Hey buddy,

Man, Hacker News was buzzing today, Friday, October 16, 2026. Had to give you a quick rundown of some wild stuff.

Tesla Roadster Drama

First up, that Tesla Roadster post blew up again. Like, it's been 10 years since they first showed it, and people are still super skeptical. The comments were brutal! Folks were saying it's a "silly idea" like flying cars, and someone even made a weird analogy linking it to "nazis" – wild, right? Lots of doubt about whether it's even real or just marketing hype at this point. Check it out: https://www.tesla.com/roadster

Claude Spills Crypto Secrets?

Then there was this crazy one about Claude (the AI) apparently revealing how NIST generated some crypto curve origins back in the day. The article got a ton of points, and the comments were going off. If Claude's right, it implies NIST staff used really bad random number generation practices, like srand(time(NULL)) for cryptographic seeds, which is a huge no-no. One guy who literally wrote a major crypto library (libsecp256k1) was in the comments laughing at someone for insulting his intelligence. So much drama around potentially sloppy crypto history! Here's the link: https://www.anthropic.com/research/ecdsa-curve-origins

AI Still Sucks at "Real" Work

Another big one was about AI, but not in a good way for the bots. A study claimed that 90% of US working hours are spent on tasks not covered well by AI benchmarks. Basically, AI is great at specific, benchmarked tasks, but most of our actual jobs are messy, human-interaction-heavy stuff it's not good at. The comments had some interesting takes, like how corporate bureaucracy slows things down way more than AI ever could, and a thought-provoking idea about whether "white-collar" desk jobs are even suitable for humans in the first place. Read it here: https://www.mercor.com/research/labor-index/benchmark-map/

Waymo Hits Toronto, Gets Roasted

Waymo is expanding to Toronto, which you'd think would be cool, but the comments were pretty negative. People were ripping on how self-driving cars just improve the experience of navigating "noisy, nasty, ugly, toxic and dangerous car sewers" (aka city streets) and questioning if it just leads to more isolation. The "enshittification" argument even popped up for ride-sharing services. Sounds like folks are tired of car-centric urban planning. Link: https://waymo.com/blog/2026/10/waymo-in-toronto/

AI Finds Too Many Bugs for Humans

This is wild: Postgres 19 got delayed because AI reviewers found more bugs than humans could fix! Like, the AI was *too good* at its job. The comments were debating if LLMs actually find real bugs or just hallucinate them. Some said they've seen AIs do a solid job finding bugs in real codebases, especially if trained on the project's history. But others pointed out that while AI might get the micro-code right, it often introduces "invented complexity" that a human wouldn't. Definitely a sign of things to come, for better or worse. Article: https://www.theregister.com/databases/2026/10/16/postgres-19-delayed-after-ai-reviewers-find-more-bugs-than-humans-can-fix/5303471

Home Assistant Secret Scandal

And finally, a bit of a security scare: Home Assistant users are up in arms because an integration was apparently scanning their secrets (like passwords) against Have I Been Pwned without clear user authorization. The original issue author even claimed they could "abuse this to loot credentials from every HA install on the planet," which is a terrifying thought. It really highlights the trust issues when local, private software starts reaching out to third parties without explicit consent. Read the drama here: https://github.com/home-assistant/home-assistant.io/issues/49314

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