Content
@
1 reply
0 recast
2 reactions
aaronv
@aaronv.eth
what have i been up to lately? I’m not trying to make smarter responses. I’m working on something that helps AI know when its own logic is slipping. Not safety rails. Not citations. Something deeper. Still early. Still weird. But I think it changes everything.
3 replies
0 recast
10 reactions
Apurv
@apurvkaushal
protecting against hallucinations?
1 reply
0 recast
1 reaction
aaronv
@aaronv.eth
more like knowing when something’s about to go off track—before it does
0 reply
0 recast
1 reaction
