@stevesperandeo.bsky.social
Reposted
I have bad news: The offshore southern portion of the Great Barrier Reef that escaped severe coral bleaching in 1998, 2002, 2016, 2017, 2020 & 2022 finally ran out of luck in 2024.
December 1, 2024 at 2:55 AM
I have bad news: The offshore southern portion of the Great Barrier Reef that escaped severe coral bleaching in 1998, 2002, 2016, 2017, 2020 & 2022 finally ran out of luck in 2024.
Reposted
We’re down to 2% of reefs on the Great Barrier Reef untouched by coral bleaching - and that’s supposed to be good news?
December 1, 2024 at 4:01 AM
We’re down to 2% of reefs on the Great Barrier Reef untouched by coral bleaching - and that’s supposed to be good news?
Reposted
I'm biased ofc but IMO the ~right theoretical framework for thinking and talking about instructable LLM-based agents is assistance games (a.k.a. Cooperative Inverse Reinforcement Learning (CIRL) games).
Setting aside the fact that people are literally re-inventing a bunch of terminology around agents when we have several decades of framing from RL (POMDPs, anyone?) that we can rely on to describe what we're doing, the main challenges are as follows., working backwards from the end point. [2/11]
November 19, 2024 at 2:07 PM
I'm biased ofc but IMO the ~right theoretical framework for thinking and talking about instructable LLM-based agents is assistance games (a.k.a. Cooperative Inverse Reinforcement Learning (CIRL) games).