Re: the recent meta discussion and ongoing chats about self hosting, open vs closed source, AI etc, I wanted to share some food for thought.
I’ll explain why this is related to self hosting at the bottom (section bolded): our lovely new mods can call it. I hope it inspires some out loud thinking.
Disclosure: I am not the content creator nor am I paid by them to signal boost. I just like their stuff and think this is an important topic, from multiple angles.
“The future of AI depends on the moral compass of five people.”
I’ve been watching “AI in context” for a few weeks (they make long form biopic content on current state of AI - really good stuff).
This dropped today; it’s about the wheeling and dealing behind closed doors at OpenAI re: Sam Altman’s firing. It’s a lot more watchable than that sounds :)
https://www.youtube.com/watch?v=_eYTkvZqbnQ
The line that brought me to a stand still was “the future of AI depends on the moral compass of like 5 people”.
I think folks here (and Lemmy generally) are more savvy about AI then the gen pop (though Lemmy is famously FuckAI)…but even if you’re training a nanoGPT model from scratch on hardware you own …you’re still beholden to outside forces.
Eg: the people’s champion - Qwen - seems to have split or gone closed weights for 3.7. That’s not a good sign.
Reason for post
In the recent [Meta] chat, I noted an undercurrent of “no man is an island” - that is, yes, you might host your own X, but you’re still dependant on external Y (eg: SearXNG).
Self hosting / FOSS / forking mitigates some of the “we changed the terms of service after the sale” enshittification we see occurring in related spaces (eg: right to repair). But there’s only so much leverage you can enact before it becomes pyrrhic.
I would like to believe “fuck you, I won’t do what you told me” is our bulwark against market forces.
At the same time, it’s sad to see so many “Don’t be Evil” mission statements not survive contact with reality (watch the vid: OAi was founded on the ideal of “don’t let AGI kill us”)
I think what happens upstream has effects down stream too (see prior X vs Y examples)…
Not sure where this leaves us. It’s a weird time to be alive.
Enjoy the video (and their others - Ai2027 is eye opening). Look forward to any productive chat this post might inspire.
im very skeptical of this channel just because of how much it anthropomorphizes AI, it feels propagandistic.
Instead of taking the statistical analysis approach, its easy to rely on the marketing of “AI scary and will kill us and we’re not ready” or whatever, like, scifi is cool, but this is not that.
I like this especific explanation on an AI responce to a PR (the whole thread is good in that sense), as to why it says what it says, its a boring and statistical framework, shaped by rules that turn it into a complex machinne.
We as people like to anthropomorphize other things, read into them and see our own reflections, and altho its not bad to have empathy, but thats not always reality, and we need to be weary of how this behaviour can be exploited.
This whole youtube channel is personally off to me, mostly because of the super professional look, how out of the blue it appeared, not knowing anyone from the team that researches any of these topics (maybe i just havent looked into the credits) and how their weariness of AI is the same as the weariness of the companies, which have the interest of selling it (see claudes secret model thats a danger to the world and must not be published).
With the amout of money in AI, i worry its a psyop from some company that want to promote the idea of AGI, so that AI business stays viable for that little bit longer.
End of rant, i hope it made some sense lmao. Fuck IA