There are more than 1.6 million skills you can install into your AI right now. You need about twenty.
That gap is a problem software already solved once, painfully. A copied skill is a dependency that runs inside the work you have stopped checking, and most people install them the way we installed packages before the lockfile existed: no version, no source, no way to switch a bad one off.
The file costs nothing to copy, which is exactly why the file was never the moat. Knowing which twenty to keep is.
Which skill is steering your agent right now that you could not, this second, name the source of?
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: https://t.co/bwn0sximKZ
There are more than 1.6 million skills you can install into your AI right now. You need about twenty.
That gap is a problem software already solved once, painfully. A copied skill is a dependency that runs inside the work you have stopped checking, and most people install them the way we installed packages before the lockfile existed: no version, no source, no way to switch a bad one off.
The file costs nothing to copy, which is exactly why the file was never the moat. Knowing which twenty to keep is.
Which skill is steering your agent right now that you could not, this second, name the source of?
Fast models make mistakes of omission. They skip details they consider unimportant.
Reasoning models make mistakes of commission. They elaborate confidently on wrong premises.
Same task, different failure modes. Fast model output needs a completeness check. Reasoning model output needs a premise check.
The verification strategy changes depending on which failure mode you are defending against.
🚨 $HIMS INSIDER BUY
Director David B. Wells bought 48.4k shares ($1.2 million) today at $24.235 average
Wells was previously CFO at Netflix
He now owns 224k shares after the purchase
THE FIRST OPEN-MARKET INSIDER BUY IN YEARS!!!
Google DeepMind's latest paper shows an AI agent solving 9 of 353 open Erdos problems, at about $200 per problem. The agent generates formal proofs in Lean and verifies them automatically. No human oversight needed. A 2.5% solve rate on problems mathematicians have been stuck on for decades, at a cost that keeps dropping. This is Law II in action: AI's real power is doing what humans cannot, not just doing what humans do faster. The paper is worth a read: https://t.co/1mj4gRd2kO