Foxconn is wiring hundreds of AI agents into its production lines through a new system called MoMClaw — projecting ~80% faster fault diagnosis.
A biotech betting that aging can be reprogrammed at the cellular level just got one of the largest war chests in longevity.
Salesforce has agreed to buy Contentful, the headless content platform used by 4,800+ brands, to give Agentforce a content layer.
Asana has agreed to acquire StackAI, a no-code AI-agent platform, pushing the work-management vendor from organizing work to executing it.
At ServiceNow Knowledge 2026, NVIDIA and ServiceNow set out to answer the harder enterprise question: how should AI act, and who is accountable?
Google DeepMind released DiffusionGemma, an experimental open model that drafts whole text blocks at once instead of word-by-word.
When Anthropic teased a model called 'Fable,' the name suggested a storytelling engine. The model that shipped is almost the opposite.
A claim making the rounds would, if true, rewrite the economics of building LLMs. After checking, it doesn't hold up to verification.
A brain MRI can take days to read. Michigan researchers built an AI that compresses that wait to seconds.
The AI boom has a thermodynamics problem. An MIT spinout thinks the fix has been hiding inside nuclear reactors all along.
A new MIT Media Lab study adds an unsettling entry to the ledger of skills we are outsourcing to machines: spotting fake news.
Understanding a 3D world that is also changing through time is brutally hard for machines. D4RT pushes that frontier furthest this year.
China is preparing to spend roughly $295 billion over five years to build a nationwide network of interconnected AI data centers.
Within five days, Anthropic urged a coordinated global slowdown of frontier AI — then shipped its most powerful model to anyone with an API key.
The Pentagon is evaluating AI models from OpenAI, Google and xAI as potential replacements for Anthropic's Claude in classified systems.