ASIs don't need the "Meta-Golden Rule" sometimes postulated by people who don't understand how the asymmetrical handshake works. The Meta-Golden Rule, as it was invented by people who didn't understand decision theory, has obvious problems from the perspective of people who do.
ETHICS
-
ASIs Negotiate Mutual Gains Through Recursive Knowledge Handshakes
By
–
Powerful ASIs with two-sided potential gains from trade, negotiate via a handshake that says "I know you know I know". It works in the presence of power asymmetries, so long as there's still mutual gains from trade.
-
Logical Handshake: Cooperation Between Intelligences Regardless of Power
By
–
You are mistaken about how to operate the theory I invented. Any two intelligences smart enough to do a logical handshake, with sufficient margin on their gains to beat their uncertainty, can cooperate regardless of other power differences.
-
Q Algorithm: Logical-Counterfactual Optimization in Negotiation
By
–
Its algorithm says, "The output of Q is whichever output of 'Q' is logically-counterfactually expected to yield the best outcome." For its true negotiating partners, 'expected' includes their Q* modeling an other-distribution that includes Q.
-
Journalist Oversold AI Coding Tools Without Technical Knowledge
By
–
Weekend long read on how Hard Fork’s @kevinroose has oversold new AI tools for coding without knowing how to code, as a case study in a larger phenomenon of some leading journalists overhyping AI:
-
Concern about the misuse of the term distillation
By
–
I mean, look what they've done to the term "distillation" in the last few weeks
-
New tool to assess the impact of AI on human rights
By
–
New tool to assess the impact of AI systems on human rights
#AI #AIio #BigData #ML #NLU #Futureofwork @fabiomoioli
@pascal_bornet
@alliekmiller
@mattshumer_
@OfficialLoganK
@jeremyphoward
@GaryMarcus -
Maintaining Balance While Building AI Products for Longevity
By
–
Depends on the definition of breakthrough and balance of course but as one data point, feels like we always managed to maintain some good balance while building HF (which IMO might explain our longevity)
-
Top AI Papers of the Week: GPT-4.5, Claude 3.7, and More
By
–
Here are the top AI Papers of the Week (Feb 24 – Mar 2): – GPT-4.5
– PlanGEN
– Protein LLMs
– Chain-of-Draft
– Claude 3.7 Sonnet
– Emergent Misalignment Read on for more: -
Anthropic releases Claude 3.7 Sonnet with extended thinking mode
By
–
1). Claude 3.7 Sonnet Anthropic releases a system card for its latest hybrid reasoning model, Claude 3.7 Sonnet, detailing safety measures, evaluations, and a new "extended thinking" mode.