hi there!
I've been a fish user for some time now, and I've really been enjoying it. It's a great shell and scripting language and I've preferred it over POSIX shells and bash for a while.
I've noticed that this file was added: https://github.com/fish-shell/fish-shell/blob/master/AGENTS.md
I want to urge you to revert the addition of this file and reconsider the use of generative AI.
It seems useful at a glance, providing better autocomplete or suggesting whole snippets or completing PRs, but all of these come with very real costs to the real world outside of the software. Environmental, social, and ultimately also in the field of technology.
To cut it short, code is being used against licenses1, datacenters are polluting communities and driving up costs, and these tools are currently being used to drive an actual, real-life war2.
I urge you to please reconsider. It serves the good of everyone to reject these tools despite their benefits on the grounds of our shared responsibility.
hi there!
I've been a fish user for some time now, and I've really been enjoying it. It's a great shell and scripting language and I've preferred it over POSIX shells and bash for a while.
I've noticed that this file was added: https://github.com/fish-shell/fish-shell/blob/master/AGENTS.md
I want to urge you to revert the addition of this file and reconsider the use of generative AI.
It seems useful at a glance, providing better autocomplete or suggesting whole snippets or completing PRs, but all of these come with very real costs to the real world outside of the software. Environmental, social, and ultimately also in the field of technology.
To cut it short, code is being used against licenses1, datacenters are polluting communities and driving up costs, and these tools are currently being used to drive an actual, real-life war2.
I urge you to please reconsider. It serves the good of everyone to reject these tools despite their benefits on the grounds of our shared responsibility.
Footnotes
because the training data contains code of an incompatible license, the outputted code may be considered tainted by some courts. This would risk making the Fish codebase entirely Public-Domain as a result in some jurisdictions, should fish ever get into legal trouble. ↩
these companies may deny it, but we've seen the deals between OpenAI and the "Department of War" play out. ↩