Man, Hacker News today was pretty wild, especially with all the AI stuff and some crazy real-world drama. Lemme give you the quick rundown:

Claude Opus 4.8 Dropped

First up, Anthropic just launched their new AI model, Claude Opus 4.8. People are going nuts for it, especially folks who used to use Codex for coding. Sounds like it's making a lot of coders super happy. Some people in the comments were debating if LLMs are hitting a wall with energy costs and memory, which is kinda interesting. And one guy even wondered if this new model is just a way for Anthropic to get more revenue by making people use more tokens, which is a pretty cynical but maybe smart take.

"Can We Have the Day Off?"

There was a big discussion about the idea of giving everyone more time off, or even some form of universal basic income. It got a lot of people talking about how the US has high taxes but not great social services. One person said that with all the recent Big Tech layoffs, it's never been easier to start your own company, especially if you slap "AI for X" on it. Kinda makes you think, right? Someone even compared it to the old promise of nuclear power being "too cheap to meter."

Bricks and Minifigs Stole a Man's $200k Lego Collection

Dude, this story is insane. Apparently, this company, Bricks and Minifigs, just straight-up stole a guy's $200,000 Lego collection. The comments were all about how blatant the executives are, basically daring the guy to sue them and saying they'll drag it out. People are shocked at how wild this whole situation is, especially since the company kept selling the Legos even after a cease and desist. Total corporate scumbags, if you ask me.

UC Faculty Wants SATs Back for STEM

This is a hot topic: UC professors are pushing to bring back SAT tests for STEM admissions because they're seeing "severe math deficits" in students. The comments section blew up with debates about equity vs. equality, and how these tests can be designed to favor certain environments. It's a real mess, with people arguing if it's about building a successful society or just getting good at tests.

AI Models Disagree on Facts

Turns out, even the top AI models often can't agree on basic real-world facts when you give them fact-checking tasks. It highlights that it's actually super hard to write good questions for AI, and that these LLMs are basically just guessing the next word, so they don't really "know" what they know or don't know. Pretty humbling for the AI hype, honestly.

Dorm Room Million-Dollar Product

Okay, this is pretty cool: a guy wrote about how he made a million-dollar product from his dorm room back in 2025. It was a component for custom keyboards called Nice!Nano. The comments had some interesting points, like how important it is to find a niche with enough motivated buyers, and how being in a supportive environment really helps. Someone also asked if he regretted open-sourcing the firmware, since it led to clones, which is a classic dilemma for makers.

GitHub Banned a Security Researcher

And finally, GitHub banned a security researcher who posted some zero-day Windows exploits. People were debating if it was fair, with some bringing up Section 230 laws and others pointing out that Microsoft has been laying off skilled security staff. It seems like a pretty vindictive move from Microsoft, according to some of the experts commenting.

So yeah, a lot of AI talk, some crazy corporate drama, and a cool success story. What a day!