Yeah, I had a good laugh at this. Half of the commits I review are coauthored by Claude, a fact that I’m sure Anthropic is thrilled to claim, but this colossal fuck up was obviously the work of a rogue intern or something.

I wonder if there’s going to be a point in the future where we all look back at this massive over-investment and kick ourselves for making so much expensive electronics waste.

It’s the missing GitHub status page.

GitHub is teetering on one 9 over an entire quarter. There are serious, systemic problems in Microsoft’s shop.

How have I never heard about Kanopy? This looks so cool!

https://archive.is/20260326121207/https://www.wired.com/story/senators-demand-to-know-how-much-energy-data-centers-use/

Makes sense to track this since it’s one of the fastest growing consumers. I think it would also be interesting to track who is buying power infrastructure and in what quantity. Even basic power poles are backordered for years, so I shudder to think about the economic impact of a data center for a blank check tech company clawing in everything they can.

For top left I’d like to humbly nominate SCO Unix.

Like anyone who’s been paying the least bit of attention, I was wondering “wait, what’s new here?”

In a letter sent Tuesday to Attorney General Pam Bondi, Raskin said the documents point to a broader risk to national security, writing: “These new disclosures suggest that Donald Trump stole documents so sensitive that only six people in the entire U.S. government had access to them, that the documents President Trump stole pertained to his business interests.”

This is more interesting than the headline, in my opinion. The claim that they “pertained to his business interests” is not exactly the same thing as selling them to the highest bidder. It suggests either that Trump’s businesses are the subject of intense scrutiny by the fed or, more likely and more worrying, Trump is much deeper into brokering sensitive information than we expected (as in, it’s one of his main sources of income).

See, this is why openai can never go public: if it has to open its books, the music stops.

Disney does not have the luxury of teasing its investors with billion dollar deals-- every quarter, they have to report where they’re stacking their chips, and any cute shit will get them sued and investigated at a minimum. They probably expect their peers to meet them halfway on that.

Openai is not a serious company. When a major prospective partner like fucking Disney wants to open a billion dollar account, what possible excuses could a real company have for not figuring it out?

I’ll be honest, after we pivoted to Afghanistan from Iraq, I assumed I’d never see American boots in Iran. It is inspiring to know that someone has the hubris to go where no one before him dared.

I assume you mean radio frequencies, and the answer is basically none. A grounded fireproof safe is basically a perfect faraday cage.

EDIT: Ok, I actually have a pedantic answer for this. If you put a microphone on a device inside the safe, you can signal it from outside by sending it vibrations, and you could encode a message in binary and thus technically send it a “digital signal”. If you wanted to be a little more analog you could use Morse code :)

Probably yeah. A fireproof safe will be airtight, so the system can only produce as much energy as whatever reagents you put inside.

Where is “nothing is real, but good luck making use of that information”?

It also says “never”, but yeah, it’s a pretty bland statement when you put the context back in (literally and figuratively).

Holy hell, the fact that those slack messages and that chatbot history ended up in court is mind blowing. I guess we should be grateful that this time, the bad guy and his hamfisted “Project X” got put in the spotlight.

Yep, you and I are operating in orthogonal spaces. I genuinely envy you.

Haha, yeah I use it as well, and like I said it makes drafting the code a lot faster, but it dramatically slows down review and validation of fit for the business purpose.

If I could, I’d put the genie back in the bottle because having ICs dump thousand line MRs on each other and then finding out in gamma that it didn’t actually solve the problem is a ton worse than making a person actually think about what they’re gonna commit for a couple hours. But alas, if we don’t take a first draft with Claude or Gemini agentic tools for every ticket we’ll get PIP’d, so I guess the AI enthusiasts and their sponsors are happy.

I think that newer models of Claude are a lot better, but they are still just chatbots and they still just generate words. As anyone in the industry will tell you: typing out the code was never the slow part.

Attorney-Client Privilege. Sorry, I should have just said it.

For anyone who might have avoided this part of the world, ACP makes communications between you and your counsel inadmissible in court. In big companies, it’s somewhat common to bring lawyers into discussions under the auspices of seeking legal advice, but primarily to ensure that if any artifact from that discussion were to be uncovered by an adversary, it couldn’t be used in a lawsuit.

That’s an impressive investigation.

It would be tough to find a better example of why lobbying in the US is fundamentally broken. An entity like Meta has ample funding to break up an operation into distinct cells that do not directly interact in public forums, while tracking the whole process in documents protected by ACP. I think it’s particularly telling that Meta lobbyists are quietly nodding along legislation pushed by “grass roots” activists and that Meta’s new OS just happens to implement the technology exactly as described in the law.

It’s that sort of coordinated effort that the RICO act was drafted specifically to address, but it’s perfectly legal.