so much of what we’ve called growth has come from blowing up what was once free and ordinary, then restoring a poor facsimile thereof for a price. we can talk to our families over zoom. what a novelty.
probably not. serves me right for reacting to a stupid x post. x.com/michaelaarou...
claims like Mississippi is richer than France because GDP per capita aren’t the zinger you think. they just (correctly!) undermine the notion that GDP per capita is a good measure of prosperity.
how does the electrical power utilization of generative AI compare with that of bitcoin?
(i'd noticed sometimes if i accidentally miss a space after a period it adds a url card and they stick around even when i fix the spacing if i forget to x them out. using them for attributions is a great hack!)
we visited the summer before last ofc can't afford to stay in central Paris stayed in Colombes, hotel had a great rooftop terrace surrounded by new apt buildings rising like beautiful mushrooms we could fantasize about living in and take the same easy train we took as tourists anywhere.
i guess the hope is that a special space would facilitate establishing a norm in a way the ability to reply-to-self thread does not. we'd become obnoxious with our "where are the attributions?" like we can be obnoxious about "where's the alt text?" 1/
a thing no microblogging platform offers, but all should: a hidden by default but click-to-revealable space for attributions of claims made in the post. just links, excluded from the character limit. many posts make factual claims that strike me either wrong or extraordinary, no way to tell which.
you’ve got to go back farther to get Google left coded. 20 years ago. but they were using precursors to modern machine learning even then to address spam! no one thinks there’s anything inherent in the tech that’s “right-coded”. it’s who wields it and to what perceived ends.
is DOGE going to be politically sustainable at this point under its proposed leadership?
i don’t disagree with this. it is a remarkably vague kind of claim.
it’s ultimately not the tech. it’s the degree to which applications are guided by, optimized toward profit maximization independently of other social goods that might once or in a better arranged society have been correlates of profitable (not necessarily maximally profitable) enterprises.
prediction of human behavior analytics, in order to better profit from such behavior.
i think to identify the social coding of the tech (a pretty vague idea!), you have to weight the applications behind the money behind them. where was the money focused during 2010s AI/classification boom? if you wanted to pitch a startup, was it gesture recognition or predictive analytics?
no. i mean, maybe it’s coming to that. but it wasn’t always thus. google was left-coded for a long time. 1/
it’s a recent development in a US context that identity as for-profit corporations maybe overwhelms the capacity of a collective to deliver social goods. we used to think in a mixed economy, for-profit cos could participate in ecosystems of overall flourishing. 2/
it’s amazing how a thing can become ever more contentious even while it becomes ever more boring.
the classifiers that preceded LLMs are also right-coded, they’re the “analytics” behind surveillance capitalism. Palantir is right-coded.
how many engineers at X will be diverted now to a project setting up a realtime somebody-else’s-voice filter for his sock puppets on spaces?