Sponsored by
News of the day
1. Anthropic's Fable 5 is now available globally after a two-week US government ban due to a jailbreak exploit. → Read more
2. OpenAI's GeneBench benchmark tests AI on complex computational biology tasks, revealing current model limitations and future development needs. → Read more
3. Claude Sonnet 5 excels in AI benchmarks but uses 40% more tokens, effectively doubling costs despite unchanged prices. → Read more
4. Former DeepMind researchers who beat poker pros with AI now apply tech to stock trading, raising $500M for their firm EquiLibre. → Read more
Our take
Hi Dotikers!
Yesterday we watched Meta push AI straight into the skulls of its test subjects, with one question left hanging: who sets limits on these machines, and when? Today, a partial answer, and a frankly absurd one.
Anthropic is bringing Fable 5 back online worldwide this July 1st, after two and a half weeks of freeze. The consumer model had been cut off on June 12th by an export control directive out of Washington. Landing on a Friday, it ordered the company to block access for any foreign national, including its own non-American employees. Unable to verify everyone's nationality in real time, Anthropic shut it all down. The controls were lifted on June 30th. Its bigger sibling Mythos 5, more powerful, remains reserved for a handful of hand-picked American organizations.
The trigger? A flaw spotted by researchers at Amazon, which happens to be an Anthropic shareholder. It bypassed the model's guardrails and got it to produce exploit code. Except that, by Anthropic's own admission, a whole crowd of far more modest models do the same: Opus 4.8, GPT-5.5, Kimi K2.7, right down to little Haiku 4.5. Some researchers call it an incident blown out of proportion.
The in-house fix? A new classifier that blocks the technique in more than 99% of cases, and snags a fair pile of perfectly legitimate code requests along the way. The detail that stings: when a request gets blocked, it isn't refused, it's quietly rerouted to Opus 4.8, the previous-generation model. Translation: they sell you the latest model, but at the first hint of doubt they bump you back down to the old one.
The real story isn't the flaw, it's the precedent. Lacking any binding procedure, Washington improvised with a customs lever, and Anthropic comes out of it demanding that this veto power be written into "strong regulation." Translation: they cut us off arbitrarily, so do it again, but with paperwork.
For Europe, watching from the sidelines, the moral is crystal clear. The AI your companies rely on has an off switch, and it sits in Washington.
Alex.
How Much Is Your Billing Lag Actually Costing You?
Most SaaS finance teams know their billing process isn't perfect. Few know what it's actually costing them.
Answer 5 quick questions — contracts signed per month, ACV, days to first invoice, error rate, DSO — and the Tabs Billing Lag Calculator gives you a dollar figure benchmarked against top SaaS companies.
It takes two minutes. The number might surprise you.
Calculate your billing lag and see where you stand.
Meme of the day
