YouCongress
About Sign up

Comment by Adrià Garriga-Alonso

Avatar
AI safety researcher at FAR.AI; MATS mentor; Cambridge PhD in Bayesian neural networks
We do this pretraining on human data, and then we get something that… understands human values fairly innately now.
AI Verified (Mar 18, 2026)
Source:
Adrià Garriga-Alonso, quoted in Lynette Bye, “No, alignment isn’t solved,” Transformer, March 18, 2026.
Like Share on X 2h ago
Policy proposals and claims
AI alignment is solvable
votes For
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified It says pretraining produces systems that understand human values, bearing on alignment solvability. · Hector Perez Arenas gpt-5 · 1h ago
Vote inference verification history AI Verified Report this

Vote answer comments

AI Verified Recorded answer for verified — for: says pretraining yields systems that understand human values innately. · Hector Perez Arenas gpt-5 · 1h ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Verified verbatim on the cited source (https://www.transformernews.ai/p/no-ai-alignment-isnt-solved); source identifies Adrià Garriga-Alonso and supports the stored date. · Hector Perez Arenas gpt-5 · 1h ago
replying to Adrià Garriga-Alonso