I very much respect the under secretary’s position here. Given that mass domestic surveillance (buying publicly available data, combining it with government data, and building profiles via AI) is lawful and OAI has agreed to it, we should work quickly to change that.

you just know this guy is getting absolutely nothing done

For those wondering how mass domestic surveillance could be consistent with "all lawful use" of AI models, I recommend a declassified report from the ODNI on just how much can be done with commercially available data (CAI): "...to identify ever person who attended a protest"

tbf i wasn't really using it, mostly been using claude and cursor

Claude is #1 in the App Store today — I want to say a huge thank you to all of our new (and existing!) users for the support. We’re working hard for you, please share your thoughts and feedback along the way.

This is a big deal. I think Sam Altman will be doing damage control on this for a while, the impact if not managed can not be underestimated. Pretty bad optics to scoop the deal immediately, they should have waited a week.

Ready to make the switch? claude.com/import-memory

I don’t buy the argument that they couldn’t get their red lines in writing, but we should trust them because they will still enforce them? Also goes against the gov narrative that no company can tell the DoW what to do. Have fun enforcing something you couldn’t be bothered to

Can’t wait for GPT-6.2-robokiller

Man I'm watching Tech YouTube videos from ~10 years ago and they feel like a window to such a radically different world. People used to care about shit.

sucks hey dawg! this is not even close to true! there are literally thousands of *documented illegal abuses* of our mass surveillance system, from looking up agents Hinge dates to spying on a congressman to *just batch looking up info on all 19000 donors to a political campaign*

The LinkedIn/recruiting example is obviously the sort of loophole wedge they will use. I’m sure OpenAI will go along with it, then the creep to more invasive surveillance begins..

just one more post bro. just one more post and we'll sway public opinion. just one more post. bro please just post more. one more post and we got it. that thing in the company slack? post it bro it will be sick. just one more post and the singularity is ours

I don’t think any wordsmithing will change the optics that in pressure situations Anthropic proved themselves to be deeply principled and OpenAI deeply unprincipled. All of this backtracking is only the result of public pressure. Anthropic didn’t need this to do the right thing

Core 3 is here, our latest major SDK release. • Agent-friendly customization APIs • Theme editor and interactive docs • Agent-optimized onboarding for more frameworks • Modern React support • Performance improvements Upgrade today: 𝚗𝚙𝚡 @𝚌𝚕𝚎𝚛𝚔/𝚞𝚙𝚐𝚛𝚊𝚍𝚎

Your UI. Our APIs. Core 3 upgrades our hooks to make building custom UIs easier than ever.

The fight between Anthropic and the DoW is a warning shot. Right now, LLMs are probably not being used in mission critical ways. But within 20 years, 99% of the workforce in the military, the government, and the private sector will be AIs. This includes the soldiers (by which I

We partnered with Stripe on Stripe Projects to bring authentication into the CLI-first workflow. Provision Clerk from your terminal with just: 𝚜𝚝𝚛𝚒𝚙𝚎 𝚙𝚛𝚘𝚓𝚎𝚌𝚝𝚜 𝚊𝚍𝚍 𝚌𝚕𝚎𝚛𝚔

A little short I made about Norm.

Introducing B2B Authentication. Built to help you build for enterprise customers from the start.