Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: How are you using GPT to be productive?

news.ycombinator.com

Post reply on HN
Print view View original

601–610 of 752 posts

Re: Ask HN: How are you using GPT to be productive?

#601

Post by gbro3n »

telchior wrote: post #571
gbro3n wrote: post #567

Earlier quoted context omitted.

I wonder if as more and more online content is AI generated, it will be harder to find human generated content to train the AI's on? Like a cumulative echo effect.

I've actually wondered if a job may exist in the future that's effectively just AI documentation. That's already what you have with power users on, say, Stack Overflow providing a ton of content that ChatGPT basically reprints; they don't even get paid for it. The cool and interesting thing about that theoretical job is that the writers of it wouldn't have to write well; they could just slop out a ton of information…

I can see how that could happen. But AI presumably knows how to output well written text because it's trained on well written text. If it's fed it's own output, I imagine that quality could degrade over time.

Re: Ask HN: How are you using GPT to be productive?

#602

Post by schappim »

I have used it to seek refunds[1] and write Swift code to OCR documents[2].

And perform a bunch of manipulation on text, including[3]:

  • Convert Text to 1st Person
  • Text to JSON arrays
  • Naming ideas...
  • Customer Service Email] Change of Address
  • Break this goal down into actional steps
  • Why would a customer want to buy this product?
  • Rewrite this product description
  • Rewrite product description for readability and conversions
  • What are the implications?
  • Inverse Thinking
  • 5 Whys
  • Convert the bullet points into a customer service email (reply email)
  • What keywords would someone searching for XYZ use?
  • Rewrite for readability
  • Convert to 3rd party todo
  • Rewrite this a customer service email
  • Summarise and Convert to Bullet Points
[1] https://files.littlebird.com.au/Screen-Shot-2023-03-21-at-8....

[2] https://gist.github.com/schappim/477125ff58f021641434d602e98...

[3] https://files.littlebird.com.au/Shared-Image-2023-03-26-10-2...

Re: Ask HN: How are you using GPT to be productive?

#603

Post by schappim »

themodelplumber wrote: post #110
jmann99999 wrote: post #15

Generally rewriting emails for clarity... but I found another neat use of GPT-4. For public APIs, I ask to make sure its aware of the api. Then I ask for endpoints. I find the endpoint I want. Then I ask it to code a request to the endpoint in language X (Ruby, Python, Elixir). It then gives me a starting point to jump off from. Thirty seconds of prompt writing saves me about 20 minutes of getting setup. Yes, I have…

You reminded me: I discovered that ChatGPT had invented an API for me. Has that happened to you yet? Since it went to the trouble of writing code for the API as well, I contacted the API developers to follow up about the topic. The code given was kind of a hand-wave anyway so I'd need to polish it up. The developers were surprised to hear they had an API. In truth, there was no such thing. I then found myself in one…

This has improved significantly between 3, 3.5 and now 4. It used to create a lot of Apple Frameworks/Classes and Methods, many of which would have been useful if they actually existed.

Re: Ask HN: How are you using GPT to be productive?

#604

Post by xkcd1963 »

VoodooJuJu wrote: post #330

Useful things: - As a thesaurus - What's the name of that "thing" that does "something" - kind of like fuzzy matching - A starting point for writing particular functions. For example, I wanted a certain string-manipulation function written in C, and it gave me a decent skeleton. However they're almost always very inefficient, so I have to optimize them. Things I've tried, that others seem to be blown away by, that I…

Yes in many ways it is just for avoiding some additional clicks

Re: Ask HN: How are you using GPT to be productive?

#605

Post by schappim »

itsuka wrote: post #123

Since I use ChatGPT regularly, I decided to create my own client. I prefer to avoid third-party services that require privilege escalation like Grammarly and Copilot. I have developed distinct profiles for different tasks, each with its own system prompt and input method. After getting the hang of it, I plan to tweak the parameters as well. Here are some of the profiles: Explainer: a default, general purpose Q&A. The…

Yeah, I built myself something similar for 1-shot tasks[1], examples[2].

[1] https://files.littlebird.com.au/Screen-Recording-2023-03-26-...

[2] https://files.littlebird.com.au/Shared-Image-2023-03-26-10-3...

Re: Ask HN: How are you using GPT to be productive?

#606

Post by bumbledraven »

VierScar wrote: post #598
thih9 wrote: post #230

Earlier quoted context omitted.

I’m unfamiliar with gpt and tokens, what would “skip prose” change?

The annoying habit GPT has of restating the question as the start of the answer. "How do I write a function in JavaScript to sort a list of cities by population grouped by region? GPT: To sort a list of cities by population grouped by region, in JavaScript, would at a high level require 3 steps...." Super frustrating when you just want the answer without the "polite professionalism"

Q: How do I write a function in JavaScript to sort a list of cities by population grouped by region? Just show me the code, no prose.

