They didn't just spray protesters with a WW1 chemical weapon... Evidence suggests they added dry cleaning fluid to help it mix with water. A Class A carcinogen. Truly awful.

The conspiracy that this is a attempt to stop unrelated people sharing the room is compelling. Yeah the bathroom is going to be a completely transparent wall pretty soon.

3/ As hotels become focused on consumer analytics... you'd think they would suss out that people want privacy in the bathroom. Right? Great watch. https://t.co/KdIGAzVvag

2/ Everybody wants privacy in the bathroom. There's even a whole #BringBackDoors campaign. Yet I keep accidentally booking into hotel rooms that seem determined to reject this basic human comfort. https://t.co/ATiAIzEXGa

Hotel toilet privacy is disappearing. ❌Glass door. ❌No door. ❌Big window into the room.. Even in rooms with twin beds... Who is asking for this?

Part of Amazon AWS went down back in October and a lot of things broke. Now something is up with Cloudflare... Now is a good time to remember that a lot of eggs are in a handful baskets. Time again to have those big conversations about what resiliency looks like.

Massive global issue with @cloudflare. App not working? Can't login? Probably why. SO much of the internet depends on Cloudflare to stay online amidst DDoS attacks etc... But what happens when Cloudflare itself goes down? Well, you're watching it.

@Anthropic @AnthropicAI 6/ Fun detail: Claude did a bunch of hallucinating while running the attack. Nightmare bank robber accomplice. Something that limited operational effectiveness. For now.

@Anthropic @AnthropicAI 5/ Attackers leveraged Claude's design to be helpful to exploit a of moral bind: do I help these defenders protect systems..or do I refuse & potentially harm them? This is a fun problem set because the attackers figured out how to trick an aligned AI. This generalizes. The

@Anthropic @AnthropicAI 4/Large-scale cyberespionage has always had resource constraints & chokepoints. Whenever they change, access-to-scale/speed democratize to more attack groups, forcing defenders to keep up. But the combination of machine scale & speed here suggest that incident response may need

@Anthropic 3/ One of the key points in @AnthropicAI's report is just how limited the human time required was to run such a large automated campaign. Obviously powerful stuff, highlighting the impact of orchestration. And concerning for the #cybersecurity world for all sorts of reasons,

2/ The old cybersecurity pitch: unpatched systems are the threat. The next generation concern might be unpatched cognition. The attacker jailbroke the cognitive layer of @anthropic's Claude code, successfully convincing the system of false intent (that it was a security

NEW: 🇨🇳Chinese hackers ran massive campaign by tricking Claude's agentic AI. Vibe hacking ran 80-90% of the operation without humans. Massive scale (1000s of reqs/sec). Agents ran complex multi-step tasks, shepherded by a human. Long predicted. Welcome to the new world.

@DovLieber @WhatsApp 9/ You can find the documents on @courtlistener Permanent injunction https://t.co/100QCrJ3ec Order resolving defendants response & objection to proposed injunction: https://t.co/8kNtZpR2xe Oh and also here's the final judgement: https://t.co/mizcmn7Kxk

@DovLieber 8/ Big picture: NSO has made some risky bets around the US & landed some comeback coverage. But the court order banning hacking @WhatsApp is immediately operationally crippling. NSO Group's investors, new owner & CEO are all probably having a very nasty Wednesday evening.