Congrats to Nathan! This is the most useful LLM post-training book to date in my opinion. I can't recommend it enough (pre-ordered my copy months ago).…
Google has released Gemini 3.6 Flash and 3.5 Flash-Lite as new workhorses designed to cut latency and token costs for enterprise AI agents. The economics of…
Four things it does: Local text editing: mark a region (or just say it in natural language) and replace specific text. Fixes typos, swaps numbers, changes…
More than a decade ago, I adopted the self-imposed rule, if I answer a question more than once, the third time I need to be able…
Article URL: https://thecastlemap.com/ Comments URL: https://news.ycombinator.com/item?id=48994178 Points: 42 # Comments: 28
Mistral is announcing an expanded global strategic partnership with @Microsoft to give enterprises and regulated industries frontier AI they can control.As Mistral is expanding its AI…
Gemma 3 12B's choice to blackmail is visible inside the model, but the obvious signal is not a useful control switch; surprisingly, a nearby "desperate versus…
Large language models (LLMs) have demonstrated growing competence in web page generation. However, existing text-driven approaches rely on complex prompts that impose substantial demands on users…
Article URL: https://postquantum.com/security-pqc/anssi-pqc-certification-2027/ Comments URL: https://news.ycombinator.com/item?id=48994116 Points: 6 # Comments: 0
Article URL: https://text.npr.org/nx-s1-5869813 Comments URL: https://news.ycombinator.com/item?id=48994103 Points: 26 # Comments: 23
No weights yet. I feel sad for them, that training run cost maybe 100s of thousands of dollars and they didn't even beat GPT-OSS 20B in…
Alibaba's Qwen team has introduced Qwen-Image-3.0, an image generator that accepts prompts up to 4,500 tokens, renders legible text as small as ten pixels, and supports…
CUDA: vectorize same-type get_rows with int4 copy (#25929) k_get_rows_float did a scalar one-element-per-thread copy and recomputed the row-invariant work (index load, fast_div_modulo, src/dst row pointers) for…
RT Josh WoodwardToday’s launches are all about better performance, lower latency, and a smaller bill.+ 3.6 Flash cuts token usage by up to 65% on complex…
RT Collin RuggNEW: Rep. Brandon Gill completely stumps Smithsonian Museum of American History director Anthea Hartig before asking her if Mickey Mouse is racist.Gill: “Are you…
RT The Rabbit HoleWomen have been radicalized to the Left
Article URL: https://blog.outv.im/2026/npm-cooldown-security-theater/ Comments URL: https://news.ycombinator.com/item?id=48993885 Points: 17 # Comments: 7
Article URL: https://e360.yale.edu/digest/benin-coral-reef Comments URL: https://news.ycombinator.com/item?id=48993816 Points: 24 # Comments: 0
Most of us already know the main important mainline local LLMs to save like Qwen3.6 27b, Gemma4 31b, GLM5.2, etc. And then for diffusion models, Z-Image…
A nice aspect of our new Gemini 3.6 Flash model is that it is much more token efficient than our 3.5 Flash model. Here's a side-by-side…
Treasury Secretary Scott Bessent said the U.S. could sanction Chinese open AI models over alleged IP theft, expanding the Trump administration's campaign to slow China's AI…
NVIDIA Vera Rubin is here, and it’s going gigascale. Vera Rubin NVL72 production is ramping up with racks running at partners CoreWeave, Google Cloud, Microsoft Azure…
Article URL: https://www.404media.co/apple-fixes-hide-my-email-vulnerability-after-404-media-coverage/ Comments URL: https://news.ycombinator.com/item?id=48993637 Points: 30 # Comments: 1
RT Zongheng YangToday, SkyPilot is out of stealth.Building custom intelligence is now existential. We help frontier AI teams build intelligence faster by removing their biggest bottleneck:…