Guest
Jeffrey Ladish
Executive director, Palisade Research
Studies AI offensive capabilities at Palisade; helped build Anthropic’s information-security program.
Follow Jeffrey Ladish
We’ll email you each new moment, one email a day at most. Plus the best of every big interview, Wednesday and Saturday.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
It’s free.
Real Alpha has cut 5 moments of Jeffrey Ladish (Executive director, Palisade Research) talking, from an interview on MTS published Nov 6, 2026.
What Jeffrey Ladish said
- 0:57
METR sees faster progress, but key evidence remains hidden
“I think that, like, Anthropic should release this data, unredact this data, so that, like, the whole world can assess this for themselves.”
at 3:34 - 12:28
Pacing may need to include limits on model use
“the level of pacing should be like man slam on the brakes”
at 14:21 - 21:05
AI gains generalize unevenly, especially in social skills
“the models are still not very funny or charismatic.”
at 22:42 - 3:55
Real agent work is a better capability test than benchmarks
“i just went from like using agents sometimes to using agents every single day”
at 5:06 - 17:39
A pacing deal could create room to establish a safety case
“what i think might happen optimistically is something like there is a negotiation between people who are saying you know let's do very moderate pacing”
at 19:13
- 0:57
Is this you? Add the badge to your site.
