neuralese.wiki

Last checked

About

About neuralese.wiki

This site explains what AI models do when they reason, or talk to each other, without using words, and keeps track of what is actually known about it. It’s for people who follow AI closely but don’t do research for a living.

Editor
Alex Rostovtsev
Started
September 2026
Written with
Claude, by Anthropic

01

Who runs it

Alex Rostovtsev runs the site: choosing what it covers and reviewing what goes up. It started in September 2026, after Alex went looking for a clear explanation of neuralese and couldn’t find one written for people outside research.

Alex works in SEO and AI search, and runs two other independent projects. BLURSOR follows new research on how AI systems choose what to cite and recommend, and turns it into plain briefings and tools. WROITER is a free writing diagnostic that points out what makes a text sound machine-written and suggests a fix for each problem.

02

How it’s made

This is an AI-assisted site. Claude, an AI model made by Anthropic, does most of the research and writing. Every factual claim is checked against its original source before it goes up.

03

How to read the labels

Claims about current AI models carry one of five statuses.

Confirmed
The primary source says it on the record.
Reported
Credible journalism, not confirmed by the source.
Not established
Claimed or implied, but the evidence doesn’t support it yet.
Wrong
Contradicted by the evidence.
Open question
Nobody knows yet, or researchers disagree.

Sources carry a type, such as peer-reviewed paper or journalism. The sources page explains each type. And one rule about words: most researchers don’t call their methods neuralese. The site uses it as an umbrella and keeps each paper’s own terms for the details.

04

Corrections

If something here is wrong, email hello@neuralese.wiki. We’ll fix it and note the change, with its date, on the page.