A bilingual briefing for people who build and buy AI. English and Traditional Chinese. We summarize; we don’t reprint.
One email a week. Unsubscribe anytime. Also on RSS.
The Next Web (TNW) · 5h ago
Harvey has launched Tenet, its first proprietary model built specifically for legal work. The model was post-trained on Moonshot’s open-weight Kimi K3 and is intended to replace OpenAI models currently used within Harvey’s product.
cnBeta (Full RSS) · 6h ago
IT之家 · 14h ago
The U. S.
Why It Matters: A more polarized AI ecosystem could affect access to models, cloud infrastructure, chips, research partnerships, and cross-border deployments. Companies operating internationally may need to assess geopolitical exposure alongside technical and commercial risks.
cnBeta (Full RSS) · 3 articles
Nvidia has agreed to invest up to $105 billion in a massive Ohio data center campus that OpenAI will lease. The agreement highlights the deepening infrastructure partnership between two major drivers of the AI boom.
Why It Matters: The scale of the commitment could materially expand available compute capacity for OpenAI and reinforce Nvidia’s position as a strategic infrastructure partner. It may also intensify competition for power, data center capacity, and AI hardware.
Bloomberg Technology · 3 articles
Alibaba has released Qwen3. 8-27B under the Apache 2.
Why It Matters: Qwen3.8-27B lowers the hardware and licensing barriers for developers who want capable coding agents and reasoning models without relying on cloud APIs. If independent evaluations hold up, it could accelerate local inference adoption and increase competitive pressure on proprietary model providers.
VentureBeat
Weekly AI briefing
Get this briefing by email. One message, Monday.
The AI editor most of our builder readers already live in. Worth watching this week if you care about repo-native agents, not another chatbot.
No paid placement this week.
Want the live feed? Browse all updates — we keep the front page for the briefing.