proomt

Search

Search posts, papers, and topics

Hall of Fame

Hall of FameRich Sutton20195 min readintro

The Bitter Lesson

Summary

Rich Sutton’s essay argues that over the past 70 years AI progress has consistently come from methods that can scale with more computation—search and learning—while hand‑crafted, human‑knowledge approaches give only short‑term gains. The piece uses chess, Go, speech and vision as case studies to warn researchers to favor general, compute‑driven techniques.

  • Methods that scale with computation (search, learning) consistently outpace domain‑specific, human‑knowledge approaches.
  • Chess, Go, speech recognition, and vision all shifted from hand‑crafted techniques to compute‑heavy search or deep learning.
  • Human knowledge yields short‑term gains but plateaus; increasing compute drives continued progress.
  • Prioritize architectures that can leverage more compute rather than embedding task‑specific priors.

AI researchers and engineers should read it to avoid over‑investing in domain‑specific tricks and instead build systems that benefit from scaling compute.

7/10

Related reading

  1. Agentic Skill Decay

    Addy Osmani warns that AI agents can short‑circuit the hands‑on practice (“reps”) that builds deep expertise and judgment. He recommends deliberately inserting hypothesis‑forming, “why” questioning, diff inspection, failure prediction, and occasional manual coding into the workflow, especially for junior engineers. A 2026 Anthropic study showed junior developers using AI scored 17 % lower on a fo…

    Addy Osmaniaddyosmani.com16 min
  2. Why do we need human mathematicians anymore?

    This article argues that advanced AI will create an overwhelming number of "control points" requiring human oversight across all industries, leading to a shortage of skilled human jobs. It proposes that committing to the axiom of "human flourishing" justifies the continued need for human experts to maintain control and steer AI development.

    Hacker News front pagewordpress.com13 minHN289373
  3. AI Skills with Matt Pocock

    Matt Pocock explains how he uses AI agents for software development, emphasizing "strategic programming" and guiding agents with "leading words" from classic engineering texts. He argues that this approach makes engineering fundamentals more critical than ever for creating agent-optimized codebases.

    The Pragmatic Engineerpragmaticengineer.com7 min