fchollet: Pure Parameter Scaling Hit a Wall; Test-Time Compute Is Necessary

fchollet · x · 2026-08-02

Keras creator François Chollet notes that despite massive scaling (100,000x since 2019), base LLMs without test-time compute still perform poorly on the ARC 1 benchmark.

He argues that the single-pass, static next-token prediction paradigm (GPT-2 to GPT-4 era) hit a capability asymptote. Without shifting to the test-time compute paradigm, AI would not be capable of the advanced reasoning seen in current SOTA systems. This test-time adaptation was a necessary evolutionary patch to bypass the plateau of deep learning.

Related event: François Chollet: Test-Time Compute is Key as Pure Parameter Scaling Hits Limits(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →