3 ms·

Do you mind explaining why? What do you think will happen? This is a serious question, I'm apparently just unaware of the horror-stories that can come out of b

Do you mind explaining why? What do you think will happen? This is a serious question, I'm apparently just unaware of the horror-stories that can come out of breaking copyright. (Not from the US so..) For me its most useful for helping with bash scripts and small simple stuff, just saves a huge amount of time I would spend googling and checking small things. Not sure if copyright is relevant there or not, it certainly isn't something I am worried about. Interested to hear what your fears are based on.

kmeisthax 4y ago
ML text and image generators have a habit of regurgitating training data, especially if you happen to use certain inputs or prompts that also match said data even superficially. For example, type in "fast inverse square root"[0] and get out a bunch of Quake source code. If you use code or art generated by an AI that was regurgitating training data, you can be sued for copyright infringement. The way that AI gets training data these days is... questionably ethical. It's all scraped off the web, because the people who make these AIs saw court precedent for things like Google Books being fair use and assumed it would apply to data mining[1]. Problem is, that does nothing for the people actually using the AI to generate what they thought were novel code sequences or images, because fair use is not transitive. [0] This won't work in current Copilot because a) I'm misremembering the comment phrasing and b) they explicitly banned that input from generating that output. [1] In the EU, this practice is explicitly legal
- RupertEisenhart 4y ago
  So, this makes sense from the perspective of writing code professionally for a company. I don't buy that the risk is significant for an individual programmer, but I suppose it is plausible that lawsuits could pose a threat to a large organization. But I find I get most pleasure out of using CoPilot for personal tasks, everyday stuff that isn't for work. Surely there is no issue there? And to counter the note about the scraping sources-- I have recently (out of work) been experimenting with using CoPilot to translate open-source MATLAB code into python and zig. If I make something useful in this way, it will be open source and I will credit those who wrote the original code. Surely a use-case like that last would have to tick all the moral and legal boxes?

kmeisthax 4y ago

ML text and image generators have a habit of regurgitating training data, especially if you happen to use certain inputs or prompts that also match said data even superficially. For example, type in "fast inverse square root"[0] and get out a bunch of Quake source code. If you use code or art generated by an AI that was regurgitating training data, you can be sued for copyright infringement. The way that AI gets training data these days is... questionably ethical. It's all scraped off the web, because the people who make these AIs saw court precedent for things like Google Books being fair use and assumed it would apply to data mining[1]. Problem is, that does nothing for the people actually using the AI to generate what they thought were novel code sequences or images, because fair use is not transitive. [0] This won't work in current Copilot because a) I'm misremembering the comment phrasing and b) they explicitly banned that input from generating that output. [1] In the EU, this practice is explicitly legal

RupertEisenhart 4y ago
So, this makes sense from the perspective of writing code professionally for a company. I don't buy that the risk is significant for an individual programmer, but I suppose it is plausible that lawsuits could pose a threat to a large organization. But I find I get most pleasure out of using CoPilot for personal tasks, everyday stuff that isn't for work. Surely there is no issue there? And to counter the note about the scraping sources-- I have recently (out of work) been experimenting with using CoPilot to translate open-source MATLAB code into python and zig. If I make something useful in this way, it will be open source and I will credit those who wrote the original code. Surely a use-case like that last would have to tick all the moral and legal boxes?