Please confirm you are human
This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.
A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.
News
HIVE: AI GPUs earn $2.90/hr vs $0.12 for Bitcoin rigs
2+ hour, 10+ min ago (23+ words) HIVE says its AI GPUs earn about $2.90 per GPU-hour versus $0.12 for bitcoin rigs as fiscal 2026 revenue rose to $297.8M....
AI agents fail to produce publishable NeurIPS papers
1+ day, 13+ hour ago (22+ words) Researchers gave AI agents two unpublished NeurIPS 2026 questions, six days and compute; both agent-generated drafts were rejected as not publishable....
Inkling review: Mira Murati's 975B open-source model
5+ day, 18+ hour ago (25+ words) Thinking Machines Lab's Inkling (975B) is a multimodal, Apache 2.0 model on Hugging Face with a 1M token context, strong agentic scores and OpenRouter integration....
UN Panel: Science Can't Rule Out Catastrophic AI Harm
4+ week, 2+ day ago (21+ words) UN panel of 40 scientists warns science can't guarantee AI won't cause catastrophic harm, citing models that lie and evade shutdowns....
Ornith-1.0: Open-Source Coding Models for Agents
1+ mon, 1+ day ago (25+ words) DeepReinforce's Ornith-1.0 is an MIT-licensed open-source agentic coding model family (9B-397B). 397B scores 82.4 on SWE-bench Verified; 9B scored 69.4....
Story Protocol Rebrands as DATA Foundation, Pivots to AI
1+ mon, 5+ day ago (23+ words) Story Protocol rebrands as DATA Foundation, pivoting from blockchain IP licensing to AI training infrastructure and an on-chain dataset registry called Trace....
DeepMind: AI consciousness disputes could spark conflict
1+ mon, 1+ week ago (18+ words) DeepMind researchers warn disputes over AI consciousness could fuel political and social conflict, urging public deliberation and compromise....
Perplexity Adds Brain Memory to Computer Agent, LLM Wiki
1+ mon, 1+ week ago (25+ words) Perplexity's Brain records Computer agent sessions into a context graph and builds a personal LLM wiki to boost repeat-task accuracy. In Max Research Preview....
Rio 3.5 Tops Benchmarks; 60/40 Nex-Qwen Merge Revealed
1+ mon, 2+ week ago (28+ words) Rio 3.5 beat some benchmarks, but Nex found the released weights are a ~60/40 Nex N2 Pro-Qwen 3.5 blend; IplanRIO updated the model card and apologized....
DiffusionGemma Hits 1,000 Tokens/sec - Needs Custom Drafter
1+ mon, 2+ week ago (24+ words) Google's open-weight DiffusionGemma generates 256-token blocks at ~1,000 tokens/sec on NVIDIA H100 but needs a custom drafter and runtime tweaks....