This isn’t a hot take. It’s objectively true

Before Claude Fable 5 went off shelves, I asked it to try one of my most challenging tests: One-shotting a Temple Run inspired voxel biker game in ThreeJS Pre-Fable 5 models across major labs struggle on pulling this off

ちょっと前まで「あー、このクソ面倒な作業、眠っている間に小さいおじさんがやっておいてくれないかなー」って思っていたのに。本物の〝小さいおじさん〟としてClaude CodeやCodexが現れた結果、仕事はラクになるどころか常に複数の判断・承認タスクを投げつけられて眠れなくなってしまった…

Rate limiting on the GLM coding plan is absurd. I am paying $65/month and I can hardly get anything done because I get rate limited so heavily. Fix this @Zai_org

cloudflare is anthropic hater. noted.

YC request for startups!

my read of the fable situation: anthropic decided to take its own customers hostage to improve its negotiating position vs the government similar to how epic games intentionally broke app store rules in 2020, forcing apple to ban fortnite, affecting 100 million ios players

“The Admin asked Dario to fix the jailbreak or de-deploy the model. Dario refused. — In their blog post, Anthropic defended its decision by saying the jailbreak isn’t serious. “ This is crazy. What are we even doing here?

Let’s all take a deep breath, put politics and tribalism aside, and not overreact t @DarioAmodei’s effervescent blog posts What we can all agree on is that: A) these recent models are extremely powerful B) we need to be thoughtful about releasing them* Through that lens the public and private marketplace is doing a solid job of managing the situation, no? [ * shout out to @elonmusk who saw this 15 years ago and created openAI to democratize access to these models — as opposed to one company controlling access to them ]

I'll be the first to say it: Dario needs to be ejected from Anthropic. He and his gang of Effective Altruists have spent years spreading fear, uncertainty, and doubt. The oldest trick in the book: FUD. Now Anthropic is paying the price, and Dario needs to take responsibility for this debacle.

JUST IN: Andrej Karpathy, a top AI scientist at Anthropic, is reportedly barred from accessing the company’s most advanced AI model because he is not a U.S. citizen.

‼️🚨 BREAKING: Amazon researchers snitched to the US government about jailbreaking Fable 5 and Mythos 5, forcing Anthropic to immediately shut down worldwide access. A security export control directive from Commerce Secretary Howard Lutnick enforced the action. Anthropic is fighting the directive and calls it a misunderstanding. This isn't the first clash. The Trump administration had already tried to get Anthropic to pause the release of its latest models before this directive landed.

Anthropic ran around to every media publication for the past 3 months, screaming that they had the cybersec equivalent to a nuclear weapon. They begged for regulation, released their "weapon" to the public, refused to patch exploits, and they're surprised that it got shut down.

there's a lot of asymmetry there. for one there's the eternal problem of the defender needing to get it right every time, while the attacker just has to get it right once. second is the ease of integrating it into the workflow. eventually it would even out more, but if an open weight mythos-tier model dropped tomorrow, integrating it into a defensive stack isn't instant, especially if there's bureaucracy issues or cost issues or whatever. whereas most attackers could reap the benefits more or less immediately.

Alright AI Safetyists, you've been working on AI Safety for years now. Here's a multi-billion dollar incentive to apply your skillset to do something actually productive for once Totally fair reaction of the government in the face of a cyberweapon with loose guardrails