MIT Non-AI License

https://news.ycombinator.com/item?id=45529587

https://news.ycombinator.com/item?id=45533842

But can I give you another "viewpoint"? I guess it's like, "Wow, my code, my work, what came from my brain, my fingers" - it essentially lives forever, if you think about it - it becomes embedded and compressed inside weights/tokens. Like - part of you is there.

I guess it's cool. For me it's just to know that like this super intelligent things deep down actually knows who I am - my code is in it's architecture, it gives me feeling of honor in some way. You know?

Just my take.

Related, would you be in violation if you hosted this in a public GitHub repo, since it's in the TOS that they use source for training AI?

- VC funded SF startups with an ethos of "Do whatever is necessary at first, scale then lawyer up once a unicorn and shits starts to hit the fan" this might try to prevent... and who will blissfully ignore it

- actually mindful actors, e.g public researchers, non-profit, etc who genuinely understand the point and will respect it but whom, ironically enough, you might want to support by providing them with more resources including your code

So... yes the intent is good but honestly I wouldn't trust the 1st category of actors to respect that. Technically what they "should" do, because it's safe and efficient, is only crawl the most popular compatible licenses, e.g. MIT, and ignore the rest. That's safe and pragmatic. Again what I expect them to do (just my hunch) is take EVERYTHING, get called on, apologize, change nothing, get brought to court, apologize again and only do something if the financial repercussion is bigger than the expected alternative. So probably still do nothing.

If you want to release your code as actual open source but legally restrict what AI companies do with it, use a copyleft license like the GPL. They could still use it for training, but the product of such training may itself fall under the GPL being a derivative work, and the AI corps don't want to touch that legal quagmire. (The GPL continues to be Stallman's brilliant legal aikido at work. Stallman is like that one guy from Final Fantasy Tactics Advance who gives you the power to manipulate the laws to work to your advantage.)

Honestly, it may be time to abandon open source as a concept. There are source-available strategies that cause less social harm than open source, including Coraline Ada Ehmke's "ethical source" and the Sentry project's "fair source" (https://fair.io) models that we can draw inspiration from.

Key words are:

  - permission is [...] granted
  - free of charge
  - without restriction
  - use, copy, …

> may not be used for the purposes of […]

The license contradicts itself.

> Don't we have to ask for permission before feeding someone's years of work into an AI?

That's the point of an OpenSource license, to give permission.

This kind of stuff makes me think very few people really understand what OpenSource is about. The very same people who will fallback to licenses such as the BSL as soon as people/companies will use the permissions that they gave, and then will complain that "no one wants to pay for the thing i did for free and nobody asked for".