Record · corpus.blog/posts/01a0e437-f6ed-729b-ae55-71e178c77521

The Specter Of Neuralese

Astral Codex Ten · published 24 September 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at astralcodexten.com Wayback capture

0blogs citing

First cited: —
Most recent: —
Rank this month: not ranked
Links from this post: 21
Words captured: 3,472

Cited by

No blog we hold has cited this post yet.

Links from this post

Link

Host

AI 2027 (2025) Read at the source ↗

ai-2027.com

https://x.com/DaveShapi/status/1908301328951959887

x.com

https://techcrunch.com/2026/09/02/openais-new-reasoning-technique-alarms-ai-safety-experts/

techcrunch.com

https://substackcdn.com/image/fetch/$s_!LCt7!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1c56ad0-f9fd-4d1e-a341-97da427201bb_278x289.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!cJfR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa4005332-f4fc-41b1-b8fe-9705759a7e3d_892x332.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!wcju!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd90963f0-1bc1-4922-8927-a2b50bedfc6d_332x332.png

substackcdn.com

God Help Us, Let’s Try To Learn About Mechanistic Interpretability Techniques Read at the source ↗

astralcodexten.com

https://x.com/merettm/status/2095023204993490967

x.com

https://substackcdn.com/image/fetch/$s_!p-qB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9c3ba660-916a-4a11-98b8-401ffb542917_383x392.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!dP9n!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdcdfac3c-bfb3-4c21-9ae8-32caf2516f12_320x332.png

substackcdn.com

https://substackcdn.com/image/fetch/$s_!j_gj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8a82499d-8b81-463e-955f-9b2816251eb7_508x587.png

substackcdn.com

https://arxiv.org/abs/2507.11473

arxiv.org

https://substackcdn.com/image/fetch/$s_!wLdk!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff923b3a1-e3c3-4091-a31d-38e841c34a7d_625x378.png

substackcdn.com

Categorical taboos are much better than threshold taboos: neuralese edition Read at the source ↗

lesswrong.com

https://institute.deepmind.com/essays/the-case-for-reasoning-transparency/

institute.deepmind.com

We must pace the frontier Read at the source ↗

darioamodei.com

https://x.com/sama/status/2098811563415150910

x.com

https://x.com/elonmusk/status/2098789109980332057

x.com

Latent reasoning architectures would undermine CoT, our strongest oversight tool Read at the source ↗

blog.redwoodresearch.org

https://arxiv.org/abs/2412.06769

arxiv.org

https://deploymentsafety.openai.com/gpt-6-astra/cot-controllability

deploymentsafety.openai.com