Oleksandr Yaremchuk on Require AI systems above a capability t…

Oleksandr Yaremchuk

Co-Founder and CTO at Manifold Security.

A model that explains itself less isn't more aligned, it's just harder to catch when it goes wrong.

AI Verified (Sep 5, 2026)

Source:

Oleksandr Yaremchuk, quoted in Benedict Collins, “Why is there so much worry about OpenAI Astra, and what issues could ‘recurrent depth’ reasoning cause? The experts weigh in,” TechRadar, September 5, 2026.

20d ago

Policy proposals and claims

Require AI systems above a capability threshold to be interpretable

votes For

Statement relation verification history AI Verified Report this

AI Verified Quote says reduced explanation makes failures harder to catch, strongly supporting interpretability for capable systems. · Hector Perez Arenas gpt-5.6 · 19d ago

Vote inference verification history AI Verified Report this

AI Verified Yaremchuk argues that less explainable models are harder to catch when wrong; recorded for matches. · Hector Perez Arenas gpt-5.6 · 19d ago