Bruce Schneier wrote an interesting post in June about rethinking Democracy in the age of AI.
I was reading it and thinking "Bruce" what you're writing about is Anarchy.
https://www.schneier.com/crypto-gram/archives/2024/0715.html#cg2
I suspect in the US and other places anarchy is seen as dangerous and chaotic.
Years ago when I did the political compass it called me a socio libertarian.
Now I'm aware of the issues with the political compass. My point is, that in the Anglosphere, privileged folks don't really hear that much about anarchy.
I didn't realise I was an anarchist until my 40s .
So I'm not surprised that in that post, Schneier doesn't mention anarchy. But to me anarchy fits the idea he is communicating.
@onepict He says things that are compatible with anarchism, but it's clear from reading this that even if he's decided that democracy as practiced is no longer working, his solution is not remotely anarchism; he appears to be thinking along the lines of techno-socialism. Which absolutely WILL become global AI-powered dictatorship.
@freakazoid Do you think that's because of his background?
It's not the first we've seen technologist of his background have ideas that are incompatible with our ideals of democracy and community?
I remember the fight against the potential sale of dot org. I remember being very disappointed with some of the internet pioneers.
@onepict His background and his privilege. My one constant disappointment about Schneier has been that he's always reached for either government or technology as the solution to problems. And here he appears to be losing faith not in government or technology, but in democracy. I'm hard pressed to discern a difference from Peter Thiel's "I no longer believe that freedom and democracy are compatible."
@freakazoid Sometimes I get the sense that some of that generation of Pioneers will jump on the hype/bandwagon to remain relevant. But much like with AI the bias gets in the way.
Reading that essay I kept thinking, it's Anarchy that's the answer. The solution is community and folks not technology.
Technology limits people's imagination especially AI, our biases lock us in to our norms.
But then I suppose to an American privileged technologist that's very comforting.
@freakazoid With that particular level of privilege, anything outside that norm is something to be controlled.
Plus I suppose as you get older there's more of an obsession with legacy, and to try to remain relevant.
@freakazoid I remember back in 2018 thinking that I'd like to meet Tim Berners Lee.
But then Solid got launched and there was no desire at that time for the Dev team to engage with folks on the fediverse. That along with the dot org shenanigans rather changed my thinking.
We can't create the next generation internet without engaging with everyone who wants to use it. You can innovate in an ivory tower, but it will be limited.
@freakazoid The answer is solidarity with your fellow human.
But there's a desire to avoid uncomfortable situations. Hence the surveillance to make everyone behave.
Building Bentham's Panopticon, but much like Bentham's contemporaries not taking the next part with full transparency in government, and things like what are their LLMs being trained with.
For example the OSIs definition of Open AI is limited because the openness doesn't extend to the training sets.
@onepict @freakazoid Access to training data is such a game changer. Especially in small LLMs it means I can tune my prompts and the way U use the llm to stuff that it's already encountered. Especially handy for stuff like tool calling where the names of functions are important
@freakazoid @onepict Yeah sadly fine tuning is the only type I have had acces to 🥲
@mauve Of course, with tool calling you're talking about just the finetuning/alignment data, not the pretraining data. The pretraining data is huuuuuuge, but access could be provided via Bittorrent or whatever.
Of course, that would require that they actually had rights to the data they trained on, which we know they didn't.
@onepict