Sutton: raw compute alone won't produce intelligence, AI science still needs new methods
What happened
The Decoder reported on November 8, 2026 that Richard Sutton said raw compute alone is not enough to achieve intelligence, and that AI science still needs clever new methods and is far from finished. He argues for AI agents with world models that keep learning from their own experience without destroying knowledge they already hold. Jerry Tworek and Jayden Teoh criticized reading "The Bitter Lesson" as a case for blindly scaling Transformers, or for ruling out new research directions.
Written by AI from the coverage · updated yesterday
Coverage
Follow the reports to see the story from different sides.
- The DecoderThe inventor of the bitter lesson has a bitter lesson for LLM scaling enthusiasts
Richard Sutton 表示，仅靠原始算力不足以实现智能，AI 科学仍需要巧妙的新方法，远未完成。他主张让具备世界模型的 AI 智能体从自身经验中持续学习，并避免破坏已有知识；Jerry Tworek 和 Jayden Teoh 则批评将“苦涩的教训”解读为盲目扩展 Transformer 或排斥新研究思路。
Heat over time
Heat now 5·Comparable peak 10（Nov 8 20:00）·Comparable change over 24 hours -50%
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.