DeepSeek reportedly raises $15B days after US restrictions

DeepSeek has closed a funding round of roughly 100 billion yuan — about $15 billion — backed by existing and new investors including Tencent and CATL, in one of the largest financings in Chinese history. The raise lands days after fresh US restrictions and reportedly positions the Hangzhou lab for an IPO as it competes head-on with OpenAI and Anthropic, sharpening the geopolitical split over open-weight models. Source: TechCrunch · Reuters

Anthropic launches Claude Duet, its bet on proactive agents

Anthropic has unveiled Claude Duet, a new app where Claude acts as a proactive agent — managing work tasks, building web apps and checking connected apps on the user's behalf. A research preview opens by waitlist today for Pro and Max subscribers in the US, as the market shifts from assistants to agents. Source: Anthropic · TechCrunch

Canva releases its first image generation model

Canva has shipped its first in-house image generation model, baking native text-to-image into the design platform used by hundreds of millions of people. The move brings Canva into direct competition with the established image-model vendors rather than relying on partners. Source: TechCrunch

Microsoft chases 'medical superintelligence' as Suleyman claims AI beats doctors on hard cases

Microsoft AI CEO Mustafa Suleyman says the company's models now outperform doctors on hard medical cases, outlining a push toward what he calls 'medical superintelligence'. The claims arrive as big-tech AI teams compete to land clinical deployments. Source: TechCrunch

ConnectWise fallout widens: Chinese hackers spied on thousands of companies

The ConnectWise Cloud Backup supply-chain compromise is broader than first thought, with Chinese state-aligned hackers spied on thousands of companies after breaching the IT giant. Microsoft attributes the campaign to Storm-0940, which used an agentic AI framework to abuse application credentials across Microsoft 365 tenants at scale. Source: TechCrunch · Microsoft Security

Uber puts all 30,000 employees on ChatGPT Work

Uber is rolling out ChatGPT Work to its entire global workforce of roughly 30,000 employees, making it one of the largest single enterprise deployments of OpenAI's business tier to date. The move signals how deeply consumer AI brands are pushing into frontline corporate use. Source: TechCrunch

🏠 Local models: flint-GGUF and OrcaCyber-Nano

Unsloth has published flint-GGUF, quantized builds of Recursive's flint coding model that run on consumer laptops — 7,400+ downloads in three days. Also new: OrcaCyber-Nano-1.0, a tiny GGUF vision-language model with 21,700+ downloads. Both are runnable on home hardware without a datacentre. flint-GGUF · OrcaCyber-Nano-1.0