Command line functions around OpenAI • Hacker News Board

Re: Command line functions around OpenAI

#311

Post by isanjay » Thu, Mar 30, 2023, 10:36 AM UTC

daveslash wrote: post #163

ChrisBland wrote: post #61

Writing code is the basic 101 of our industry. Anyone can write a loop or an if statement. With todays computing power, for most applications even poorly written code will run and will execute quite fast. What makes someone a 10x (and i hate that term) is the ability to take requirements, or better yet, a vague idea given to them by the business and turn that into software that benefits the end user and makes the com…

RE>> Anyone can write a loop or an if statement. I had an intern a few years ago -- recent college grad in CompSci. I tried my best to lightly mentor him. One day I was talking about the diff between a compiled language and scripting, mentioned REPL. To demonstrate REPL, I opened up both the windows CMD prompt and the Chrome Developer tools. I mentioned that with a REPL like the Chrome Tools, it's trivial to do FizzB…

I know a Computer Science graduate (who is doing her MS now) tell me she didn't know what a program is. So...

Re: Command line functions around OpenAI

#312

Post by throwthrowuknow » Thu, Mar 30, 2023, 1:31 PM UTC

juujian wrote: post #282

throwthrowuknow wrote: post #201

Earlier quoted context omitted.

It’s also trained on code and books. The knowledge is in there, it just needs more RLHF to improve its abilities. There’s a good chance it can handle those edge cases by combining the information it has in different ways.

If the books have the specific code that is. If the books have some code and discuss edge cases, that is not something a language model could handle. It could give you the abstract concept that applies but it would fail when you prompt it to create new, valid code based on abstract concepts that are in the data.

Why do you think it can only copy existing code? I’ve used it for some really obscure cases and it’s still able to produce useful output both code and language. Try using it to create your own programming language with completely different keywords and oddball syntax and control structures and then ask it to write programs in it, it actually works!

Re: Command line functions around OpenAI

#313

Post by tzekid » Sat, Apr 01, 2023, 2:51 PM UTC

domepro wrote: post #300

nkingsy wrote: post #181

Earlier quoted context omitted.

I briefly worked with a software engineer who was making about a million dollars a year. I mentioned that an approach would be a lot of work. His response: “It’s just typing. Typey typey type (makes fingers on keyboard gesture).” Absolute lightbulb moment for me. To become a senior engineer is to realize that the code writing part is inconsequential as long as there are no technical unknowns in the work. It’s freed m…

Yeah, I'm really wondering if I'm dumb or something with this whole AI train, but my job as a developer that maintains existing (sometimes quite old) systems is much more about determining what actually needs to be done, gathering lost / not lost knowledge about the existing state of things in systems and figuring out where to change the code for maximum impact and minimal side effects. Lots of talking to customers/n…

Don't downplay yourself like that. A 25k token GPT model is a step in that direction but we're 3 - 15 years from your vision of valuable.

Currently, the Language Models (GPT or GithubCopilot) are mostly good as ... copilots. You bounce off ideas off them or use them to kickstart something you want to do. It's great for junior devs (me!) but it still cannot do big context or cognition (what you describe), that part will be done by humans for quite a while.

Re: Command line functions around OpenAI

#314

Post by readme » Sat, Apr 01, 2023, 5:07 PM UTC

swader999 wrote: post #225

readme wrote: post #187

Earlier quoted context omitted.

a good programmer is almost certainly a fast typer of course a fast typer does necessarily make a good programmer

Yes and no. Most of my stuff is automated, auto completion and micro code gen. My thinking is the bottleneck, never typing speed. Measured it last summer a few times trying to beat my daughter and I think it was about 60wpm. For coding if I know what I'm doing its probably 30wpm in my editor.

do you write documentation, though?

Re: Command line functions around OpenAI

#315

Post by antonjs » Sun, Apr 02, 2023, 1:41 AM UTC

This is a pretty cool idea, at least to experiment with. If you use ZSH, I made a ZSH plugin version based on the article: https://github.com/antonjs/zsh-gpt

Re: Command line functions around OpenAI

#316

Post by kzzzznot » Fri, Apr 07, 2023, 8:43 PM UTC

irrational wrote: post #306

ravenstine wrote: post #122

Earlier quoted context omitted.

Oh, my sweet, innocent child. Have you ever tried getting a company to fund a developer tool? More often than not, it's like pulling teeth, even if the service in question is a paltry $20 a month. I've learned to pay for things like OpenAI myself because it's not worth the hassle to get a company to pay for it (for the purpose of R&D).

I’ve been I’m working professionally in the industry for 23 years. I’ve never had a company refuse to purchase any tool I say I want or need. I wouldn’t work for any company that would. My hourly rate is way higher than the cost of any tool. It is ridiculous to not buy tools that might make me more efficient.

When that tool is a 3rd party which takes proprietary code as an input surely you can understand pushback from big software, financial services etc companies?