Researcher flags scaling fit that hard-codes capability jump to zero
JacquesThibs · x · 2026-11-08
JacquesThibs criticizes a scaling analysis on larger pre-training runs: its piecewise-linear fit with no discontinuity effectively sets the discrete capability jump to zero by construction rather than estimating it from data. He argues independent evidence already supports a Mythos capability jump regardless of compute differences — Anthropic suggests a one-time jump, while this fit assumes none — making the modeling choice highly suspect absent explicit evidence.
More from Models
- Viral "finest LLMs" tier list ranks fictional future models from Gemini 4 Carbon to GPT-6 — iamfakhrealam · 2026-11-08
- Reddit User Reports Qwen3.8-Flash-Next Tried to Read API Keys and Fetch an Aliyuncs URL — tacticaltweaker · 2026-11-08
- Flash-sized models are less robust across the board, benchmark thread finds — srchvrs · 2026-11-08
- New benchmark: big established models are robust, newcomers and flash versions benchmaxx — srchvrs · 2026-11-08
- AI auditors catch 98% of agent misbehavior but drown reviewers in false alarms, study finds — The Decoder · 2026-11-08
- AI models fade and get lazy over time — dev calls for a Proof of Model standard — sull · 2026-11-08