🤖 Weekly AI Roundup
Your weekly roundup of the top stories in the fast moving world of Artificial Intelligence
Hello,
Here’s your weekly roundup of the top stories shaping the fast moving world of Artificial Intelligence. Top ten news item, summarized and linked for you to stay on top of things.
I share this with my followers as added value to the conversation of what makes and keeps us humans. The frontier of humanity is moving fast and what was uniquely human last decade is no longer our exclusive domain. So what is and what will remain Humanly Possible?
If you enjoy this and the rest of my content consider following, subscribing and sharing this email with your network.
1- Stripe, Anthropic, the OpenAI Foundation, and Bill Gates formed Intercept, a $500M nonprofit to prevent respiratory viruses like colds and flu.
2- OpenAI and Broadcom unveiled Jalapeño, OpenAI’s first inference chip, built in 9 months with industry-leading performance per watt.
3- The Trump administration urged OpenAI to limit GPT-5.6’s initial release to government-approved partners over national security concerns.
4- Anthropic launched Claude Tag for Slack, a persistent AI teammate that builds context over time. Claude now writes 65% of Anthropic’s product code.
5- Sakana AI launched Fugu, a multi-agent system that routes to the best models through one API and dynamically swaps vendors to bypass export-control risks.
6- Z.ai released GLM-5.2, an MIT-licensed open-source model with 1M context that topped long-horizon coding benchmarks, including FrontierSWE (74.4%), surpassing GPT-5.5.
7- Alibaba Group‘s released AgentWorld, an open-source language world model spanning 7 agent domains that outperformed GPT-5.4 on AgentWorldBench.
8- Meta launched Meta Glasses from $299 with EssilorLuxottica, featuring Muse Spark AI for navigation, live translation, and 26 styles.
9- SpaceX signed a $6.3B compute deal with Reflection AI, leasing $150M/month in Nvidia GB300 chips from its Colossus supercomputers starting July 2026.
10- Google DeepMind invested $75M in A24 to co-develop AI filmmaking tools without requiring access to the studio’s film library.
11- Anthropic accused Alibaba of the largest known AI model theft, citing 28.8M Claude chats from ~25K fraudulent accounts targeting agentic coding.
12- OpenAI launched GPT-5.5-Cyber and Patch the Planet to scan 30M commits across 30K+ open-source codebases and automatically fix security flaws.
13- Nobel Prize-winning AlphaFold co-creator John Jumper left Google DeepMind for Anthropic, following Gemini co-lead Noam Shazeer’s move to OpenAI.
14- Nous Research added Blank Slate to Hermes Agent, letting developers build AI agents from zero default tools for secure, reproducible workflows.
15- SpaceX acquired Cursor for $60B in stock days after its IPO and unveiled Composer 3, a 1.5T-parameter model trained on Colossus with no third-party base model.


