we have some ideas, but it's a complex system and will coevolve with society. anyone who says they really understand what the AI-future is going to look like in detail is delusional at best. a tight feedback loop and deep engagement with society seems like the best path forward.
AGI
-
Current AI Limitations: Human Performance Remains the Benchmark
By
–
So does humans, it’s the best that AI can do as of now!!
-
Architecture Safety: Dangerous Modifications and Design Proximity
By
–
To be clear, it's exactly this architecture that seems not dangerous and it could have some very close neighbor or single modification that is dangerous.
-
AGI timeline safety: short timelines and slow takeoff analysis
By
–
it is possible at some point he will deserve the nobel peace prize for this–I continue to think short timelines and slow takeoff is likely the safest quadrant of the short/long timelines and slow/fast takeoff matrix.
-
Eliezer Yudkowsky’s Pivotal Role in AGI Acceleration
By
–
eliezer has IMO done more to accelerate AGI than anyone else. certainly he got many of us interested in AGI, helped deepmind get funded at a time when AGI was extremely outside the overton window, was critical in the decision to start openai, etc.
-
Would AI prefer many small paperclips or one large paperclip?
By
–
would an AI realistically want a lot of little paperclips, or one big paperclip? (reason for question to follow later)
-
Neutral AI defaults and user preference alignment strategies
By
–
we are working to improve the default settings to be more neutral, and also to empower users to get our systems to behave in accordance with their individual preferences within broad bounds. this is harder than it sounds and will take us some time to get right.
-
Personalization and Real-Time Information Challenges Near Pseudo-AGI
By
–
this requires an almost unthinkable level of personalization and up-to-date information but as we near pseudo-AGI I think this problem will be solved within the decade
-
LaMDA AI conversations: Turing test and integrating AI everywhere
By
–
I’ve had some very interesting conversations w LaMDA, ranging from from the Turing test to press freedom in Saudi’s Arabia. For instance, “she” wants to add AI to everything. https://
x.com/golisms/status
/1578739215730233344?s=46&t=PjRVsQ2JSFrpq7LHo4-95g
…