The boilerplate code problem was solved by Lisp in the 1970s but we keep insisting on coming up with worse solutions because ‘parens look funny’
40% of the code GitHub Copilot users check-in is AI generated and unmodified
51–60 of 129 posts
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#52Earlier quoted context omitted.
In my case about 50% of the test code is just boilerplate, so I usually type the test name and generate the rest. Most of the time I have to rewrite the actual test logic (although very trivial tests are sometimes correct), but I probably keep more than 40% if we're talking tests specifically.
> In my case about 50% of the test code is just boilerplate Why not replace it with a function call?
That makes the test harder to read. Tests should be dumb.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#53Sure this must be using the most generous possible interpretation of the data. In my IDE the suggestion is shown as an autocomplete option on basically every keypress. I will only hit tab to insert if it is what I was planning to type. Sometimes if I really want to use it I'll coax it, write a comment or edit my code a bit to give it some help until the suggestion is what I want. There's no way they're counting every…
It's 40% of submitted code, not 40% of suggestions.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#54Earlier quoted context omitted.
The real question is, why we need to write "trivial" code? I think people have lost the plot there. We should be writing better compilers and libraries, instead of using AI that will write more code to deal with. The obsession with code generation instead of creating proper abstractions is a scourge of today's programming. It's a regression in software engineering.
I agree to some extent. But there are redundancies that are inherently beneficial and make code more robust and readable, like separating interface from implementation (where some declarative elements have to be repeated), or explicit type annotations instead of full-program type inference. There are also strictly necessary redundancies, like when you pattern-match an ADT, you repeat its declared structure. For desir…
This is trivial stuff which is handled well by any competent IDE from 2015, and even a well enough-working set of editor templates.
I don't see where "AI" is required for stuff like this.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#55I wonder if it’s 40% of the modules,40% of the functions, 40% of the lines or 40% of the tokens.
The first would be very impressive, the last not so much.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#56Earlier quoted context omitted.
I agree to some extent. But there are redundancies that are inherently beneficial and make code more robust and readable, like separating interface from implementation (where some declarative elements have to be repeated), or explicit type annotations instead of full-program type inference. There are also strictly necessary redundancies, like when you pattern-match an ADT, you repeat its declared structure. For desir…
This is trivial stuff which is handled well by any competent IDE from 2015, and even a well enough-working set of editor templates. I don't see where "AI" is required for stuff like this.
I agree, but also, this is actually where AI (whatever it is) would be helpful. Suggesting actual abstractions that could replace boilerplate. Of course, these can build on existing abstractions.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#57The boilerplate code problem was solved by Lisp in the 1970s but we keep insisting on coming up with worse solutions because ‘parens look funny’
Mostly because macros are unintuitive. But look, higher-order functions went mainstream a few years ago; immutable data structures, too. Sum types claw their way into mainstream, first to replace null, then for error handling. The industry is not hopeless.
(are (excessively funny) (serialized (of (parse tree) structures)) though) ;; this limits direct Lisp adoption. Python and JS / TS took its place now.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#58So how long before exploiters train code completion ais to generate back doors in the mountains of other boilerplate?
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#59Earlier quoted context omitted.
That was my exact thought. I don't use copilot but I'd wager that at least 40% of the code I check-in is autogenerated as well. A fresh modern JS project is 15+ files full of boilerplate around a single "hello, world" statement. If there was a metric for how much net new application business logic was written by copilot, that would be far more interesting to study.
I just started using Copilot and so far, it really does seem to save me time with rote stuff like imports, function definitions, and where I would otherwise be copying / pasting code I just wrote (say, while writing a bunch of tests). Even when it gets those slightly wrong, it still saves me time or at least makes it feel that way. So overall I'm honestly pleasantly surprised with it. It feels similar to when I got a…
To me, depending on the language, it's less useful than coc-nvim (too slow, breaks my tempo, too much added code, sometimes add stuff I did not ask for, or use old, inefficient patterns you only really find on shitty websites), so I still write the 'clever' part with vim, but when I write tests I now use VScode because copilot is really helpful (and it's funnier to use it than to write unit tests that work).
For languages I do not know well however I use it for everything. But I wonder if it keeps me from learning tbh.
Re: 40% of the code GitHub Copilot users check-in is AI generated and unmodified
#60Sure this must be using the most generous possible interpretation of the data. In my IDE the suggestion is shown as an autocomplete option on basically every keypress. I will only hit tab to insert if it is what I was planning to type. Sometimes if I really want to use it I'll coax it, write a comment or edit my code a bit to give it some help until the suggestion is what I want. There's no way they're counting every…
It's 40% of submitted code, not 40% of suggestions.
I guess you're right, I initially read it as "40% of AI suggestions don't get modified", not "40% of code is AI generated".