Why did this fall off the first page so hard? Regardless of the quality of the post, it was getting good engagement I thought.

At least it isn’t flagged? Negative takes on AI do poorly here

____ does good/poorly on HN is often repeated but usually never true. I see constant negativity on AI, and in fact I see it overwhelmingly more negative than positive takes. but i also have my own bias in what i see, and probably falling to the same phenomenon as you

The rate at which AI is capable of producing code is intractable for humans to deal with. Right now the bottleneck is human reviewers. If AI ever becomes effective at generating provably correct code, it's joever.

The good news is there are several thousand newly unemployed software engineers on the market…

It's also likely that there are thousands who aren't hustle-culture oriented who have been looking for so long that they don't even count as unemployed anymore.

Or maybe I'm just projecting.

But will humans understand the proofs?

Oh…

The proofs are not meant for human consumption. It's for the AI to know to try again rather than spit out hallucinations. Of course there's a leap of faith somewhere here.

The proof either passes the SAT solver in a reasonable amount of time, or it doesn't.

Somebody will have to write proof verifier, and that in many ways will be harder than writing some CRUD app that they want proof verifier to validate.

We might even end up increasing the demand and pay for devs if this happens to pass.

We have many of those that are perfectly fine. Writing proofs is still quite hard, especially proofs that actually say something about your program.

Proving something is correct is a far harder exercise, than writing a broken but acceptable version of that thing.

Telling people in tech to touch grass is not the best recipe for upvotes it appears.

What would scare me is if this becomes economically viable enough to release to the public, rather than staying an unlimited budget type of demo.

Your current self won't like it. Your latter self won't care. It's like trying to shame a heroin addict for cheating the dopamine system. When they inject that needle, they are exactly where they want to be.

Sure, "when they inject", but the rest of the time? Generally no, they are not. This is why people DO try, if unsuccessfully, to get off of their addictions -- you don't like waking up in an alley, you don't like how broke you are, you don't like how you feel when you come down from the high, you don't like how you're hurting your family.

Yes, but will AI give you a chance to get sober and reflect? Maybe once you hit your daily usage limits :)

What happens when usage limits go away?

Way back in the day there were usage limits on everyone's internet service. Nowadays, not so much. Inference won't be expensive forever.

I mean, just Devil's Advocate, but I could see this becoming an addiction crisis like none we've ever seen in the past. Only since it wouldn't be as public, no one would really be aware of it. (Assuming most people won't broadcast what they're doing in their homes during their waking hours.)

Its OK, advertisers will pick up the tab.

Except they’re also useful idiots for a company who wants to push heroin on everyone, and has a lot of money to do so.

As a mostly fraudulent software developer, I've always considered it a privilege to earn copious amounts of money sitting in my bedroom.

Most popular languages were designed with human writability in mind. I’m wondering how languages will evolve if that becomes less of a factor going forward. Of course, for the time being, we need a middle ground to satisfy both humans and LLMs writing code.

Edit: tweaked the title slightly

I see human readability as being a key criterion moving forward since humans will be reviewing the code for the foreseeable future.

I also see functional and pure functional languages gaining in popularity because they're easier for LLMs to reason about, and they're easier to apply automated verification rules.

This is in the application space. It'll be interesting to see if any significant changes happen in the systems development space. Will LLMs drive the further adoption of Rust into the systems space?

It won't work out for riders initially since wait times for women drivers will be egregious so people will just switch back to no preference.

But I'm curious how many women would now feel safe enough to sign up as drivers given this option.

If it does take off, male drivers won't get as many riders but that's ok since their demand was inflated by lack of choice anyway.

Are drivers required to provide sex assigned at birth? If not, we might see male drivers conveniently identifying as women to circumvent this (presumably to minimize wait times between rides). Although I guess they'd get canceled on a lot.

This is interesting, it made me think of trans genders and other minorities.

It would be nice to select LGBTQ friendly rides for example.

I also wonder if you transition, can you change your "sex" on Uber? how would that work and how would they prevent abuse?

I think the attack vector is traditionally opportunistic. Horny driver suddenly has a vulnerable woman in his car and they're in a secluded area, boom there's an assault.

Any driver who is so premeditated about his assault plans that he would sign up to Uber pretending to be a woman probably has easier and more direct ways to access victims that are less likely to blow up in his face.

The premise here is not that men would "pretend to be women" (and sooner or later, a trans activist will decide that this charge has been levied at the wrong target) as part of a "premeditated assault plan".

The premise is that men would do it either in order to protest the policy, or in order to retain access to business that they had before.

This has been the case for a while with search engines. I'm convinced our brains have evolved (atrophied?) to avoid having to remember things that you can simply look up on your phone in a matter of seconds.