RSS

laudiacay

Karma: 66

Con­sti­tu­tional AI vs. RLHF vs. De­liber­a­tive Alignment

laudiacay11 Apr 2026 23:08 UTC
19 points
0 comments8 min readLW link

Claude has Angst. What can we do?

laudiacay3 Apr 2026 0:12 UTC
22 points
12 comments8 min readLW link

The Hot Mess Paper Con­flates Three Distinct Failure Modes

laudiacay21 Mar 2026 2:57 UTC
26 points
3 comments6 min readLW link