crowdsourced QA team, and both model capability and guardrails improved through that live feedback. Since humans are endlessly creative at finding exploits, no amount of closed-door testing can substitute for real-world interaction.
More from AGI Musings
- Greg Egan's 1995 AI classic 'Learning To Be Me' resurfaces as free full-text — ZeroStateReflex · 2026-09-10
- AI won't kill everyone: blogger rounds up essays arguing doom debate targets wrong threats — binarybits · 2026-09-10
- tszzl mocks Tegmark IV doomsayers: 'Platonic entities are welcome in my backyard' — tszzl · 2026-09-10
- Programmed love still feels real: an analogy in the AI companionship debate — MajmudarAdam · 2026-09-10
- AI-run interviews reveal a split: childfree cite freedom, would-be parents cite cost — soumitrashukla9 · 2026-09-10
- Cambridge prof David Krueger puts AI catastrophe risk above 50%, says everyone is understating it — KatjaGrace · 2026-09-10