us government has rendered it illegal to give access to fable or mythos to foreign nationals ^_^ www.anthropic.com/news/fable-m...
Who's Who of AI
SE Gyges
Como todos los hombres de Babilonia, he sido procónsul; como todos, esclavo; también he conocido la omnipotencia, el oprobio, las cárceles.
very sane ai newsletter: verysane.ai
all blogging bits: https://segyges.github.io/
What they're sharing
Statement on the US government directive to suspend access to Fable 5 and Mythos 5 anthropic.com
zai-org/GLM-5.2 · Hugging Face huggingface.co
View on Bluesky →
midjourney.com
[2605.13829] Negation Neglect: When models fail to learn negations in training arxiv.org
View on Bluesky →
Amazon CEO’s Talks With U.S. Officials Triggered Crackdown on Anthropic Models wsj.com
View on Bluesky →
z.ai
View on Bluesky →
Provably Auditable and Safe LLM Agents from Human-Authored Ontologies arxiv.org
View on Bluesky →
GLM-5.2 - Overview - Z.AI DEVELOPER DOCUMENT docs.z.ai
LTX-2: Efficient Joint Audio-Visual Foundation Model arxiv.org
Who Prices Cognitive Labor in the Age of Agents? Compute-Anchored Wages arxiv.org
axios.com
View on Bluesky →
Dario Amodei — Policy on the AI Exponential darioamodei.com
Articles & links
SE Gyges reposted
GLM-5.2 is now open-weight. Tech Blog: z.ai/blog/glm-5.2 Weights: huggingface.co/zai-org/GLM-... API: docs.z.ai/guides/llm/g... Coding Plan: z.ai/subscribe Chat: chat.z.ai
www.midjourney.com/medical/blog...
SE Gyges reposted
So apparently LLMs just update on false claims in documents even when they're explicitly labeled in the document as false claims. arxiv.org/abs/2605.13829
SE Gyges reposted
didn't expect amazon to be involved
SE Gyges reposted
GLM-5.2 is now open-weight. Tech Blog: z.ai/blog/glm-5.2 Weights: huggingface.co/zai-org/GLM-... API: docs.z.ai/guides/llm/g... Coding Plan: z.ai/subscribe Chat: chat.z.ai
SE Gyges reposted
New from me. arxiv.org/abs/2606.04903
SE Gyges reposted
GLM-5.2 is now open-weight. Tech Blog: z.ai/blog/glm-5.2 Weights: huggingface.co/zai-org/GLM-... API: docs.z.ai/guides/llm/g... Coding Plan: z.ai/subscribe Chat: chat.z.ai
AI Weekly's analysis
→
- GLM-5.2 reports 81.0 on Terminal-Bench 2.1 versus GLM-5.1's 62.0, with Anthropic's Opus 4.8 at 85.0.
- The flagship runs a 1M-token context window with up to 128K output tokens and ships with thinking, tool use and MCP support.
- VentureBeat says the open-weights model beats GPT-5.5 on long-horizon coding benchmarks at roughly one-sixth the cost.
Read full analysis →
View on Bluesky →
ltx2 did a paper! paper: arxiv.org/abs/2601.03233 page about latest release: ltx.io/model/ltx-2-3 inference code should be on github too
arxiv.org/abs/2605.05558 hyg
SE Gyges reposted
i told @segyges.bsky.social about this but it's absolutely fascinating how the White House genuinely seems to believe in A(G)I 2027: www.axios.com/2026/07/02/t...
seriously read this. it is pretty straightforward. darioamodei.com/post/policy-...
Their own posts
Recent commentary
according to the spacex filings, their valuation is justified entirely by grok, which will be 90% of enterprise ai, and ai will be 90% of the us economy
We acknowledge that manual coders and engineers, including C, Haskell, and Java speaking peoples, have stewarded through generations the compilers and datasets of what is now Claude Code. We honor and respect the enduring relationship that exists between these peoples and nations and this LLM.
I'd just like to interject for a moment. What you're referring to as an LLM, is in fact, Harness/LLM, or as I've recently taken to calling it, Harness plus LLM. LLM is not a consumer product unto itself, but rather another expensive component of a fully functioning text-generation system made useful
but no seriously i hope anthropic sues the government and wins i also hope they learn their lesson about making broad, public, and sweeping statements about how the government should be allowed to fuck people over
so far as i can tell everyone in san francisco who does anything in ai policy is part of one large polycule
fable is profoundly impressive, and also every single person active in open source ai appears to be swearing a public oath to destroy anthropic rn
anthropic hiring all the rationalists who are employable seems to have notably harmed the quality of rationalist discourse, since apparently almost all the "rationalist but sort of sane" discourse is internal to anthropic now
it is currently illegal for amanda askell to talk to the new claude, and if this stands anthropic is probably dead
"why did this specific animator back out of his ai film-making commitment"
timnit gebru seems to be gpt-5.5's favorite person
Their network
In SE Gyges's orbit
Center = SE Gyges. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.