robin elisha robinson retweeted
1/4
Meet Qwenputer: Qwen3.5-4B with a full RISC-V CPU inside its weights, using the model's own Gated DeltaNet layers as RAM.
It runs C and MicroPython in its forward passes, no external runtime.
On DeepMind math: base 55–90%, base + Python tool 95–100%, Qwenputer 97.5–100%.
robin elisha robinson retweeted
At @lanyon_ai, we’re completing Hilbert’s (other) dream.
From November 1st, we’re rolling out pre-release versions of Lanyon to early customers and testers, and will be hosting simultaneous public demos of Lanyon capabilities in San Francisco and Chicago.
From moon bases to modular reactors to atomic-scale programmable matter, we're on a mission to make the abstract concrete, to make the digital physical, and to formally verify reality. Join us!
robin elisha robinson retweeted
I hope you continue to have wonderful adventures!
Two years ago today saw the premiere of the special commemorative anniversary episode, ONE PIECE FAN LETTER! 🏴☠️ 💌
Repost and share your favorite part to celebrate! 🎉
robin elisha robinson retweeted
We placed the fruit fly brain inside Odyssey-3, a frontier world model.
The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
robin elisha robinson retweeted
“there aren’t any cryptography results”
“there aren’t any cryptography results”
robin elisha robinson retweeted
Day 3/
The big one is GPT-6 in Chat, but today is also a little celebration day with a new high of 40M active users across Codex and ChatGPT Work.
Loading a banked reset in everyone's paid accounts. See you again tomorrow!
Roundup of Day 2/
2.1/ Approve for me (auto-review) is now included and does not use usage. Can be between 2-10% of plan when used. Also better for you.
2.2/ Simplified API for builders.
2.3/ Meeting notes integrated.
2.4/ Decisions API live for builders. Will use in the app to improve the experience.
robin elisha robinson retweeted
Irkutsk, Russia: The U.S. Embassy is aware of press reports from the Irkutsk region of suspected pneumonic plague resulting in the death of an individual and of associated quarantine measures in, and closures of, hospitals in Irkutsk. We continue to track the situation closely. The U.S. government has limited ability to help U.S. citizens in Russia, especially outside of Moscow. The State Department’s Level 4 Travel Advisory (Do Not Travel) for Russia remains in place and states U.S. citizens currently in Russia should depart immediately. U.S. citizens planning travel to Russia are again strongly advised not to travel to Russia due to risk of wrongful detention and security concerns. More at: ru.usembassy.gov/health-aler…
robin elisha robinson retweeted
Small bird, fast wings, Kolibri is here.
78B parameters. 3.46B active. Up to 1M tokens of context. Built in Europe.
Now the weights are yours. Run it on your own hardware, under Apache 2.0.
robin elisha robinson retweeted
First time I’ve seen this one
Chinese chain of thought in official Claude app, Fable 5.1
Reminds me of my friend GLM 5.3 flash, he thinks in Chinese a lot too
robin elisha robinson retweeted
Denzel explica por qué usa la IA.
Un pequeña prueba explorando Minimax H3 en ComfyUI con mis nodos y métodos de inpainting.
robin elisha robinson retweeted
GALNERYUSのニューアルバム、"A CRY FROM THE SKY ABOVE" 発売まであと少し‼️詳細はこちらからどうぞ❗️
galneryus.jp/
アルバムからのリードトラックMV
"THE LIGHT OF HOPE"
youtu.be/7dmupD_uqc0
robin elisha robinson retweeted
Fable with low reasoning is roughly comparable to GPT-5.5 medium on the 20 puzzles they both solved.
Fable scored higher in 14 of those, and I think Warming Tonic illustrates how different quality alchemical machines look. Solve rate isn't everything, quality is important!
Are AI agents shape rotators? In this new benchmark, we let the models play campaign puzzles in Opus Magnum, a puzzle game by @zachtronics.
Ironically, Claude Opus 4.8 performed poorly, being beaten by GPT-5.5, Gemini 3.5 Flash, and GLM 5.2. Claude Fable 5 crushed them all.



























