Anti-AI, AI Club — we track the data most coverage skips:
Takes 10 seconds. One question, no email needed.
Anthropic released Claude Fable 5, a model from a new tier they call Mythos-class. It's the strongest thing they've ever put in front of the general public. The launch post is full of capability claims. The parts worth tracking are the access, the price, and the catch.
The honest part is in the fine print. Anthropic says the safeguards are tuned to be cautious, so some harmless requests will get caught and silently routed to Claude Opus 4.8. A near-identical model with the cyber safeguards removed, called Mythos 5, goes to a small set of government and infrastructure partners, not to you. And the free ride on subscription plans is temporary by design. It's "available everywhere today," then rationed in two weeks because they expect demand they can't predict.
None of that is a scandal. It's just the shape of a frontier launch once you read past the benchmark charts: priced to spread fast, gated where it's dangerous, and free only until the capacity math stops working.
Read Anthropic's announcement →Most AI coverage either sells you a dream or sells you dread. We're not doing either. We just track the numbers most people skip: how big the market actually is, what it's doing to jobs, and who's using it for power.
A running mix of articles, research, tweets, and repos. All on AI.
Lots of software now ships with AI features switched on. Companies don't always flag it, settings sometimes get reset after an update, and opting out is often only available to paying customers. We track the specific ones and where the toggle actually is.
Training on your interaction data
Recall, Copilot, Click To Do
AI training data collection across Jira, Confluence, Rovo
Pick what you actually care about. That's the whole ask — takes about 10 seconds.
Got it — thanks for telling us what matters to you. If you left an email, you'll only hear from us when there's something real worth sharing.
Want to help shape it more? Send us an article worth reading: antibsai@gmail.com.
No — the name's half a joke. We're critical of the hype and honest about the harms, but we accept AI is here to stay. Clear eyes, not a crusade.
No. We skip the rage-bait. The aim is calm and informed, not anxious.
Numbers. We lead with data and sources instead of opinions and vibes.
No. The only thing we ask is which topics you care about. Email is optional, and if you leave one we'll only use it when there's something real worth sending.
A regular person tired of hype who started collecting the good, data-backed critical stuff. You're early enough to shape it.