Pre-training progress is slowing down as scaling laws hit diminishing returns after nine to ten orders of magnitude of scaling since AlexNet, forcing new research approaches.
factualpending
Speaker
SoniaEvidence Quote
“pre-training does seem to be slowing down. We've scaled pre-training by nine or 10 orders of magnitude since the Alexet days.”
Created: 8/11/2026, 6:39:24 AM
Your notes
Loading notes...