GPT-4:

  function sortCitiesByPopulation(cities) {
    return cities.sort((a, b) => {
      if (a.region === b.region) {
        return b.population - a.population;
      }
      return a.region.localeCompare(b.region);
    });
  }
  …

Re: Ask HN: How are you using GPT to be productive?

#607

Post by hypertele-Xii »

HDMI_Cable wrote: post #587
hypertele-Xii wrote: post #572

Earlier quoted context omitted.

> "How do I select rows of a dataframe where a column of lists of strings contains a string?" Literally just googled that and the first result: https://stackoverflow.com/questions/53342715/pandas-datafram... You're not using it like Stack Overflow. It's actually regurgitating Stack Overflow, except with errors hallucinated in.

Even with that caveat, using GPT in this way is still useful. The amount of time spent to simply ask GPT-4 is a lot lower than to search StackOverflow, and while this problem is so basic that the first result often works, once one gets into complex problems that massively benefit from input context, I think GPT-4 would save massive amounts of time.

Since you're gonna have to search Stackoverflow anyway to verify that ChatGPT didn't hallucinate garbage, I'm very dubious that it actually saves any time, let alone "massive" amounts of it.

Re: Ask HN: How are you using GPT to be productive?

#608

Post by toomanyrichies »

boredtofears wrote: post #585
toomanyrichies wrote: post #520

I’m using it as a technical editor for a book I’m writing on bash for beginner programmers. Compiling this book from scratch is a one-man operation; I don’t have the budget or time to hire a technically-competent bash programmer who is also willing to act as a proofreader and editor. Instead, I plug in certain paragraphs and ask “Does the following paragraph about file descriptors / environment vs. shell variables /…

> But I can ask ChatGPT that same question, and get a correction in my understanding without the loss of those sweet, sweet internet points. How do you ever know when the correction it makes is wrong?

Mostly I just Google the information it tells me, and see if I can find confirmation from sources known to be written by humans, such as official docs or other equally-trustworthy sources.

For example, just today I asked ChatGPT "In UNIX, what is the difference between redirecting with symbols like > and , and | confused Google, and the first-page search results were all irrelevant.

However, ChatGPT told me that the difference was that were used for "redirection" of input and output into or out of a file, where as the character | was used for piping output into a command. Based on this, I was able to Google "piping vs redirection" and find confirmation from Stack Overflow.

Re: Ask HN: How are you using GPT to be productive?

#609

Post by hosteur »

TacticalCoder wrote: post #52

Tangentially related but... There's an issue related to training models with data that doesn't have a license allowing it. I don't know if it'll hold up in court but here's my prediction: we'll see an open source license protecting code welcoming using the repository to be used to train future models IFF the models are then made public. And private models are going to be given the finger by the license and won't be a…

Here's a potential amendment that could be added to the GPLv3 to allow for the use of source code in training AI models:

"In addition to the permissions granted by Section 2 of this License, the source code distributed under this License may be used in the training of artificial intelligence models, provided that:

a) The resulting models are made available to the public under a free software license that allows anyone to use, modify, and distribute the software without any additional restrictions; and

b) The models are not used for any commercial purposes, including but not limited to training proprietary models or selling access to the trained models.

Any use of the source code for training proprietary models or for commercial purposes is strictly prohibited. This amendment shall be effective immediately upon adoption and supersedes any conflicting terms of this License."

Re: Ask HN: How are you using GPT to be productive?

#610

Post by jhugo »

imiric wrote: post #595
jhugo wrote: post #409

Earlier quoted context omitted.

> Think of the amount of boilerplate and tests you write, and tedious API documentation lookups you do daily; that all goes away with GPT. At work we have really worked hard to minimise boilerplate and manually-written/repetitive tests, so I don't write much of that. Getting GPT to write it would certainly be worse: we would still have the deadweight of boilerplate/repetition even if we didn't have to write it, and s…

> At work we have really worked hard to minimise boilerplate and manually-written/repetitive tests, so I don't write much of that. There's boilerplate in any codebase, even if you make an effort to minimize it. There are always patterns, repeated code structure, CI and build tool configuration, etc. If nothing else, just being able to say "write a test for this function", which covers all code paths, mocking, fuzzing…

This may be highly dependent on problem domain or programming language (see the other article about GPT tending to hallucinate any time it is given problems that don't exist in its training set). My experience has mostly been that the output (including simple stuff like "test this function", though we generally avoid unit tests due to low benefit and high cost) is consistently so flawed that the time to fix it approaches the time to write it.
Post reply on HN