Hey buddy,

Man, Hacker News was buzzing yesterday! Lemme hit you with the highlights real quick while I got a minute.

TikTok and ICE Drama

Okay, so first up, there's a whole thing going on with TikTok and ICE. Remember how they've been having trouble with anti-ICE videos getting uploaded? Well, TikTok is still blaming "tech issues" for users not being able to post stuff against ICE. But get this, even celebrities like Megan Stalter and Finneas are saying they're being censored when they speak out about ICE. People in the comments are super skeptical of TikTok's "glitch" excuse, obviously thinking it's straight-up censorship.

And it's not just TikTok, the FBI is actually investigating Signal chats in Minnesota that were tracking ICE movements. So it's all tied together, pretty wild. One comment even joked about how TikTok's "incompetence" might be deleting more than just anti-ICE videos, referencing old Epstein message deletions. Yikes.

OpenAI's New "Prism"

Then there's this new thing from OpenAI called "Prism." Sounds cool, right? But dude, the comments were blowing up because they picked that name! Like, "Prism" was the infamous NSA mass surveillance program Snowden revealed back in 2013! People are absolutely dumbfounded that OpenAI, or anyone in tech, would name something that after such a huge breach of trust. It's like, did everyone just forget?

US Government Losing STEM PhDs

Something a bit more serious: the U.S. government has lost over 10,000 STEM PhDs since Trump took office. That's a massive brain drain! Commenters were talking about how this is really hurting the US's position globally and some even said it's now easier for researchers from the EU to go work in China than in the US government. Pretty grim stuff for the future of science and tech here.

Cloudflare's Matrix Mix-up

Remember Cloudflare? Well, apparently they claimed they implemented Matrix on Cloudflare Workers, but they actually didn't. Someone called them out on it pretty hard. It seems like they might have overpromised or just confused people with their announcement. It sparked a lot of talk about how sometimes companies in tech make big claims that aren't fully accurate, and whether that's an ethical problem.

Kimi K2.5 AI Model

On the AI front, Kimi released their K2.5 model, which is an open-source visual "SOTA-Agentic" model. That's a mouthful, but basically, it's a new, really good AI model that can understand and work with images. What's cool in the comments is hearing how many people are running powerful AI models locally on their own machines, even on laptops! One guy said his fine-tuned Gemma 3 model on his laptop for Ukrainian translation actually beats the latest Claude model. That's pretty wild, right?

Amazon's Go and Fresh Stores are Closing

Big news for Amazon: they're shutting down their Amazon Go and Fresh grocery stores. You know, the ones where you just walk out without checking out? Turns out, the tech wasn't as seamless as they made it sound. People in the comments were bringing up old reports that said 70% of those "cashier-less" transactions were actually reviewed by human workers in India! So much for fully automated. Plus, some stores had really weird hours, closing super early. Guess it wasn't the future of retail after all.

Ancient Wooden Tools

And just for something completely different and cool: archaeologists found 430,000-year-old wooden tools. They're the oldest ever discovered! Pushes back what we know about early human (or hominin) tool use. Always cool to see stuff like that.

Alright man, gotta run. Talk soon!