LLM Coding Tools in Xcode
Kade Killary (via Hacker News):
Therefore, in this article I’ll cover how to catapult your productivity to the heavens via a series of command line wrapper functions around the OpenAI API.
[…]
I think the real differentiator in the coming months/years will be developer productivity. Good, bad or indifferent there seems to a be an unavoidable advantage to those who aggressively adopt workflows that lean into AI versus fighting, or ignoring, it. I think there is an oncoming paradigm of rapid prototyping combined with simulation that will help to isolate and test various components of a project. Additionally, the ability for translating languages and developer productivity in new languages.
[Google Codey is available] through extensions in all major IDEs. Except of course xcode. Because Apple protected us from a rogue plugin by killing plugins.
(yes - technically extensions are still a thing, but they’re so limited they might as well not be)
Xcode source editor extensions are not at the same level, but there is one for Copilot. Christian Tietze:
Just yesterday, I used the Copilot Xcode plugin to write a lot of boilerplate for me. I can confirm it does its job.
I used it to help me write a unit test spy for an umbrella repository, or façade, that conforms to 5 protocols. In my human mind, there was rhythm and similarity, but code completion can’t help there.
And there are a variety of tools that give you a text box outside of Xcode:
At a high level, S-GPT is a shortcut that lets you ask ChatGPT questions from an input box on your iPhone, iPad, or Mac; answers are returned by ChatGPT and displayed in an alert on your devices. You can ask whatever you want, wait a couple of seconds, and get a response back from the assistant. S-GPT only supports text, and there are no limits in terms of question length.
Easily access ChatGPT from the menu bar, Dock, or using a keyboard shortcut.
Joe Fabisevich (via John Gruber):
Short Circuit works on on your iPhone, iPad, and Mac. Now you can have a premium ChatGPT experience no matter where you are or what you’re doing.
Previously:
- Google Codey
- Cheating Is All You Need
- ChatGPT Plug-ins
- GitHub Copilot X
- Petey 1.1.2
- ChatGPT for Apple Platforms Development
- Xcode 14: Swift Package Plug-ins
- Xcode 8 Tips and Issues
- An Xcode Plug-in for Unsmoothed Text
- Xcode’s Plug-In Interface
Update (2023-05-29): Adam Wulf:
There are many tasks during my development day that take up significant time, not because they’re complicated, but because they’re tedious. With Developer Duck, these tasks are significantly faster. I use it for adding DocC comments, writing simple extensions, writing tests for those extensions, and even researching unfamiliar APIs.
1 Comment RSS · Twitter · Mastodon
I wonder how many regrets people will have if the training of these large language models is ruled not-fair-use in the markets these people's products are sold in. (e.g. the EU).
In the old days, one had to have one team document an API by reverse engineering it, and a separate team to implement it, if one wanted to legally duplicate functionality: the documents made it clear that no copyright violation occurred, otherwise the presumption would have been the code was copied. (The reproduction of IBM's BIOS by Compaq IIRC).
Nowadays? A vast neural network no one understands, but basically interpolates between examples it saw which form the underlying structure of the neural network's manifold (see Pedro Domingo's work).
Do we really believe the legal profession has changed its mind this far? Is it A-OK to make a Mickey Mouse movie based on a "fair use" sampling of Disney's work?
