Gary Marcus

Rank 14 of 47
|
Score 83
5/1/2025, 10:43:56 PM
X
In reply to:
Jordan Moren
@PazneriaJ
·
245d
@GaryMarcus It was Satya Nadella I heard say that - in his interview with Dwarkesh. Yes I'm pretty sure he used those exact words
Gary Marcus
@GaryMarcus
·
245d
@PazneriaJ The chart says performance is up; it’s important to understand what that means and the graph doesn’t convey that.

Who said intelligence is “log of compute” btw, in exactly those words? Would be useful.
Jordan Moren
@PazneriaJ
·
245d
@GaryMarcus Gary I really do value your perspective, but I feel like you're disingenuous sometimes. They've always said that intelligence is the log of compute, they never claimed it scales linearly with compute.
Gary Marcus
@GaryMarcus
·
245d
GPT-4 came out two years ago; is GPT 4.1 6.25 times smarter?

FLOPs/s have scaled faster than intelligence.
Epoch AI
@EpochAIResearch
·
245d
We've documented over 500 AI supercomputing clusters from 2019 onward.

Among leading AI supercomputers, performance has been increasing at a rate of 2.5x per year.

Thread of our insights 🧵

The conversation involves a discussion on AI performance and intelligence scaling, referencing a statement by Satya Nadella. It touches on the interpretation of data and the validity of claims about AI development. The tone is analytical, with some disagreement on the interpretation of intelligence scaling.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The discussion is mostly respectful and focused on understanding AI performance, aligning with the principle of doing no harm. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The conversation promotes understanding by discussing the nuances of AI intelligence and performance metrics. [+1]
  3. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    There is a slight tension in the exchange, with a participant feeling another is disingenuous, which could be seen as a minor violation of engaging in constructive criticism. [-1]