Live data from Hacker News

GPT-4 is phenomenal at Code

github.com

181–190 of 221 posts

Re: GPT-4 is phenomenal at Code

#181

Earlier quoted context omitted.

Yes. The privacy concerns are huge. With that said, id use the shit out of this for open source development where privacy is not a concern.

Instead you have legal concerns when you are re-licensing code you don't know where it comes from. Someone could have published a snippet of code with retained copyright (not FOSS) and GPT-4 slurped it up, splitting it back to you and somehow the law firm representing the author comes across it.

This sounds like plausible deniability, and not legal problems, at the hobbiest level. Not something I am going to worry about.

Re: GPT-4 is phenomenal at Code

#183

Earlier quoted context omitted.

Yes. The privacy concerns are huge. With that said, id use the shit out of this for open source development where privacy is not a concern.

Instead you have legal concerns when you are re-licensing code you don't know where it comes from. Someone could have published a snippet of code with retained copyright (not FOSS) and GPT-4 slurped it up, splitting it back to you and somehow the law firm representing the author comes across it.

Have there been any tried cases of this happening yet, or is everyone still waiting for the legal precedent to be set for this situation?

Re: GPT-4 is phenomenal at Code

#184

> is phenomenal at Code Can it - attend bullshit meetings - lick my manager's boot - tell them that their ideas are good but in the masochistic, self-deprecating way they like to hear it from me in Not yet? Darn it, guess my job is safe for now.

I think it can actually do all those things reasonably well. I’d like to start getting meeting summaries that tell me my action items based on a transcript. And it can generate infinite boot-licking butt kissing if you just ask it. Generate and click send to your boss!

Teams is already halfway there: https://www.theverge.com/2022/10/12/23400332/microsoft-teams...

Just needs a little more bootlicking.

Re: GPT-4 is phenomenal at Code

#185

I'm the "really good", but definitely not "phenomenal" camp. That is, it works - but what is with the repeated class declaration for every h3 element? 1. Pruning The whole point of CSS is to have to declare properties like this once, and only once, for a group of functionally similar elements. Also note that switchTab has to be told the position of the tab it's switching to (it should be able to infer that the elemen…

They're using TailwindCSS for this, which is full of one-liner utility classes that effectively "replace" CSS classes by moving the equivalent CSS into your HTML.

However, what they're not using is Tailwind's `@apply` directive [1] which basically encapsulates these repeated declarations with a repeatable, named CSS class.

Doing so would replace their h3 tags with something like

    1. Pruning
and CSS of

    h3.header {
        @apply text-xl font-bold mb-2;
    }
It's not the biggest change and it's debatable how much cleaner the code ends up being. In my experience with TailwindCSS, I don't see a lot of companies using @apply until late in the code lifecycle, if at all. It's often way faster to just use the utility classes directly, and then you find yourself with a site that's already built and looks great -- and very little reason to allocate dev time to go back and abstract away repeated CSS.

Tailwind also has this to say WRT using it:

> Whatever you do, don’t use @apply just to make things look “cleaner”. Yes, HTML templates littered with Tailwind classes are kind of ugly. Making changes in a project that has tons of custom CSS is worse. If you start using @apply for everything, you are basically just writing CSS again and throwing away all of the workflow and maintainability advantages Tailwind gives you.

[1] https://tailwindcss.com/docs/reusing-styles#extracting-class...

Re: GPT-4 is phenomenal at Code

#187

Earlier quoted context omitted.

Maybe it can translate python code to C++. That is interesting. I need to try that.

It definitely should be able to. Any example you would like me to try. Put in an issue and we will do it.

I was about to write an old-fashioned Wu anti-aliased circle-drawing routine for a simple C-like language (it has no structs nor unions yet, for instance), is that something that might be an interesting test case? Worth opening an issue?

Great work BTW. :)

Re: GPT-4 is phenomenal at Code

#188

Earlier quoted context omitted.

> Unfortunately it will also mean a lot of junior folks may never get back into software jobs. Why do you suppose this age-based partitioning is going to happen? There are a lot of senior people who are less effective than bright juniors. Being a good software engineer can always be learned. I think what's more likely is that the quality bar in general will be raised across all age bands.

I don't at all mean "junior" or "senior" as "age" but capabilities. I know lots of young senior/staff :) And plenty of new developers in their 40s plus.

At this point these titles really have become pretty meaningless... A good developer will build and use the tools they have access to, a bad developer will fall behind.

Re: GPT-4 is phenomenal at Code

#189
I just asked GPT-4 to "generate an ansible playbook that will install oh-my-zsh on a linux computer?" It did a much better job of the playbook than the one I created last week. GPT-4 did it in a few seconds, and mine took about an hour of Googling.

Re: GPT-4 is phenomenal at Code

#190
post #50

Earlier quoted context omitted.

Yes, it can. - Automated relations, whether with your employer or a customer, are trivial - These relations can be fine-tuned for exactly the level of syncophancy your boss prefers - So yes, it can tell your boss their ideas are good in precisely the masochistic, self-deprecating way, etc. Your post is not longer than a system prompt. Think it through.

But it will have zero credibility, so no.

It will absolutely have great credibility. Did you read any of the code samples in the GH repo linked by OP?

Here's the algorithm to get you appropriately scared.

- Measure the amount of time it took to go from GPT 2 to GPT 3

- Measure the amount of time it took to go from GPT 3 to GPT 4

- Measure the code quality improvement for the same spans. Go ahead. Pick a metric.

You'll notice that, not only is GPT-4 really fscking good, it represents a year-over-year improvement that is accelerating.

What you are likely looking at is a classic hockey-stick exponential graph. And there is no ceiling yet in sight.

Post reply on HN