i can admit it's possible i'm being overly cynical here and it is just sloppy journalism on Raffaele Huang/his editor/the WSJ's part. but i still think that it's a little suspect on the grounds that we have no idea how many times they had to restart training due to the model borking, other experiments and hidden costs, even before things like the necessary capex (which goes unmentioned in the original paper -- though they note using a 2048-GPU cluster of H800's that would put them down around $40m). i'm thinking in the mode of "the whitepaper exists to serve the company's bottom line"

btw announcing my new V7 model that i trained for the $0.26 i found on the street just to watch the stock markets burn

That's the opposite of what I'm saying. Deepseek is the one under scrutiny, yet they are the only one to publish source code and training procedures of their model.

this has absolutely fuck all to do with anything i've said in the slightest, but i guess you gotta toss in the talking points somewhere

e: it's also trivially disprovable, but i don't care if it's actually true, i only care about headlines negative about AI

"the media sucks at factchecking DeepSeek's claims" is... an interesting attempt at refuting the idea that DeepSeek's claims aren't entirely factual. beyond that, intentionally presenting true statements that lead to false impressions is a kind of dishonesty regardless. if you mean to argue that DeepSeek wasn't being underhanded at all and just very innocently presented their figures without proper context (that just so happened to spurn a media frenzy in their favor)... then i have a bridge to sell you.

besides that, OpenAI is very demonstrably pissing away at least that much money every time they add one to the number at the end of their slop generator

consider this paragraph from the Wall Street Journal:

DeepSeek said training one of its latest models cost $5.6 million, compared with the $100 million to $1 billion range cited last year by Dario Amodei, chief executive of the AI developer Anthropic, as the cost of building a model.

you're arguing to me that they technically didn't lie -- but it's pretty clear that some people walked away with a false impression of the cost of their product relative to their competitors' products, and they financially benefitted from people believing in this false impression.

i think you're missing the point that "Deepseek was made for only $6M" has been the trending headline for the past while, with the specific point of comparison being the massive costs of developing ChatGPT, Copilot, Gemini, et al.

to stretch your metaphor, it's like someone rolling up with their car, claiming it only costs $20 (unlike all the other cars that cost $20,000), when come to find out that number is just how much it costs to fill the gas tank up once

i would love to read someone more familiar with historical fascism and imperialism who is able to articulate the link between the two. like, an analysis of imperialism as the construction and maintenance of borders across which exploitation is enforced and economic value is extracted -- paired with the observation of fascism as the attempts to construct borders "across peoples" from within the imperial core

i wonder which endocrine systems are disrupted by not having your head sufficiently stuffed into a toilet before being old enough to type words into nazitter dot com

if you put this paragraph

Corporations institute barebones [crappy product] that [works terribly] because they can't be bothered to pay the [production workers] to actually [produce quality products] but when shit goes south they turn around and blame the [workers] for a bad product instead of admitting they cut corners.

and follow it up with "It's China Syndrome"... then it's pretty astonishingly clear it is meant in reference to the perceived dominant production ideology of specifically China and has nothing to do with nuclear reactors

A WELL TRAINED AI can be a very useful tool.

please do elaborate on exactly what kind of training turns the spam generator into a prescription-writer, or whatever other task that isn't generating spam

Edit: to add this is partly why AI gets a bad rap from folks on the outside looking it.

i'm pretty sure "normal" folks hate it because of all the crap it's unleashed upon the internet, and not just because they didn't use the most recent models off the "Hot" tab on HuggingFace

It's China Syndrome but instead of nuclear reactors it's AI.

what are we a bunch of ASIANS?!?!???

very nice :3

the post is fake regardless, its just another rightwing shitpost account. doubt there's a kid at all

f4mi's channel is fantastic btw, fun little deep dives on old hardware and games. highly recommend checking her stuff out

i went and bought it, and yup, the revisited version is the one i was thinking of. time to walk around inside a picture of Sam Altman so i can absorb his raw intellect and business acumen

goddammit! you have no idea how many variations of "first person walking simulator projected image texture trippy visuals" i slapped into every search engine!

but yes, that was the one i was thinking of

yeah, that "most of the internet will be Al-generated" nonsense is tanking my ability to take them as domain experts seriously.

still, something gets me about completely generated, transient-when-you're-not-looking, constantly shifting worlds. might have to collect more examples

maybe i'm a weirdo but i actually really like this a lot. if there weren't armies of sycophants chanting outside of all our collective windows about how AI is the future of gaming... if you look at this "game" as an art object unto itself i think it is actually really engaging

it reminds me of other "games" like Marian Kleineberg's Wave Function Collapse and Bananaft's Yedoma Globula. there's one other on the tip of my tongue where you uploaded an image and it constantly reprojected the image onto the walls of a first-person walking simulator, but i don't recall the name

because it encodes semantics.

if it really did so, performance wouldn't swing up or down when you change syntactic or symbolic elements of problems. the only information encoded is language-statistical

oh gods they're multiplying

"blame the person, not the tools" doesn't work when the tools' marketing team is explicitly touting said tool as a panacea for all problems. on the micro scale, sure, the wedding planner is at fault, but if you zoom out even a tiny bit it's pretty obvious what enabled them to fuck up for as long and as hard as they did