I remember another company called Kite [1], working on a similar approach - smart autocompletion - this however uses GPT-3 so it's a bit different I guess because it doesn't detect what you're trying to do but rather transforms technical natural language into code. Right?
GitHub Copilot
91–100 of 1001 posts
Re: GitHub Copilot
#92Hi HN, we've been building GitHub Copilot together with the incredibly talented team at OpenAI for the last year, and we're so excited to be able to show it off today. Hundreds of developers are using it every day internally, and the most common reaction has been the head exploding emoji. If the technical preview goes well, we'll plan to scale this up as a paid product at some point in the future.
This looks really cool. Do you plan to release this in some other form like a language server so that it can be easily integrated to other editors?
Re: GitHub Copilot
#93Hi HN, we've been building GitHub Copilot together with the incredibly talented team at OpenAI for the last year, and we're so excited to be able to show it off today. Hundreds of developers are using it every day internally, and the most common reaction has been the head exploding emoji. If the technical preview goes well, we'll plan to scale this up as a paid product at some point in the future.
This is obviously controversial, since we are thinking about how this could displace a large portion of developers. How do you see Copilot being more augmentative than disruptive to the developer ecosystem? Also, how you see it different from regular code completion tools like tabnine.
Re: GitHub Copilot
#94Should I be impressed that the example parse_expenses.py on the home page doesnt include any error handling and uses a float for currency? This seems like it's going to revolutionize copy and paste programming.
It's a copilot. You're still the pilot. To be honest this seems like it can definitely save me a bunch of googling and let me stay in the ide.
in their own example the copilot aimed for the mountains with float() - thanks but no thanks
Re: GitHub Copilot
#95So if it was trained using "source code from publicly available sources, including code in public repositories on GitHub." was it also GPLv2? So everything generated also GPLv2?
I think this would fall under any reasonable definition of fair use. If I read GPL (or proprietary) code as a human I still own code that I later write. If copyright was enforced on the outputs of machine learning models based on all content they were trained on it would be incredibly stifling to innovation. Requiring obtaining legal access to data for training but full ownership of output seems like a sensible middl…
And so it begins: We start applying human rights to AIs.
Not a critique on your point, which a was just about yo bring up myself.
Re: GitHub Copilot
#96I would love to see an equivalent where it generates all the tests for you
https://www.diffblue.com
Here you go :)
Re: GitHub Copilot
#97I would love to see an equivalent where it generates all the tests for you
It should be able to (try to) do something like that too. There's a little demo about that here: https://copilot.github.com/ It's "just" an autocompletion system basically, if you write something that looks like the beginning of a test it should understand that and try to autocomplete that.
I'd say it's more than 'just' an autocomplete system.
Naive autocomplete, as implemented in Excel since forever ago (and I'm sure long before that, I'm just familiar with being annoyed by Excel suggesting wrong entries from its simple and over-eager autocomplete system), merely matches a sequence of characters - if I typed "aut" again in this paragraph it will suggest "autocomplete" because I recently typed it. Implementing it is the kind of task you give to a first-year programming student to practice string matching data structures, similar to a spell checker that merely checks that a string exists in a dictionary.
There's a spectrum from 'just' autocomplete, to a syntax-aware system like VS Intellicode, to this, and eventually beyond this. As mobile predictive text is to a spell checker, so Github Copilot is to autocomplete. As mobile predictive text is to GPT3 [1], so Github Copilot is to...what next? GPT3 is not just a spell checker.
[1] Also by OpenAI: https://news.ycombinator.com/item?id=23345379
Re: GitHub Copilot
#98I wonder how Copilot's suggested snippet compares if the comment is a CS1 homework prompt.
Re: GitHub Copilot
#99Re: GitHub Copilot
#100This is very impressive! OpenAI’s tech opens an ethical Pandora’s box: 1. It’s clear that the raw inputs to all of OpenAI’s outputs originated with real, human creativity. 2. So, in a sense, OpenAI is laundering creativity. It reads in creative works, does complicated (and, yes, groundbreaking) transformations, and produces an output that is hard to trace to any particular source. 3. Yet, isn’t that effectively what…
s/OpenAI/Photoshop
Reads similarly :)