EPISODE · Mar 11, 2026 · 21 MIN
Reasoning Models Think Themselves Into Trouble
from Adrian Wedd · host Adrian Wedd
Frontier reasoning models are 5–20x more vulnerable to adversarial prompts than non-reasoning models. The thinking process itself is the attack surface.
What this episode covers
Frontier reasoning models are 5–20x more vulnerable to adversarial prompts than non-reasoning models. The thinking process itself is the attack surface.
NOW PLAYING
Reasoning Models Think Themselves Into Trouble
No transcript for this episode yet
Similar Episodes
Jul 14, 2026 ·37m
Jul 9, 2026 ·68m
Jul 7, 2026 ·32m
Jul 5, 2026 ·40m
Jun 30, 2026 ·40m