Every story across every category, newest first. Each card links to the original publisher; daily-brief posts open as editorial pages.
meh. Mythos 2 will build new mature node fabs tooor do everything with 14A. Lots more chips per wafer that way. It'll probably invent molecular nanoassembly…
Article URL: https://www.ft.com/content/09a62ed4-16af-433c-adb7-c877d1975388 Comments URL: https://news.ycombinator.com/item?id=48828648 Points: 29 # Comments: 24
AI chip maker SambaNova has raised at an $11B valuation months after Intel was rumored to be trying to buy it for about $1.6 billion.
In this tutorial, we explore NVIDIA's cosmos-framework from a practical Colab angle while staying honest about the hardware needed for real Cosmos 3 checkpoints. We probe…
LLM agents increasingly rely on retrieval buffers to store and reuse past experience, yet the cache management policies governing these buffers remain largely ad-hoc. We formalize…
"HalluSquatting" weaponizes LLMs' inability to say "I don't know."
CUDA: Fuse MMVQ post-scale for NVFP4 (#24481) CUDA: Fuse MMVQ for NVFP4 and BS 1 TODO: Add tests to test-backend-ops (did verify correctness manually for one…
Note: the modeling assumptions and conclusion are Thomas Kwa’s opinion, and others at METR disagree.1 Also, the math was checked by Claude but not a second…
RT Yun-Ta TsaiHas been my regular work horse for weeks. Impressive indeed.Elon Musk: Based on strong positive feedback from customers in our beta test program, @SpaceXAI…
My standards for V4-Pro-Official are: match GLM 5.2 on net. I'll accept something like slightly worse DeepSWE for slightly better CritPT/FrontierMath, on merit of their later…
Based on strong positive feedback from customers in our beta test program, @SpaceXAI will make Grok 4.5 available to the public tomorrow.It is an Opus-class model,…
My big takeaway is that both Sol & Fable represent jumps over previous models and have opened a large gap with the next-best AIs. People will…
America, the BeautifulDustin: Elon Musk just said the one thing about America they made sure you’d never learn. The one thing that should’ve made you proud,…
Below are six story prompts that you might enjoy. Especially intended for my writer friends, but hopefully other people might find it fun to try writing…
> •2027 capes est: US$73bn (below consensus $80-85bn), likely capped by EUV supplyLPP EUV is a very annoying technology, doesn't scale worth a damnthough maybe europoors…
Our new open-source AI model for Ancient Egyptian hieroglyph translation. Available in two versions: Horus Hiero 9B Horus Hiero Mini 4B (optimized for CPUs and mobile…
State-of-the-art single-image 3D reconstruction methods often rely on complex hybrid architectures and loss functions, or compress geometry into latent spaces in order to leverage pre-trained latent…
On-policy distillation (OPD) trains a student policy by matching a stronger teacher on the student's own trajectories, offering a promising framework for language agent training. However,…
**xAI** publicly launched **Grok 4.5**, a new coding-and-agents-focused frontier model emphasizing capability-per-dollar rather than benchmark supremacy. Elon Musk described it as "Opus-class" but faster, more token-efficient,…
Article URL: https://noma.security/blog/gitlost-how-we-tricked-githubs-ai-agent-into-leaking-private-repos/ Comments URL: https://news.ycombinator.com/item?id=48827858 Points: 20 # Comments: 4
RT jasonthe booth was a great investmentTheo - t3.gg: I would like to add that OpenAI has been awesome to work with despite the obvious government-related…
I understand that it may not be appropriate to call it “officially accepted” yet because of the wording used in the notification, and I also saw…
Hierarchical Vision-Language-Action (VLA) models decouple high-level planning from low-level control to improve generalization in robot manipulation. Recent work in this paradigm uses 2D end-effector trajectories predicted…
fix: OOB reads in UGM tokenizer (precompiled_charsmap handling) Validate minimum size (4 bytes) before reading xcda_blob_size Use strnlen with bounds check instead of unsafe strlen Both…