Hacker Newsnew | past | comments | ask | show | jobs | submit | dalmo3's commentslogin

I hereby propose the Turquoise Trolley Problem (TTP):

If you say the trolley is blue, it goes straight, where there's a baby in the tracks. If the trolley is green, two grandmas die.


If you say it's neither, the trolley derails ;)

Ah, yes, 2001 but on land.

I consider the scene with Dr. Chandra and SAL 9000 to be a fairly realistic predictive description of how experts interact with LLMs. SAL even has a somewhat obsequious personality.

I assumed the author was one of those HN haters and was filtering by referrer. Now I'm curious.

No, I tried going to the page on a fresh incognito session. Same thing.

At this rate Lee Child must have already secretly invented AGI.

AI or not, I liked this bit:

> Averages lie about the tails, and a person who gets a 15 cm bust error doesn’t care that the mean is 4 cm.

A variation of that sentence should be mandatory in every scientific paper.


That's a poor analogy because practicing an instrument is a single-player game.

Unless you mean you can practice socialisation at the mirror, or that you're willing to practice scales or solfege in public.


Ohh they finally ship to New Zealand. It's a shame they wouldn't last year, and I ended up with a Lenovo :/

The new UI is literally opt-in. Nothing changed for me.

> I'm familiar with the newer carousel CSS/JS stuff but it's kind of clunky

I think you've just answered yourself why.

Browsers are the most used applications on the planet. They have one chance to get an API right ("don't break the web"). They iterate on a new API for years. When they finally launch it, devs call it "clunky".


Until you take a baby to the vaccine clinic, the nurse googles which vaccine to give at his age, and blindly trusts the highlighted AI snippet at the top.

Not a fictitious example.


I saw a doctor use AI to do some maths recently. I checked it and it was right, but trusting LLMs to do statistics is not a good idea.


Sure, but that's an incredible level of incompetence, that I can only see would be expose otherwise outside of AI use. The entire list of what vaccines to give to what age group could fit on a single piece of paper.


And much of what people use AI for now could be easily done without it. How many steps of a Claude Code /plan are just running basic ls commands, all for a few thousand tokens?

The entire thing reeks of laziness and incompetence. It's neat an all but its a giant sucking maw that is threatening to gobble up whats left of anything good.


Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: