Tech News
← Home  ·  All topics

Model

191 GoKawiil briefs on this topic

Google Home opens smart-device controls to third-party AI agents via MCP

Google has launched Google Home MCP, letting outside AI agents like Claude, Google Antigravity, Hermes, and Open Claw access and control devices in a user's Google Home ecosystem through the Model Context Protocol. The integration lets these agents view device event history, analyze camera footage across cameras, track usage patterns like laundry cycles or light usage, send voice updates through Google Home speakers, and build custom control dashboards. Gemini for Home remains the primary assistant for the Home app and Nest speakers, with MCP adding an extra layer of third-party access rather than replacing it.

UTe2 shows electrically tunable superconducting memory tied to vortex dynamics, not heating

Researchers studying the heavy-fermion superconductor UTe2 examined whether a newly observed memory effect—where past current pulses alter later superconducting behavior—could simply be an artifact of Joule heating from the applied current pulses. By analyzing temperature during pulse protocols and testing how the effect scales with current amplitude, duration, and especially the speed at which current is switched off, they found the memory effect strengthens with faster relaxation rates, which actually reduces total energy deposited, undermining a simple heating explanation.

New cosmological simulation traces black hole and 'little red dot' formation from dark matter halos

Researchers built a high-resolution dark-matter N-body simulation, seeded with Planck cosmological parameters, tracking halo mergers from redshift 35 to 7.5 using ROCKSTAR and CONSISTENT-TREES tools. They layered a semi-analytic galaxy-formation model onto this framework to simulate gas cooling, star formation, supernova feedback and radiation, showing that overmassive black holes and compact 'little red dot' galaxies emerge naturally from these early-universe conditions without requiring exotic physics.

Open-source toolkit Scikit-decide optimizes flight paths to cut jet fuel costs

An open-source project called Scikit-decide, combined with TU Delft's OpenAP fuel consumption model and NOAA wind data, can calculate more fuel-efficient flight paths for airliners. The author tested it against a commercial fuel-cost example of a Boeing 787-9 flying nonstop from Newark to Rome, showing how wind-adjusted routing could trim thousands of dollars in fuel spend. The comparison also included an Airbus A320 test using OpenAP's performance model on the author's own high-end workstation.

Meta launches WhatsApp Business Tools MCP for AI agent-driven setup

Meta introduced a new MCP (Model Context Protocol) server called WhatsApp Business Tools MCP, letting AI coding agents like Claude, Cursor, Codex, or ChatGPT directly configure WhatsApp Business accounts. Instead of manually navigating the Developer Console, Business Manager, and API references, businesses can now simply describe what they need to an AI agent, which handles account creation, phone verification, Cloud API registration, and template management.

GraafHenk releases WangNet, a 1.8MB neural network that judges 'Numberwang'

A developer known as GraafHenk has published an open-source project called Numberwang, built around a tiny 1.8MB JSON model and roughly 100 lines of pure Python with no external dependencies. The tool parses digits, number words, arithmetic expressions, Roman numerals, currencies and more across eleven languages, then returns one of four comedic verdicts inspired by the British sketch of the same name, including 'Numberwang' and 'Wangernumb'. It runs via command line or can be imported directly as a Python function returning probability scores for each verdict.

TypeSafe AI launches Jev, a decision-focused model claiming 40-400x lower cost than LLMs

TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, has released Jev in early access, a new type of AI model designed to make fast structured decisions rather than generate text. The company says Jev matches existing large language models on decision-making tasks while running 20-200 times faster and 40-400 times cheaper, using a new architecture and a training method called Reinforcement Learning for Calibrated Decisions.

Gut infections in mice train immune memory cells that settle in brain's protective lining

A mouse study led by Cambridge researcher Menna Clatworthy found that immune cells activated during intestinal infections, such as those caused by Citrobacter rodentium bacteria and the parasitic worm Schistosoma mansoni, travel to the meninges surrounding the brain. There, the cells take up long-term residence and mount a faster, stronger response when the same infection recurs more than a month later, indicating they retain a memory of the original illness.

Anthropic's Amodei and OpenAI's Altman call for coordinated 'pacing' of AI development

Anthropic CEO Dario Amodei published an essay proposing that frontier AI companies adopt independent evaluators, coordinate on safety standards with peer firms in democratic countries, and eventually seek global coordination including authoritarian governments, while stressing this does not mean halting AI progress. OpenAI's Sam Altman quickly echoed the sentiment, agreeing on the need to 'pace the frontier,' endorsing third-party evaluators, and calling for a federal framework establishing consistent safety requirements across the industry.

Enterprise clients like Nvidia, Palantir, C Spire tighten limits on OpenAI and Anthropic tools

Large corporate customers of OpenAI and Anthropic are pushing back over how their data is handled, with some restricting which AI models employees can use and demanding contractual guarantees against training on their proprietary data. The friction follows a June policy change by Anthropic allowing it to retain customer data from its Claude models to prevent misuse, which some clients fear could expose sensitive business information. Companies like C Spire have secured agreements barring model training on their data, though both AI firms still collect technical usage metadata.

New Technical Book Teaches Foundation Model Engineering End-to-End

A new textbook titled Foundation Model Engineering has been released, aimed at AI engineers and research-minded readers who want to understand foundation models beyond basic API use. It ties together topics like attention, mixture-of-experts, RLHF, multimodality, long-context inference, retrieval-augmented generation, and agents into a single engineering narrative, using PyTorch examples, quizzes, and interactive visualizers.

TechCrunch Disrupt 2026 to host panel on AI platform risk with Airbyte, Radical Ventures, Webflow

TechCrunch Disrupt 2026, running October 13–15 at Moscone West in San Francisco, will feature a Builders Stage session titled 'What Happens When OpenAI Ships Your Roadmap.' The panel brings together Michel Tricot of Airbyte, Rob Toews of Radical Ventures, and Linda Tong of Webflow to discuss how AI startups navigate the risk of foundation model providers absorbing their core features. The event is expected to draw over 10,000 founders, investors, and operators across 250+ sessions.