Newsletter
Five minutes. What happened, and what to do about it.
OpenAI opens its agent harness to everyone, at no extra fee
The managed harness behind Codex, now behind one API call.


Redis's LangCache stores AI answers so you stop paying for the same one twice

OpenAI stopped selling Pro subscriptions because too many people wanted them

Microsoft finds Irish workers rethinking career routes as AI changes the job

Microsoft is using AI to make leaving Salesforce easier
Anthropic says Chinese labs are systematically copying its models
Anthropic alleges China-based labs, naming Alibaba, Moonshot AI and DeepSeek, ran five campaigns totalling nearly 200 million exchanges to copy Claude's reasoning into their own models. Allegation, not finding. For a small business it is mostly a spectator sport, unless the cheap open model you run later turns out to have a licence fight attached.
- Nothing to do today. Useful if you run open-weight models and want to understand where they came from.
- If you run an open model, keep a note of who published it and under what licence.
Source: TechCrunch. This is our short summary of their reporting. The commentary in it and the things to do are SeedFoundry's opinion. Picture from the source, unless it is our own graphic. How we source and credit

The EU's cybersecurity agency now has Anthropic's Mythos 5 and is testing it
These are short summaries of other people's reporting, with our own commentary. Each story names its source and links to the original. How we source and credit