Moonshot AI's use of model distillation to create high-performance small models from larger ones like GPT-4 has sparked debate over intellectual property, but the process involves training a student model on the teacher'…
Building production-ready AI products requires far more than calling an API, according to an experienced engineer. The actual LLM interaction accounts for only 10% of the work, while 90% involves infrastructure and orche…
AI companies are buying and destroying physical books at massive scale to train large language models, with one database company, ISBNdb, now facilitating bulk purchases of 1,000 to one million books per order. The pract…
AMD announced an integrated platform for robotics and autonomous systems, targeting physical AI computing with standardized architectures and open ecosystems to simplify development and deployment across markets.
Y Combinator is calling for startups to build an AI-powered adaptive tutor for young children, inspired by the Primer from Neal Stephenson's The Diamond Age, aiming to deliver one-on-one tutoring quality at consumer scal…
A developer building LLM agents with the DevSwarm architecture, which splits tasks across specialized open-weight models including GLM-5.2, Qwen3-Coder-480B, and Kimi-K2.7-Code, implemented self-hosted SigNoz observabili…
A developer outlines a simple four-step pattern—trigger, data, action, deliver—for building reliable AI agents, arguing that most tasks don't need an agent and that the real engineering lies outside the LLM call. The pos…
Nvidia and SK Group are launching a more than $500 billion AI initiative in South Korea, including a 2-gigawatt AI data center from SK Telecom powered by Nvidia's Vera Rubin chips and SK Hynix's HBM4 memory, with the fir…
Librarians Charlie Bailey in Philadelphia and Hannah Cyrus in Maine are hosting viral 'Avoiding AI' workshops that teach people how to disable AI features like Apple Intelligence and Gemini on their devices. Bailey's wor…
Microsoft has open-sourced its ML Video Codec (MLVC), a machine learning-based video compression technology designed to improve efficiency and quality. The release on GitHub includes the codec's source code, models, and …
Ringup, a new service for voice agents, lets callers pay and be recognized by phone number alone, charging saved cards in seconds without reading digits aloud. The system integrates with 144+ payment processors and settl…
Aniruddha Adak, an AI Agent Engineer and Full-Stack Developer, merged 373 pull requests across open source projects during a bug-smashing spree. Notable fixes include a critical security vulnerability in the cognee AI me…
A growing number of professionals are using AI transcription tools like Granola to record conversations without informing the other party, according to a Wall Street Journal report. Granola, which raised $125 million at …
Principles of Intelligence (PrincInt, formerly PIBBSS) is launching PIRAMID, an internal research division that uses statistical physics to build scientific foundations for ambitious mechanistic interpretability. The div…
A growing number of people are using AI to generate fake evidence and deceive the justice system, according to a report by El País. The article highlights cases where AI-generated documents, images, and audio have been s…
Farmers in Paraná, Brazil, tokenized 10 dairy cows on the B3 stock exchange, raising nearly $20,000 in credit backed by their livestock. The project by agtech firm Cowmed uses AI-powered tracking collars to create encryp…
AWS EC2 compute is evolving to meet demand from agentic AI and physical AI workloads, with customers using AMD-based instances for AI inference alongside traditional HPC tasks, according to Art Baudo, principal product m…
A developer created Agent-Up, an open-source desktop app and local server for managing multiple coding-agent environments on one machine. It provides per-workspace process lifecycle, port allocation, Docker services, and…
Opus 5 achieved a 0% success rate against browser prompt injection in tests, compared to a 3.7% rate without protection layers, according to a report. The system's ability to distinguish between external web data and int…
Broadcom is shifting production of high-end AI accelerators from TSMC to Samsung, betting on Samsung's process nodes to match yields and power efficiency. If Samsung delivers, the move could lower AI compute costs and di…