WoolyAI released a new private multi-agent inference server for DGX Spark clusters, achieving up to 90.83 decode tokens per second on the Nemotron 3 Nano Omni 30B model without quantization or speculative decoding in a 2…
A 2024 PDF report titled 'Bing Copilot (ChatGPT-4) Flunks Math' claims that Microsoft's Bing Copilot, powered by OpenAI's GPT-4, fails to correctly solve a range of mathematical problems. The document presents evidence o…
Spacebar, an infrastructure platform for real-time multimodal AI agents, has been unveiled by founders Ayush Agarwal and Amogh Asgekar. The model-agnostic runtime combines audio, video, cursor activity, and canvas object…
AMD and Cerebras Systems announced separate AI chip deals on Monday, with AMD securing a $1.5 billion contract from an unnamed hyperscaler and Cerebras landing a $100 million agreement with a Middle Eastern oil company f…
Filmmaker Neill Blomkamp's new 13-minute short film 'Nightborne,' generated almost entirely using ByteDance's Seed 2.0 video model, has received largely negative reception from fans, who criticize it as 'soulless slop.' …
Tesla missed analysts' profit forecasts for the second quarter and reported negative free cash flow for the first time in over two years as Elon Musk's AI spending surged. Adjusted profit was 33 cents per share, below th…
AI-generated resumes sound generic because cheap 'unlimited' models fall back on buzzwords like 'results-driven professional' and 'leveraged cross-functional teams,' creating a resume monoculture that recruiters easily s…
MarkTechPost published a tutorial on EdgeBench, a benchmark for evaluating AI agents across task categories, runtime environments, and interaction-time budgets. The analysis covers dataset download from Hugging Face, tas…
Meridian Labs and Anthropic have developed an extension for the open-source AI safety evaluation framework Petri that enables multi-agent evaluations, addressing a key limitation of the original single-agent system. The …
A 5-day project sprint during ARENA 8.0 found that prompt rephrasing can measurably impact alignment-relevant properties even on frontier models, though data were noisy and inconsistent across evals, models, and variatio…
A developer built a state-management workflow for long AI-assisted projects, featuring three modes: Normal, Quick Check, and Lockdown. The workflow is designed to help users track progress and avoid losing context in ext…
Jeff Cui and collaborators announced Patch Policy, a robot-control method that beats OpenVLA-OFT with 0.7% of its parameters and trains on a single NVIDIA RTX 5090 GPU. The approach freezes a Vision Transformer pretraine…
MergeForge, an open-source VS Code and Cursor extension, brings a JetBrains-style three-pane conflict resolver to the editors, paired with an AI assistant that reads the repository context before suggesting fixes. Built …
A Twitter post by user joefioti highlights the dominance of a key player in the AI chip market, though the specific company or chip is not named in the available text. The post has garnered 1 point on Hacker News with no…
NotchAgent, a native macOS notch control surface for AI coding agents running under herdr, is now available. The app provides a color-coded status line below the MacBook notch, allowing users to monitor, approve, deny, a…
Onibex ASK, an open platform from the Onibex team, turns natural language into governed SQL over SAP data using a curated semantic layer instead of raw database schemas. The system employs a Bronze/Silver/Gold medallion …
Google burned through cash last quarter as it ramped up spending on artificial intelligence infrastructure, according to the Financial Times. The company's capital expenditures surged as it invested heavily in data cente…
AMD announced plans to sell tens of billions of dollars worth of AI servers to Anthropic and invest up to $5 billion in the AI startup, with delivery of AMD's Instinct MI450 AI servers expected to begin in the first half…
AI tools have become the third most-used channel for product research among consumers, behind search engines and retailer websites or apps, according to a PwC survey of 1,000 consumers. The survey found 44 percent of con…
Google Research published a first-of-its-kind national-scale study (n=13,917) showing that a conversational AI agent powered by Gemini Flash 2.0 can conduct end-to-end symptom interviews and differential diagnosis with a…