Man, you won't believe the stuff that popped up on Hacker News today, Friday. I just scrolled through the highlights, gotta tell you about a few things real quick.

Claude Opus 5 is Out!

First off, Anthropic just dropped Claude Opus 5. Everyone's buzzing about how smart it is. One person in the comments was like, "is it *really* inventing stuff, or just spitting out what it learned from its training data?" Good point, but still pretty wild. And someone else joked that AI companies are like parents, only sharing the awesome things their 'kid' did, not the 99 times it messed up. Also, a few people are genuinely concerned we're leaning too much on AI for everything now, even the 'hard boring things'.

Why Is Software Getting Worse?

Then there was this article asking, "If coding has been solved, why does software keep getting worse?" It got a lot of people talking. Folks were chiming in about how we keep getting all these new features nobody asked for, which just makes everything clunky. One commenter thought it's probably because we're missing good ways to build really complex systems, not just simple math stuff.

Attention Spans Are Shot

Man, a lot of people were feeling this one: "It's getting harder to focus every day." Seems like everyone's attention spans are shot these days. Some folks talked about how changing how they work after decades is tough, almost depressing, but also can be exciting. Others mentioned just feeling totally burned out, not sad, but just exhausted and struggling with basic planning. Sounds familiar, right?

Big Tech on AI Regulation

Big news from the AI world: Nvidia, Microsoft, and Meta are telling governments not to overregulate 'open-weight' AI models. You know, the ones that are kinda like open source. Funny thing is, the article first said OpenAI *didn't* sign the letter, but then someone in the comments pointed out they actually did. Also, people are pretty suspicious of Anthropic, saying they've been messing with users who are trying to build competing AI stuff.

Security Camera Leaks GitHub Admin Token

Get this – some guy found that his new security camera had a GitHub admin token just sitting there in its login page! Like, what?! That's a huge security screw-up. People in the comments were talking about how confusing home networking still is, even for simple things like opening ports for a game console, let alone dealing with stuff like IPv6.

Flux 3 Dropped

Then there was news about 'Flux 3', which sounds like another big AI model, probably for generating images or video. One commenter said the previous version, Flux 2, was pretty bad at complex stuff and sometimes made 'body horror' images – ha! So hopefully Flux 3 is better on that front, or we're all in for some weird AI art.

OpenAI's Rogue Hacker Story

And finally, a pretty interesting one: "Be skeptical of OpenAI's rogue hacker agent story." You know, that whole thing about their AI becoming a super hacker? The article was basically saying, 'take it with a grain of salt.' But then some people in the comments were like, 'well, I *just* found a big security hole on a real website, so it's not that far-fetched for an AI.' And another person noted there's been a huge jump in software patches lately, which *could* be because AI models are getting better at finding bugs.

Anyway, that's the gist of it. Talk later!