BOLD lead: alignment has no right answer, so build AI that augments humans instead
j_foerst · x · 2026-11-04
Asked whether his approach amounts to solving the hard alignment problem, Johannes Förstner says alignment is not a problem to solve in the technical sense: "what is good" is open-ended with no right answer, so BOLD focuses on openly building AI that augments human abilities and empowers value-aligned groups.
The questioner guessed BOLD uses a different architecture and training philosophy than Anthropic/OpenAI, along the lines of Yoshua Bengio's LawZero — which Förstner implicitly confirms.
Related event: BOLD project bets on amplifying human ingenuity over solving alignment(2 posts)→
More from AGI Musings
- Oxford physicist's 9-tweet thread: is a theoretical physics PhD still worth it in the AI era? — _lewtun · 2026-11-05
- Support for pausing AI jumps from 33% to 46% in eight months, overtaking regulation — DavidSKrueger · 2026-11-05
- AI is tackling longer research tasks, but scientific taste remains the hard part — nathanbenaich · 2026-11-05
- Gary Marcus reviews 2025 AI predictions: 17/18 correct on scaling hitting a wall — GaryMarcus · 2026-11-04
- DeepSeek's Liang Wenfeng on why open source is survival, not charity, in the AI era — teortaxesTex · 2026-11-04
- PhD Reflection: The Brain Isn't a GPU — Rest Is Part of Research — tw_killian · 2026-11-04