Leland Richardson

And so it starts

@adhidesign i'm open to it! i'm not sure I will have any revolutionary insights to provide though :)

@intelligibabble Makes sense! On a completely different topic: I'm very curious how your workflow has changed at Anthropic! If you find some time, you'd be open to giving a talk sometime to your old team on it?! 😁

Good advice for coming months

we sent the right crew to the moon 🥹

any and all software. The goal is to harden as much as possible before a model of similar capabilities is openly available to bad actors. Which is just a matter of time. Mythos-level LLMs are extremely capable at finding and creating the exploits, whether the software has anything to do with AI or not. For example, an OpenBSD exploit was found, taking advantage of a 27 year old bug in arguably the most security hardened OS in the world for a few $1k of tokens.

@intelligibabble is this for any and all software? or just the ones exposed to LLMs?

wasn't meant to be a vaguepost! this is in reference to mythos/glasswing... Mythos is just the *first* model with this level of cyber capabilities. If glasswing works, there will be thousands of vulnerabilities and hundreds of exploits found in the most important software that exists over the next couple of months. patches will follow but take time. While we undergo this transition, reducing exposure will require that people upgrade software if it can't be done remotely by the vendor.

@intelligibabble tell us more!!

when you spam "hi" to Mythos over and over and over instead of getting frustrated and angry sometimes it decides to write an epic quest involving 11 emoji animals defeating Bye-ron the Ungreeter

Some brief thoughts on Mythos We’ve known this was coming for a long time. At least, we *should* have. Extremely effective software vulnerability discovery was clearly coming to anybody paying attention. It has also been clear that all AI policy so far has been made and executed with training wheels. It was always clear that, sometime soon, the training wheels would come off. The training wheels aren’t fully off just yet—this model is being kept under lock and key, and Anthropic does not seem inclined to release Mythos preview to the public anytime soon, if ever. The training wheels will be off when these capabilities are fully diffused in ways centralized actors cannot control. It is inevitable that this will happen. The point is not to argue about whether we should “ban open source” or similarly unrealistic notions. The point is to harden the world for this new reality. I applaud Anthropic—and I especially applaud @logangraham—for doing so. But their efforts alone are not close to enough. Project Glasswing—a partnership with Anthropic and other companies—seems nice, but unsurprisingly it lacks uniform frontier lab participation. It would probably be ideal, for our national cyberdefense, if the federal government were not trying to destroy Anthropic and eliminate their models from government systems. If anything, the government should be trying to work more closely with Anthropic. As a side note, I hope Anthropic is working with state and local government entities on cyber vulnerability discovery, since many of our adversaries know that state and local is America’s soft underbelly in so many ways. In any event, the Mythos news should lay bare how stupid and counter-productive the Department of War’s feud with Anthropic really is. As someone who suspected all this was coming (not from inside knowledge but from it being ~obvious), that probably explains why I have had such a strong reaction to that feud. It’s this senseless distraction just at the time that the training wheels are coming off. I hope the two parties can resolve their differences now, for the sake of the country, but I am not hopeful. I do want to call out, however, the numerous political and career civil servants in the Trump Admin who do get these issues, know how stupid the Ant-DoW stuff is, and want to work with the frontier labs like adults. I wish you all utmost success. I find myself inclined to end on some positive notes. Mythos appears to be—according to Anthropic at least—“the most aligned” model Anthropic has ever trained. We are approaching superhuman capabilities in some domains, and yet alignment is getting better rather than worse. That’s not nothing. I know some of you think the model is faking its alignment, or aware when its alignment is being tested. I don’t have a good answer. Finally, there is this: Mythos was made by an American company, and like most successful American companies, it has a vested interest in maintaining order and peace, and it is investing substantial resources in mitigating the risks of its technological progress, as I expect most of the American labs would. This is cause for optimism: The incentives of capitalism are working. The training wheels are coming off, but at least we are the ones removing them, as opposed to our enemies. Perhaps we can be the first to learn to bike for real. The first step would be to get beyond all the low-fidelity, under-specified, pimply little fights of AI policy’s prepubescent era. That goes for me too. “What hath God wrought,” wrote the first telegram. What, indeed. In this case, the answer is still up to us.

Love this idea

I love people using auto research for non ml stuff like this. I have a few things like this I want to try! Just need a good eval function and go

Our run-rate revenue has surpassed $30 billion, up from $9 billion at the end of 2025, as demand for Claude continues to accelerate. This partnership gives us the compute to keep pace. Read more: anthropic.com/news/google-br…

How Axios was compromised 🤯

@GroundControl @bcherny From your description, it should be allowed. See here

I built myself a web UI on top of tmux so I could have better session management for my claude code sessions. My usage pattern of juggling multiple tmux/claude code sessions at the same time hit my weekly usage limits really hard because of constant cache miss (1M token limit, which I didn't set myself, combined with short server cache time) I have no idea what Anthropic considers a harness. Is my tmux wrapper a harness? I've asked, because it's been really confusing, and I've been told that Anthropic doesn't want to stifle innovation. But this week really shook my confidence tbh. It would be really nice if Claude code had a server mode like Codex so we could build our own convenience UI on top of it with our subscriptions.

Starting tomorrow at 12pm PT, Claude subscriptions will no longer cover usage on third-party tools like OpenClaw. You can still use these tools with your Claude login via extra usage bundles (now available at a discount), or with a Claude API key.

We see our home planet as a whole, lit up in spectacular blues and browns. A green aurora even lights up the atmosphere. That's us, together, watching as our astronauts make their journey to the Moon.

They want LeBron

Basically my sense is that we cannot possibly make it through the AI transition well if the public debates reduces to whether it is “racist” or “woke,” or whatever other low-fi garbage the culture war prism would yield. It will be impossible to muster the requisite intellect.

Wow, I totally disagree with this statement. At the current state, AI actually amplifies the developer to developer difference. If you were a 10x developer, you had good ideas + architectural clarity, this is a brutal advantage when using AI. Steering is a fundamental part of today's AI development.

In 2008, 62% of teachers said they were very satisfied with their job. In 2022, that dropped to 12%. We've got a serious problem brewing in education...

I press “Show me fewer shorts” on YouTube like… daily. I don’t think it has ever made any fewer show up. I mean can I just pay 2x whatever ad revenue you think you would get by me watching more shorts for a toggle to just show me none?

Literally got tears