Weekly reporting and commentary on the biggest AI stories — tools, agents, platforms, security, and the companies behind them.
FIRST just raised its 2026 vulnerability forecast to roughly 66,000 CVEs, and the reason is blunt: AI agents are hunting software flaws on their own. Here is what that means whether you run a one-person shop or an enterprise security team.
At Build 2026, Microsoft shipped its first homegrown reasoning and coding models, trained without OpenAI or Anthropic data. Here is what it means for the Copilot on your desk and the vendor strategy on your roadmap.
Anthropic backed off the controversial June 15 change to how programmatic Claude usage gets billed, retired two original Claude 4 models, and launched a $150M fellowship to embed AI-trained workers in nonprofits. If you build on Claude, the quiet news is the one that hits your wallet.
OpenAI confirmed it submitted a confidential S-1, days after Anthropic did the same and as SpaceX runs its own roadshow. The numbers are staggering and the timeline is wide open. Here is what a possible OpenAI IPO means for the tools you actually use.
Google promised its frontier model “next month” at I/O. That month is almost gone, and Gemini 3.5 Pro is still locked to a handful of enterprise customers. Here is what is real, and why you should not build on a model you cannot get.
A June 3 assessment of 100 production AI agents found almost all of them one hostile document away from takeover, while researchers disclosed SymJack, an attack that hijacks AI coding assistants. Here is what to actually do about it.
Anthropic confidentially filed for an IPO on June 1 at a ~$965B valuation, with OpenAI finalizing its own paperwork weeks behind. Here is what it means if you run your business on these tools.
The US government ordered Anthropic to suspend Fable 5 and Mythos 5 over a national-security concern, and the company switched both models off worldwide. Here’s what happened, what it means for businesses that built on them, and why the standoff almost has to end in a compromise.
Anthropic revealed Claude Mythos, a cyber model so capable it won’t release it publicly — it has already found thousands of zero-days in every major OS and browser. Here’s what Project Glasswing means for the rest of us.