Tag: Microsoft
Microsoft and Alphabet both trim the edge off their spending plans
Capex guidance came in under analyst estimates for the first time in several quarters, and neocloud stocks paid…
Microsoft offers buyouts to 60,000 staff as AI spending bites
Roughly 27 percent of Microsoft's workforce is eligible to leave with a package before year end in the…
Microsoft lets a thousand agents organize themselves on one job
The Agensh harness drops the central orchestrator and lets workers claim their own share of a coding task.
Microsoft’s panel of AI doctors beats GPs on hard cases
A Science paper reports an orchestrator that questions, tests, and debates reached 83% on complex case reports where…
Microsoft turns its security agents inward on the agents it sold
Guardian Agents watch enterprise deployments for insider-style behaviour, built on the premise that frontier models cannot be trusted.
New benchmark says agents pick badly on long tasks
Taste-Bench tests whether an agent chooses the better path when outcomes are hidden, and the strongest model managed…
Microsoft builds a 2T-parameter base model on a cleaner diet
Microsoft's research group says its new MAI-Base-2 base model beats the best open rivals while using far less…
Microsoft bets its own silicon can carry Azure inference
Microsoft unveiled Maia 300 and said its homegrown accelerators will move to multi-gigawatt deployments beginning in 2027.
A new class of model scores options instead of writing text
Microsoft and Nace both shipped small decision-scoring models that rank choices rather than generate prose.
Nadella wants an emergency brake bolted onto frontier models
Microsoft's CEO says advanced AI needs a kill switch that authorized operators can pull mid-task.
Microsoft builds a benchmark that checks what agents actually change
A new Microsoft and Hugging Face benchmark grades agents on the records they leave behind, not the reports…
Microsoft’s streaming speech model transcribes while you talk
Microsoft released MAI-Transcribe-2-Streaming, a real-time speech model it says tops Artificial Analysis accuracy while returning draft text in…