RSS

Tuna

Karma: 34

Ac­cess to agent CoT makes mon­i­tors vuln­er­a­ble to persuasion

Jul 25, 2025, 4:09 PM
18 points

5 votes

Overall karma indicates overall quality.

0 comments4 min readLW link

Les­sons from a year of uni­ver­sity AI safety field building

Jun 6, 2025, 2:35 PM
28 points

17 votes

Overall karma indicates overall quality.

3 comments7 min readLW link