The BaseEssayLessWrong

"Holy shit" graphs and how to get more of them

According to the excerpt, June Jimenez argues the most powerful models may be internal ones never meant for release, so gaps between internal and public capability can distort how the public and regulators see frontier risk. The post proposes short-term ways to improve visibility.

Open the source ↗

Why it matters

The same thing, for different readers.

For lawmakers

Argues policymakers may be judging frontier capability from models that are behind the internal frontier, and suggests low-cost ways to improve visibility.

For researchers

Highlights graphs, such as the METR time horizon, as a way to communicate capability trends.

The record

Questions
Who can demand evidence, delay a release or stop an unsafe deployment? · What happens when AI helps build more capable AI?
People
June Jimenez (Wrote it)
What we read
We have read only the description the publisher gives. No quotations until someone reads the source itself.
Found by
The Radar, run of October 18, 2026, via lesswrong (tier 1: the source’s own channel or an official record)
Cite this
LessWrong, “"Holy shit" graphs and how to get more of them”, October 18, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/lesswrong-2026-10-18-holy-shit-graphs
Our notesLight lane: opening only. The specific proposals and graphs are not visible.