Icon by @ebu


friend enjoyer and aspiring game developer

aromantic/bisexual

25 years of being chaotic and counting

send pats!!


i also do custom magic item commissions

for ttrpgs and such


frequently NSFW, sometimes I write porn and sometimes I draw it of myself
๐Ÿคท


ฮ˜ฮ” am creature (dragon edition)


I enjoy doing worldbuilding a lot and have a big sci-fi setting with magic I've been building for over a decade now


If you want to know more places to find me,
ask me on discord! I don't bite (unless you want me to :3)

jnnnn
@jnnnn
This page's posts are visible only to users who are logged in.

ebu
@ebu

this is good, but there is one main critique i would launch at this video. "ai" proponents believe that their technology is going to usher in whatever techbro utopian revolution, and i think McGee explains very well that the world they are describing is a morally-bankrupt dystopia. but i think the error that both make is assuming that this world even can be real; that "ai" is going to change that much at all.

(i wrap "ai" in scare quotes here because "ai" is a marketing term. it has no formal definition by design, because any such definition might get in the way of the sales pitch.)


firstly: we already live on the planet where computers decide who gets healthcare, loans, and jobs. a sizeable chunk of "ai" being marketed toward or used for these purposes is a fancy coat of paint over the already well-established statistical systems that automate decisions towards the goal of profit. this is an incremental change at best, and at worst, this is swapping out some of those systems with demonstrably worse versions. ("worse", as decided by corporate interests here. think less "this is bad for humanity" and more "our new chatbot called a customer a slur")

secondly: a lot of output of the genuinely-novel "ai" systems is... put bluntly, not good. anyone who's tried or been asked to, say, generate images or text knows there is a ton of garbage that has to be edited, remade, or thrown out. the stuff that survives to go viral on twitter or make it to promotional material is the stuff that survived the cutting room floor, and even then, those results can vary wildly in quality.

thirdly: i believe that most of the tasks that "ai" is fit-for-purpose for is Graeberian bullshit. since people can't trust the output of the machine to be of quality, the only places where it is applicable are the places where quality does not matter. stock images, generic copywriting, background art, lorem ipsum, boilerplate code, listicles, spam.

i think sometimes the "ai" advertising is so strong we believe it even as we despise it. but i think enough "ai" has gotten into the hands of enough people to evaluate it and determine that it really doesn't live up to the hype. proponents might hand-wave away flaws and say "it will get better", but that's just another narrative, another sales pitch. hell, when was the last time an update didn't degrade something? even the people using ChatGPT and other text models the most notice and complain when updates make the technology worse to interact with. (the "update" narrative also makes little sense. they already scraped the internet for images and text -- there isn't an internet 2 to scrape for "ai 2". well, besides that internet 2)

however i think the broad conclusion of the video is correct. i think people in general will begin to psychologically reject "ai" the way they do NFT's and cryptocurrency nowadays and once that potential rejection starts to appear on corporate cost-benefit analyses, they will drop "ai" like a rock.


also a minor nitpick: i very strongly disagree with the notion (as presented in the video) that blockchains have genuine use cases. every practical problem that blockchains attempt to solve (single source of truth, immutable history, decentralization) are already solved by existing database engines (quorums, append-only databases, replication), and don't bring on the loads of problems that blockchain does, like the massive compute cost, ludicrously slow transaction times, and overall fragility. if it were useful to track physical goods, amazon would use it. (they don't, but they are happy to sell you the footgun, see Amazon Managed Blockchain)


You must log in to comment.