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

Using agents to edit code. And Helix doesn’t support live update of files. This is the reason it’s not my first choice.

AI can produce thousands of line of code. But that’s not the goal.

What is the goal?

Producing code that does what's intended. The metric is fuzzy and based on the usage of the software, not the scale of lines of code. The extent of the importance of the code itself is that I'm practice software tends not to be "one and done", so you need to be able to go back and modify it to fix bugs, add features, etc., and it turns out that's usually hard when the code is sloppy. Those needs still should stem from the sandal actual user experience though, or else we've lost the plot by treating the mechanism as the goal itself

Would my user rather have a program that works 100% in 2 weeks, or a program that works 80% in one day?

When the user needs a change made, would they prefer I spend another two weeks extending my perfect program, or throw a few LLMs at their sloppy code and have it done in a day?


That would depend on who your users are and what they're using their program for. My point is that the context of who is using the program, how they're using it, and what they're using it for are what actually matters, because most of the time, software that no one uses is by definition useless. There are circumstances where that might not apply, like code used as part of education or training (whether in a formal course or a project someone writes specifically because they're trying to learn something from the process) or when the purpose is aesthetic or humorous, but I'd argue that whatever process makes sense for them doesn't necessarily bear any resemblance given how different the goals are.

You're really asking if a user would want a program that fails a fifth of the times?

In some cases it might be better to have some crap right away and more cheaply, but even you would probably not like a 20% failure rate in most of the software you use.


Try this argument in any field where the stakes are high (e.g. medical diagnostics) and see how far this gets you.

This is really basic decision-theory stuff, often the cost of an error is far greater than the benefit of correctness.


What is art? What is the point of anything? Why write code instead of eating bananas all day? There is no answer to your question.

How is this different than RAG?

It doesn’t


Next.js without Tailwind


Wasn’t he fired by MS?


No. He is a Microsoft Technical Fellow and still a core contributor to the TypeScript compiler.


I used trial of Manus and it was in no way better than GPT. And they are using publicly available models only.


This material is best one you can find out there. You can always report typos.


> manual memory management

Rust has automatic memory management.

> Complexity tax

Could you be more specific?


The trait/type system can get pretty complex. Advanced Rust doesn't inherit like typical OOP, you build on generics with trait constraints, and that is a much more complex and unusual thing to model in the mind. Granted you get used to it.


OOP inheritance is an anti-pattern and hype train of the 90's/00's, especially multiple inheritance. Especially the codebases where they create extremely verbose factories and abstract classes for every damn thing ... Java, C++, and the Hack (PHP-kind) shop are frequently guilty of this.

Duck typing and selective traits/protocols are the way to go™. Go, Rust, Erlang+Elixir... they're sane.

What I don't like about Rust is the inability to override blanket trait implementations, and the inability to provide multiple, very narrow, semi-blanket implementations.

Finally: People who can't/don't want to learn multiple programming language platform paradigms probably should turn in their professional software engineer cards. ;)


> Rust has automatic memory management

Sure, if you define "automatic memory management" in a bespoke way.

> Could you be more specific?

The lifetime system, one of the most complex and challenging parts of the Rust programming model, exists only so that Rust can combine manual memory management with memory safety. Relax the requirement to manage memory manually and you can safely delete lifetimes and end up with a much simpler language.

Have you ever written a Java or Python program?


I doubt those specs are enough for running games at good graphics settings.


There is Quetta browser that is stable and support extensions on mobile.


Quetta's design is really bad, takes multiple clicks to switch from Private back to Normal browsing mode.


Quetta is unusable for me from how much they redesign the UI for no real reason. What I wanted was just an up-to-date version of Kiwi Browser which this project seems to aim for.


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

Search: