4 ms·

Hm odd takes here. It's really weird for software engineers to judge something by its current state and not by its potential state. To me, it's clearly solvab

Hm odd takes here. It's really weird for software engineers to judge something by its current state and not by its potential state. To me, it's clearly solvable by Copilot filtering the input code by that repository's license. It should only be certain open source licenses, maybe even user-choosable, or code-creators can optionally sublicense their code to Copilot in a very permissable way. Secondly, a way for the crowd to code review suggestions would be a start.

slumdev 5y ago
> It's really weird for software engineers to judge something by its current state and not by its potential state. No, we're not afraid of Copilot replacing us. The thought is ridiculous, anyway. If it actually worked, we would be enabled to work in higher abstractions. We'd end up in even higher demand because the output of a single engineer would be so great that even small businesses would be able to afford us. Yes, we are afraid of Copilot making the entire industry worse, the same way that "low-code" and "no-code" solutions have enabled generations of novices to produce volumes of garbage that we eventually have to clean up.
- vmception 5y ago
  Sounds like projecting because thats not what I was referring to I’m saying copilot can be better with very simple tweaks
gpm 5y ago
Practically every open source license requires attribution, if copilot has a licensing issue, training a model on only repositories with the same license won't fix it except for the extremely rare licenses which do not require attribution.
- vmception 5y ago
  why not? it can just generate an attribution file or reminder
  - gpm 5y ago
    Because it's an opaque neural network on the backend, it doesn't know if or from whom it copied code.
    - buu700 5y ago
      Could they handle this by generating a collective attribution file that covers every (permissively licensed) repository that Copilot learned from? Of course this would be massive, so from a practical consideration the attribution file that Copilot generates in the local repository would have to just link to the full file, but I don't think that would be an issue in and of itself.
      - gpm 5y ago
        Maybe? Might depend on the license, I doubt the courts would be amused. Almost certainly a link would not suffice, basically every license requires that the attribution be directly included with the modified material. Links can rot, can be inaccessible if you don't have internet access, can change out from underneath you, etc. (I am not a lawyer, btw)
        buu700 5y ago
        Makes sense. Maybe something like git-lfs/git-annex would be sufficient to address the linking issue, but it seems like the bigger concern is whether a court would accept this as valid attribution. In a sense it reminds me of the LavaBit stunt with the printed key.
        joeyh 5y ago
        I think a judge could be persuaded that a list of every known human does not constitute a valid attribution of the actual author, even though their name is on the list. The purpose of an attribution is to acknowledge the creator of the work, and such a list fails at that.
mimixco 5y ago
I've been in the business a long time and I just don't believe in generalized AI at all. Writing code requires general (not artificial) intelligence. All of these "code helping" tools break down quickly because they may be searching for and finding relevant code blocks (the "imperative hell" referred to by another commenter), but they don't understand the context or the overall behavior and goals of the program. Writing to overall goals and debugging actual behavior are the real work of programmers. Coming up with syntax or algorithms are 3rd and 4th on the priority list because, lets face it, it's not that hard to find a reference for correct syntax or the overall recipe implied by an algorithm. Once you understand those, you can write the correct code for your project. I do think Copilot has potential as a search engine and reference tool -- if it can be presented that way. But the idea of a computer actually coming up with the right code in the full context of the program seems like fantasy.
- gpm 5y ago
  If we're coming up with potential uses, I think they got the direction wrong. Don't tell me what to do, tell me what not to do. "this line doesn't look like something that belongs in a code base", "this looks like a line of code that will be changed before the PR is merged". Etc.
  - mimixco 5y ago
    That would be fantastic! Imagine if it could catch common errors before you make them. So many things in loops and tests that we mess up all the time. My favorite is to confuse iterating through an array vs an object in JS. I'd love to have Gazoo step in and say, "Don't you mean, this, David?"

slumdev 5y ago

> It's really weird for software engineers to judge something by its current state and not by its potential state. No, we're not afraid of Copilot replacing us. The thought is ridiculous, anyway. If it actually worked, we would be enabled to work in higher abstractions. We'd end up in even higher demand because the output of a single engineer would be so great that even small businesses would be able to afford us. Yes, we are afraid of Copilot making the entire industry worse, the same way that "low-code" and "no-code" solutions have enabled generations of novices to produce volumes of garbage that we eventually have to clean up.

vmception 5y ago
Sounds like projecting because thats not what I was referring to I’m saying copilot can be better with very simple tweaks

gpm 5y ago

Practically every open source license requires attribution, if copilot has a licensing issue, training a model on only repositories with the same license won't fix it except for the extremely rare licenses which do not require attribution.

vmception 5y ago
why not? it can just generate an attribution file or reminder
- gpm 5y ago
  Because it's an opaque neural network on the backend, it doesn't know if or from whom it copied code.
  - buu700 5y ago
    Could they handle this by generating a collective attribution file that covers every (permissively licensed) repository that Copilot learned from? Of course this would be massive, so from a practical consideration the attribution file that Copilot generates in the local repository would have to just link to the full file, but I don't think that would be an issue in and of itself.
    - gpm 5y ago
      Maybe? Might depend on the license, I doubt the courts would be amused. Almost certainly a link would not suffice, basically every license requires that the attribution be directly included with the modified material. Links can rot, can be inaccessible if you don't have internet access, can change out from underneath you, etc. (I am not a lawyer, btw)
      - buu700 5y ago
        Makes sense. Maybe something like git-lfs/git-annex would be sufficient to address the linking issue, but it seems like the bigger concern is whether a court would accept this as valid attribution. In a sense it reminds me of the LavaBit stunt with the printed key.
        joeyh 5y ago
        I think a judge could be persuaded that a list of every known human does not constitute a valid attribution of the actual author, even though their name is on the list. The purpose of an attribution is to acknowledge the creator of the work, and such a list fails at that.

mimixco 5y ago

I've been in the business a long time and I just don't believe in generalized AI at all. Writing code requires general (not artificial) intelligence. All of these "code helping" tools break down quickly because they may be searching for and finding relevant code blocks (the "imperative hell" referred to by another commenter), but they don't understand the context or the overall behavior and goals of the program. Writing to overall goals and debugging actual behavior are the real work of programmers. Coming up with syntax or algorithms are 3rd and 4th on the priority list because, lets face it, it's not that hard to find a reference for correct syntax or the overall recipe implied by an algorithm. Once you understand those, you can write the correct code for your project. I do think Copilot has potential as a search engine and reference tool -- if it can be presented that way. But the idea of a computer actually coming up with the right code in the full context of the program seems like fantasy.

gpm 5y ago
If we're coming up with potential uses, I think they got the direction wrong. Don't tell me what to do, tell me what not to do. "this line doesn't look like something that belongs in a code base", "this looks like a line of code that will be changed before the PR is merged". Etc.
- mimixco 5y ago
  That would be fantastic! Imagine if it could catch common errors before you make them. So many things in loops and tests that we mess up all the time. My favorite is to confuse iterating through an array vs an object in JS. I'd love to have Gazoo step in and say, "Don't you mean, this, David?"