- cross-posted to:
- opensource@lemmy.ml
- programming@lemmy.ml
- aneurysmposting@sopuli.xyz
- cross-posted to:
- opensource@lemmy.ml
- programming@lemmy.ml
- aneurysmposting@sopuli.xyz
With the OSI publishing their abysmal - explicitly not open source - “Open Source AI” definition I thought I’d post my argument, why it is bad and why “Open Source AI” currently probably does not exist.
A pretty solid piece on how AI is closed-source by nature, and a solid takedown on the OSI’s FOMO-fuelled dumpster fire of an Open Source AI definition.
I’ve also thought a bit about AI’s relationship with open-source as well - to expand my views a bit, I view AI as having a hostile relationship with open source, stealing whatever it wants and damaging open-source projects when it quote-unquote “gives back”, and I suspect that we will see a severe decline in the FOSS ecosystem because of it.
With AI bros treating “publicly available” to mean “theirs to steal” (sometimes openly saying it, oftentimes suggesting it with their actions) and more-or-less getting away with it for the past two years, people have been given plenty of reason to view FOSS licenses (Creative Commons, GPL, etcetera) as not worth the .txt files they’re written in, and contributing to it as asking to have their code stolen.
The recently-released Stallman Report (which you mentioned) definitely isn’t helping FOSS either - all the diversity initiatives and codes of conduct in the world can’t protect against a PR nightmare on the magnitude of “your movement’s unofficial face becomes the Jeffery Epstein of coding”.
Baldur Bjarnason’s also talked about open-source’s rocky financial future - I’d recommend checking it out.