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

It doesn't have problems with undefined behavior, memory safety, or especially thread safety?

That has not been my experience when using Codex, Composer, Claude, or ChatGPT.

Things have just gotten to the point over the last year that the undefined behavior, memory safety, and thread safety violations are subtler and not as blindingly obvious to the person auditing the code.

But I guess that's my problem, because I'm not fully vibing it out.


You just tell it the problem and it'll fix it. It's almost never been an issue for me in Zig.

Do you really think the user didn't try explaining the problem to the LLM? Do you not see how dismissive the comment you wrote is?

Why are some of you so resistant to admit that LLMs hallucinate? A normal response would be "Oh yeah, I have issues with that sometimes too, here's how I structure my prompts." Instead you act like you've never experienced this very common thing before, and it makes you sound like a shill.


Yeah. I've got some EE coworkers that are vibe coding their way through everything and nothing in the codebase is understandable.

We're going to have to go through another quality hangover I suspect.

But since people that have never coded are now coding and think it's the best thing ever the only way out is through.


It has ever been thus. There are multi-million dollar businesses propped up by .NET applications on a foundation of shunted-around files, and at best, SQL used as APIs/queues. "Working" code is, in the long run, a liability outside the hands of those doing real engineering.

Yeah. I unfortunately moved to an APU where code size isn't an issue so I never got the chance to see how well that analysis translated to the work I do.

Provocative talk though, it upends one of the pillars of deeply embedded programming, at least from a size perspective.


This feels a lot like trying to sanitize database inputs instead of using prepared statements.

What's the equivalent of prepared statements when using AI agents?

Don't have the AI run the commands. You read them, consider them, and then run them yourself.

It only ends up in the DNA if it helps reproductive success in aggregate (at the population level) and is something that can be encoded in DNA.

Your comparison is nonsensical and simultaneously manages to ignore the billion or so years of evolution starting from the first proto-cell with the first proto-DNA or RNA.


You're communicating ideas across unknown thousands of miles with a stranger in near realtime and are able to comprehend each other, for one.

No cat or dog has managed that feat yet.

No cat or dog has managed to reproduce fire to the degree that evolution has changed their gut to adapt to the increase in available calories.

The big brain comes with down sides, but one thing it does have is utility.

Germ theory of disease has made it so a scratch isn't fatal anymore. Why, after all, do cats play with their prey? To tire it out so there's less chance of injury when they go in for the kill.

We just figure out how to farm it instead and mold it to our needs.


I don't disagree with any of this, but what is the utility of viewing this ability as "more advanced"?


What is the utility of denying it?

What do you or anyone else actually get from such obvious absurdity, I wonder?

If it helps - and I have doubts - does (say) a working knowledge of Galois theory require more advanced mathematical cognition than arithmetic?

Would it be immoral to introduce such ghastly, hierarchical language? Etc.

I see you ignored the obvious rejoinder downthread, which stated that the utility of classifying behaviours or capacities is to help you predict outcomes.

How much more help do you need here? It’s not very complicated, but you prefer to showboat.


> What is the utility of denying it?

Speaking in material terms allows clearer communication of meaningful concepts than floating signifiers. "Advanced" is just a meaningless concept.

> I see you ignored the obvious rejoinder downthread, which stated that the utility of classifying behaviours or capacities is to help you predict outcomes.

It also helps you mispredict outcomes


Let's say you're about to embark on a cross-oceanic sailing voyage. For safety reasons, you think it's best to bring another living being with you who can help if things go south or you are incapacitated.

Are you going to bring another human, or a goat? Can a goat navigate while you sleep? Can it apply first aid to you? Can it respond on the VHF radio if you get hailed? Can it operate the bilge pump?


Embarking on a cross-oceanic sailing voyage seems to be a particularly human brand of tomfoolery. Why not just stay at home with the goat?


I honestly can't tell if you think you're being funny, deep, or just trolling.


No, it was a serious question


In that case, why do anything? Why leave the house? Why build a house? Why not lay in the dirt? Why exist?


> I don't disagree with any of this, but what is the utility of viewing this ability as "more advanced"?

Because that's the most accurate description of what it is. The more accurately you describe something, the more effectively you communicate, an aspect of more advanced cognition.


It's only accurate if you understand what the meaning of advanced is, and it has no clear semantics or referent. It's a floating signifier.


It's solving the wrong problem, and doesn't help the typical situation of being on hills, pedestrians, bicyclists, etc.

Just turn the damn maximum output down.


I have a car with LED lights. It's easily the best car I've had for vision at night. We very occasionally get someone flashing us at night, wrongly believing our high beams are on.

However, from a safety point of view, I'm not convinced the trade off is actually in favour of reducing illumination for everyone.


If I flash a car, it means they're blinding me. I don't care if it's their high beams or not. It doesn't matter.


No one is truly blinding you. Even old incandescent headlights can be unpleasant. Some people are more sensitive to it than others and things like a car coming over a slight hill or bend in the adverse direction can change the alignment of the lights in such a way that they appear much brighter.

The point I was trying to make is that reducing the brightness isn't a simple trade off. How many accidents are caused by people being "blinded" vs people not seeing something until it was too late?

If it needs regulation to fix then that regulation should try to balance those things. Perhaps by automatically adjusting the headlights when another car is detected (maybe matrix style headlights, or a simple angle adjustment).


That's still ignoring the impact on bicyclists, pedestrians, and cars it can't detect because it's not a spherical cow on a uniform plane.

Look at the output of a car from 10 years ago, 20 years ago, and 30 years ago compared to today.

Each is progressively dimmer with their low beams. Modern low beams are brighter than the high-beams of yesteryear!


The difference is THEY DID IT.

Photoshop is not real.

This was real.

This was recorded.

The value is in the authenticity and execution of a cool idea no one else has done before.


DNG isn't exactly rarely used, it's Adobe's open raw format and lots of image processing programs read and write it.

It's not a codec made for one game back in 1994, it's still in very active use today if you're using such a rare and uncommon program as Photoshop.


I agree it is a widely used format, however, it is rarely used on iOS devices.


Old reddit is unfortunately just a rounding error. I weep for the day they decide to kill it.


I don't, the utility of reddit has declined over time for me but there are still a handful of reddits that I enjoy but them killing old.reddit.com is absolutely what will push me off the platform entirely.

Though at this point I spend (or waste depending on PoV) much less time on reddit than I used to.


I'm sure they're going to eventually. They say they won't, but they deprecated private messages. Yeah, I did use those lol.


weep? I'd finally be free. I wish this site would disappear too. Whoever designed these algorithms got me good, at a young age, and I don't think these sites have been a net positive overall or on me personally


We are already free. If we keep returning to a few subreddits, it's because we can't find an equivalent community elsewhere. If they kill the old interface, we'll eventually use the new one if there are no other alternatives, no need to lie to ourselves.


Nah, I just leave. Because I want too, and because I don't like Reddit's AI training ppolicies.


Iam nearly 15 years on reddit now and I would miss it if i cant use old. or a good client. Iam sure sooner or later it will happen and ill most probably leave.

Reddits quality went downhill over the years but there is more or less no successor/competitor. It will be over and buried forever. Eternal september gets them all.

Side note: be free if you want to and dont make it dependent on decisions others do for you.


I am in the same boat. I think I will get a lot more of time back if they kill old


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

Search: