Comment by myaccountonhn
7 hours ago
I think Anthropic just have no legitimacy to being the stewards of AI. They don't have a good track record. They are a private company without any governence that puts my interests into the equation. I am not US-based, thus I can't democratically influence them. Why should I want some batshit crazy, US-based technocrats deciding what I can and can't do with AI?
Not to mention they are the first case of a code repo leak I ever heard.
Reverse engineering, disassembling, cracking - we all heard. I never heard a company leaking the entire codebase of their product.
And, when people read the code, they aren’t even impressed.
Code leaks are pretty common in gaming. But they happen when someone hacks servers or pays off employees, not accidentally like Claude Code.
Cheat developers all want server code so they can analyze the cheat detection and sell more reliable hacks. Modders, pirates and preservation activists can skip a lot of complex RE work. Competing game studios might learn some new techniques. Abused workers want revenge. Script kiddies want clout.
It's a notoriously secretive industry with complex products, hostile work environments, lots of media exposure and heavy competition. There are way more incentives for a leak compared to some ordinary web app or business tool.
Claude Code could be another case where someone working at Anthropic decided to expose them. Or they did in on purpose because they wanted to show there is no secret sauce and Claude is really that good, but they were embarrassed to open source it and make themselves look like hypocrites.
Who would you prefer?
It's not really about who, it's about having institutions with good governance. I don't believe in technocracy/theocracy.
Sure, but that's what I'm asking. What institution, with what governance? Someone ultimately has to decide what these systems will and won't do, and there isn't an obvious democratically legitimate institution with global jurisdiction over AI as far as I'm aware?