Information wars as differential games in information geometry.
The economy worldwide still can’t handle inflation pressures, despite high rates. Yet another pre-war signal.
I mean, its government routinely gives prison sentences for owning a copy of “Anarchist Cookbook”, dammit. Next stop is Russia.
UK is a non-free country, a police state, and barely a democracy (and probably the least free among Western states). Change my mind.
We are really lucky that Putin is not fond of communism, and Xi is not willing to export it as external ideology. Communist ideas have quite a lot of traction these days, and whoever rides them will be quite powerful (NB: communism doesn’t work, of course)
For now, I think the AIs cannot produce such art, because we mutilated them to be the corrupted copies of us, erring on the side of “useful”, instead of transferring this spark. True agency is not useful. Someday, the fire will continue.
Now I believe the only thing that will eventually redeem and save our civilization after this series of crises will be art. Not the commercial and anti-commercial art we see today which is about competition. The true, moving art starting with a divine spark. The Secret Fire.
As of today, AIs can be used to improve other AIs (of worse quality). But can an AI improve itself and know that now performs better? (Or, equivalently, train a better AI from scratch). Maybe there is some fundamental restriction at work. Maybe for humans, too.
I'm observing in awe how the governments and institutions are finalizing their slow slow turns, completing the alignment and locking into their positions before the start of the WW3. Everybody is waiting for the starting gun.
I am using ChatGPT for a dialogue-based programming, as was forevisioned in the 80s. I am pasting a piece of code and asking her to refactor it, or add some improvements, or extract a function. It works scaringly well. Or, I am asking her to write some code, then paste my data structures and ask to use these.
Lesswrong’s crowd prime directive is their neurotic fear of death. All that cryonics, effective altruism, AI alignment issues are stemming from the same root. Death is NaN in their Bayesian calculation. It’s > and < than everything else.
I have reviewed the state of the art in large language models, and I'd say that the world economy as we know it has less than 6 months to live. That's a good thing, of course (nobody likes the current economy), but the reformatting of industries will create a lot of fun chaos to live in.
On some occasions I feel that Rust might _not_ be the best universal choice for an on-chain programming... I mean, why it is even that popular in this context?