#AI

Researchers using a machine-learning pipeline called RAVEN report more than 100 newly validated planets and over 2,000 vetted candidates from TESS data, highlighting how AI is reshaping the search for worlds beyond the太阳

Google has quietly released an iOS dictation app built around on-device speech recognition, signaling a more direct push into AI-assisted writing tools that work even without a cloud connection.

Anthropic says a preview version of its upcoming Mythos-class AI model will be used by major technology and open-source organizations to identify vulnerabilities in critical software infrastructure.

OpenAI, Anthropic, and Google are reportedly sharing information through the Frontier Model Forum to detect unauthorized distillation of their AI systems by Chinese competitors.

Venture firm Eclipse has closed $1.3 billion in fresh capital and says it will use the money to back and build startups that apply AI to transportation, energy, infrastructure, compute, and defense.

Arcee, a 26-person U.S. startup, is promoting a new reasoning model called Trinity Large Thinking as it tries to offer Western companies an open-weight alternative to both Chinese and closed-source AI systems.

Singapore-based Firmus has raised another $505 million at a $5.5 billion post-money valuation, underscoring investor appetite for AI infrastructure and for data center projects built around Nvidia’s next-generation plans

A new analysis found Google’s AI Overviews answered benchmark questions correctly about nine times out of 10, but the remaining error rate could still translate into a vast volume of wrong answers.

MIT Technology Review reports that some economists now see current measures of AI job exposure as too weak to predict displacement, and are calling for new worker-level data before policy falls further behind.

Tools such as Alibaba’s Accio are compressing product research and factory discovery into a chat workflow, giving small sellers a faster route from idea to launch.

A report on Suno says the AI music platform’s copyright safeguards can be bypassed with minimal effort, raising new questions about imitation, monetization, and enforcement in generative music.

According to the supplied candidate metadata, researchers from the University of South China and Purdue University used artificial intelligence to create ultra-strong, rust-proof steel aimed at 3D-printed components.

Folk musician Murphy Campbell found AI-generated tracks uploaded under her name and then faced a copyright claim over public-domain material, exposing failures in both platform verification and online enforcement.

Netflix has open-sourced VOID, a video AI framework that removes objects from scenes and adjusts the physical interactions those objects originally caused.

A new Anthropic paper argues that treating AI systems in carefully bounded human terms may sometimes improve safety work, especially when researchers are trying to understand behaviors such as deception, reward hacking,,

Meta has paused work with data vendor Mercor after a security incident, while other AI labs assess whether proprietary training-data workflows were exposed.

A newly published Science paper points to a machine learning approach aimed at discovering prokaryotic immune systems, underscoring how AI methods are being applied to one of biology’s most active frontiers.

Generalist AI says its new GEN-1 system is a general-purpose model for physical AI and a major step toward its goal of building general intelligence for the physical world.

DuckDuckGo’s privacy-focused chatbot posted a sharp traffic jump in February, suggesting that data handling has become a more visible competitive factor in consumer AI.

The cancellation and discontinuation of a novel suspected of heavy AI use is intensifying pressure on publishers and agents who admit current detection tools are far from reliable.