RSS

Parv Mahajan

Karma: 1,145

parvmahajan.com

Ge­or­gia Tech AI Safety Ini­ti­a­tive Ret­ro­spec­tive 2025-2026

24 Jul 2026 11:55 UTC
60 points
3 comments7 min readLW link

Sev­eral fron­tier mod­els are sub­stan­tially pre­fill aware

17 Jun 2026 17:41 UTC
62 points
2 comments5 min readLW link

AI Safety’s Biggest Ta­lent Gap Isn’t Re­searchers. It’s Gen­er­al­ists.

13 Apr 2026 14:47 UTC
60 points
2 comments6 min readLW link

The cur­rent SOTA model was re­leased with­out safety evals

8 Mar 2026 1:51 UTC
111 points
12 comments5 min readLW link

Three ways to make Claude’s con­sti­tu­tion better

Parv Mahajan2 Feb 2026 21:48 UTC
41 points
3 comments2 min readLW link