"Any definition of safety antagonistic to capability will never work" sparks debate
thisismyhat · x · 2026-11-06
A viewpoint circulating in AI circles: any definition of safety that is antagonistic to capability will never work — "Capability finds a way." It challenges the approach of trading capability for safety, echoing the long-running debate between safety and acceleration camps over whether safety should be embedded within capability development rather than opposed to it.
More from AGI Musings
- Apollo: Still No Sign of AI Boosting Overall US Productivity — michalmalewicz · 2026-11-06
- AI May Worsen Market Efficiency Short-Term but Optimize Allocation Long-Term — jzlegion · 2026-11-06
- Garrison Lovely's New Book 'Obsolete' Argues to Halt the Race to Replace Human Labor — GarrisonLovely · 2026-11-06
- Every asks whether AI's time savings go to better creative work or just more output — every · 2026-11-06
- Jensen Huang: engineers using agents are busier, not less, because they must verify output — shashib · 2026-11-06
- Canadian grant applicants may use AI but reviewers may not, researcher says that's wrong — KordingLab · 2026-11-06