The lacklustre performance of ChatGPT 4.5 provides further evidence that the era of steady pretraining scaling is over. This has many implications for AI governance.
The lacklustre performance of ChatGPT 4.5 provides further evidence that the era of steady pretraining scaling is over. This has many implications for AI governance.
www.tobyord.com
Inference Scaling Reshapes AI Governance — Toby Ord
