CLS

Richard Socher Verified account @RichardSocher · Oct 21

Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.

• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from scratch in under 4 months

Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.

View quoted post on X

alphaXiv Verified account @askalphaxiv · 6d

Flint, Recursive’s new 40B open model, can run interpretability experiments on itself!

It autonomously found a goalkeeper-related neuron, then amplified its output direction resulting in every new response mentioning “goalkeeper"

View quoted post on X

Laude Institute @LaudeInstitute · 6d

Collaborative Intelligence for the Future of Work | @Stanford

@Diyi_Yang, @tatsu_hashimoto, @chrispiech, and @erikbryn are building AI systems that complement human expertise to equip the workforce of the future.

laude.org/moonshots/one/…

View quoted post on X

Ricursive Intelligence Verified account @RicursiveAI · Nov 4

Today we release Ripple-65B Preview, our 65B dense LLM we pretrained from scratch.

Ripple sets a new state-of-the-art for U.S. open-weight models in agentic coding tasks, surpassing significantly larger models including Muse Spark Open-816B, Beam-501B, Inkling-975B, and Nemotron 3 Ultra-550B.

Learn more: ricursive.com/ripple

View quoted post on X

John Langford @JohnCLangford · Nov 3

The Bytedance folks released Recycler seed.bytedance.com/en/public_pape… which is the Full Bandwidth Transformer arxiv.org/abs/2608.08888 up to inessential details.

Experiments are more comprehensive, cover RL, work at larger scale, and compare across model sizes.

Nice work! :-)

View quoted post on X