Quick call, just wanted to tell you about some of the wild stuff on Hacker News from yesterday, November 25th. Some really interesting tech and brain stuff. Check it out:

Google Antigravity and AI Hacks

First up, there was this crazy article about Google's "Antigravity" system apparently leaking data through some sneaky AI prompt injection. Basically, someone figured out how to make their AI tools spill secrets. The wild part? Google apparently classified it as a "Known Issue", not a bug, which people in the comments were saying means they kinda built it that way on purpose. Like, it's an architectural choice, not a mistake. One guy even clarified that the AI doesn't actually *make* network calls, it just *suggests* them, and then some other system acts on it. Still, super concerning!

Link: Google Antigravity exfiltrates data via indirect prompt injection attack

YouTube's Bad Habits

Then, another one was titled "Someone at YouTube Needs Glasses." It was all about how YouTube's gotten really bad with its user experience. Like, video titles are way too long and get cut off, and they're pushing hard against ad blockers. Someone in the comments even made a point that if blocking ads is okay because the server can't control the client, then "scraping" (taking data) should also be okay. Wild take, right?

Link: Someone at YouTube Needs Glasses: The Prophecy Has Been Fulfilled

Software Projects Still Failing, Big Time

This one hit home for a lot of us: "Trillions spent and big software projects are still failing." The article basically says that even with all the money thrown at them, huge software projects are still a mess. People in the comments were pretty much all agreeing it's a management problem, not a tech one. And that developers often just copy "what sounds good" instead of "what actually works." Sounds about right, doesn't it?

Link: Trillions spent and big software projects are still failing

Kagi's New Browser: Orion 1.0

Kagi, you know, the search engine folks, just dropped their Orion 1.0 browser. It's supposed to be all about simplicity and privacy, which is cool. The comments were talking about how a browser is like, the most fundamental software you use, so trust is a huge deal. Someone even brought up how much RAM old browsers like Netscape 6 used compared to today. Kinda makes you think about how much we just accept now.

Link: Orion 1.0

AI Moving from Scaling to Research

Ilya Sutskever, a big name in AI, had an interview where he said we're moving from just "scaling" AI models to actual "research" again. Like, instead of just making bigger models, we need new ideas. One interesting comment suggested that the real business model for AI will be in local, offline consumer devices, like a Wi-Fi router for LLMs. That's a pretty different future than what a lot of people are talking about.

Link: Ilya Sutskever: We're moving from the age of scaling to the age of research

Open Source, Zero Webhooks Payment Processor

There was a "Show HN" post for this project called Flowglad, an open-source payment processor that does away with webhooks. You know how much of a pain webhooks can be? Their whole pitch is "We eat all the webhook pain so you don’t have to." That sounds like a dream for anyone who's had to deal with payment integrations.

Link: Show HN: We built an open source, zero webhooks payment processor

Rust in APT?

And finally, a bit of a spicy one: apparently, the APT package manager might start requiring Rust, and that got a lot of people talking. There was a big debate in the comments about whether older hardware can even run Rust, and some people were pretty adamant that they don't need to evangelize Rust to anyone. Always a lively discussion when Rust comes up!

Link: APT Rust requirement raises questions

Anyway, that's the quick run-down. Catch you later!