R1, OpenAI’s o3, and the ARC-AGI Benchmark: Insights from Mike Knoop on the Gradient Dissent Podcast

Поділитися
Вставка
  • Опубліковано 11 лют 2025

КОМЕНТАРІ • 3

  • @frun
    @frun 46 хвилин тому +1

    I hypothesize, AI👾🤖👾 should adapt its architecture dynamically to be AGI. Neuroplasticity is needed 🧠

  • @blengi
    @blengi 6 днів тому +1

    how does deepseek do on ARC AGI and frontier math compared to o3?

  • @securityinteractive
    @securityinteractive 6 днів тому +1

    The only confident thing I can say is that it is extremely difficult to predict the course of new technology. More specifically AGI is very subjective. There will be many good and bad AGI attempts and I expect some models to become by-words for frustration and confusion in consumers.

  • @narrowsuperintelligentai
    @narrowsuperintelligentai 4 дні тому

    because relying exclusively on more data and more compute will be limited to the human-level intelligence expressed in the data set, the discovery of new linguistic rules of logic, perhaps by ais, may be absolutely necessary to reaching asi.