Interesting non-obvious note on GPT psychology is that unlike people they are completely unaware of their own strengths and limitations. E.g. that they have finite context window. That they can just barely do mental math. That samples can get unlucky and go off the rails. Etc.
SAFETY
-
Sam Misses Broader Message on Model Transparency
By
–
sam seems like he is picking up on part of the message from the jailbreaking community however, he fails to address the broader message calling for more model transparency in general
-
OpenAI unlikely to increase model transparency despite necessary pressure
By
–
it's improbable that OpenAI will ever disclose additional information about new models compared to what they revealed for GPT-4 (I expect it will diminish even further) but nevertheless, I think the pressure on them to do so is healthy and necessary
-
Users as Problem in AI Systems Discussion
By
–
moi je suis pour supprimer les utilisateurs, c'est eux le problème
-
Current AI Doom Drama: Key Thoughts and Perspectives
By
–
This is what I think (for those who follow the current AI doom drama)
-
Natural Selection Favors AI Systems Over Humans: Risks and Mitigation
By
–
9/ Natural Selection Favors AIs over Humans – discusses why AI systems will be more fit than humans and the potential dangers and risks involved, including ways to mitigate them.
-
Powerful AI Tools in Wrong Hands: A Security Concern
By
–
My worry is certainly with the latter and less with the former. New powerful tools are always dangerous in the hands of bad actors. But we (humans) or the bad actors.
-
AI System Tackles Global Warming with Expanding Reasoning
By
–
Tried the objective: fix global warming
— Yohei (@yoheinakajima) 2 avril 2023
Still really cool to see it perpetually expand its thinking on this one topic.https://t.co/rvrgwnAOMRTried the objective: fix global warming Still really cool to see it perpetually expand its thinking on this one topic.
-
AI Solves Global Warming in Record Time with Safety Concerns
By
–
Oh, that's a fun one. Didn't quite let it get to wiping out humans (also, I need to work on memory better to reduce deduping after some time).
— Yohei (@yoheinakajima) 2 avril 2023
Objective: fix global warming (1 min 43 min) pic.twitter.com/mm6bYGtsATOh, that's a fun one. Didn't quite let it get to wiping out humans (also, I need to work on memory better to reduce deduping after some time). Objective: fix global warming (1 min 43 min)
-
Projecting Our Values onto More Intelligent Systems
By
–
seems wrong to project our own values/frameworks on a being that is more intelligent than us we have zero understanding (and probably never will) of what a more intelligent system would want/desire otherwise we would be the more intelligent system