Explore
Enter a keyword in the search bar above to find articles, or use the filters to browse.
3,592 results • Page 78 of 300
Following its liftoff from Cape Canaveral on July 21 aboard a SpaceX Falcon 9 rocket, the Mission Robotic Vehicle (MRV) hosting the NASA-supported Robotic Servicing of Geosynchronous Satellites (RSGS) payload is now en route to geosynchronous Earth orbit, where it will use its advanced robotics to service spacecraft. RSGS leverages in-space robotics expertise from NASA, […]
Anthropic leaped to a $47 billion revenue run rate by May, compared to $9 billion in 2025. It’s the kind of growth that Menlo Ventures’ Matt Murphy says he’s never seen in 25 years of investing, not in the internet wave, not in mobile, not in the first cloud boom. Menlo led Anthropic’s $500M Series D, and Murphy has had a front-row seat as the company went from a pre-revenue, […]
Hey HN, Henry & Roman here from Cactus. A small, on-device model is fast and private, but sometimes wrong, but frontier models are getting expensive pretty fast. So, we post-trained Gemma 4 E2B post-trained to know when it's wrong. Every response comes with a confidence score between 0 and 1. Developers can accept the on-device when it's high, hand off to a bigger cloud model when it's low. By routing only 15-35% of queries to Gemini 3.1 Flash-Lite, Gemma-4-E2B matches Gemini 3.1 Flash-Lite on most benchmarks. - ChartQA: 15-20% - LibriSpeech: 25-30% - MMBench, GigaSpeech, MMAU: 30-35% - MMLU-Pro: 45-55% We were always frustrated by the routing signals hybrid apps rely on: asking the model to rate itself in text (unreliable, and you're parsing prose), or token entropy heuristics (barely better than a coin flip in our tests). So we did mechanistic studies on small models, Gemma 4 particularly, and found the hidden state for different layers carry meaningful self-awareness signal for various situations. SO we extended the model with a 68k params probe layer (LayerNorm, low-rank projection, attention pooling, small MLP head) reads one intermediate layer during decoding and predicts p(wrong); confidence = 1 - p(wrong), returned as structured data, never parsed out of the answer text. Across 12 hold-out benchmarks spanning text, vision and audio, the probe averages 0.814 AUROC vs 0.549 for token entropy. The result that convinced us this is real: the probe was trained on zero audio data, yet scores 0.79-0.88 AUROC on four audio benchmarks where entropy is near-random or worse (0.32-0.52). It's reading a modality-independent correctness signal from the hidden state, not memorizing patterns from its training data. We published all weights on HuggingFace and provide copy-pase codes to run it on Transformers, MLX, Llama.cpp or Cactus. With Ollama, vLLM, SGLang etc in the works. For llama.cpp we ship a patch series you compile in once (upstreaming is planned). The code is MIT licensed; Gemma model use remains subject to the Gemma terms. GitHub: https://github.com/cactus-compute/cactus-hybrid Weights: https://huggingface.co/collections/Cactus-Compute/cactus-hyb... Some caveats: - The probe scores single-sequence decoding only, up to the first 1024 generated tokens. - Handoff works best when routing per task in a multi-step process, not per step. - Hierarchical routing is still in the works: try on-device, then DeepSeek v4 Flash, before Fable/GPT5.5/Gemini/Muse/Grok. - The technique is boutique for each model, we will share each weights as they roll out. These issues are currently being tackled at Cactus and updated weights will be shipped directly into the HuggingFace collection and GitHub repository straight up. Please let us know your thoughts, it helps us find ways to improve the design progressively. Thanks a million!
The New York City mayor instead called on US authorities to act on International Criminal Court's arrest warrant against the Israeli prime minister.
The company said it is reducing its headcount by 20%, or about 630 staff, to "support a leaner, more focused operating model" as it focuses on its AI Work Platform.
Out of Dyson’s two new affordable stick vacuums, this one is easily my favorite.
626 points 235 comments on Hacker News · mitchellh.com
Need more screen space on the go? The AOC Portable Monitor V50 is down to $99.59, with Full HD and a travel-ready design.
A digestion of the Jacobian conjecture counterexample - https://news.ycombinator.com/item?id=48998362 - July 2026 (133 comments) Claude Fable produced a counterexample to the Jacobian Conjecture - https://news.ycombinator.com/item?id=48973869 - July 2026 (508 comments)
Touch Surgery bridges pre-operative planning and training, intra-operative tele-mentoring and tele-proctoring, and post-operative insights. The post Medtronic to launch AI compute platform for the operating room appeared first on The Robot Report .
604 points 119 comments on Hacker News · github.com
663 points 240 comments on Hacker News · dylancastillo.co
Hacker News
“GPT-5.5's API pricing is reshaping how startups build AI products”
