What do you guys think?

I personally think it's very good, for something free. Yes, I know the app is Chinese and does censor certain sensitive topics, but if you run it locally on your machine you can strip it off of it's censorship so that's not too big of an issue.

I mean, being able to make Nvidia's and other big tech companies' stock to slump is impressive, and the thing is FREE and OPEN SOURCE.

I swear, if the geniuses in China ever pulls a DeepSeek again but this time on Autodesk and Adobe, I'm going to finally learn Mandarin and sing that one song.

it exists.

I ain't trusing no trumped-up transformer to do-- whatever hell the usecases these AIs are purported to have. up to this point I've been thoroughly unconvinced as to the utility of any language model beyond its function as a trivial novelty, and even in this, its adequacy has decayed as the systems have "improved".

if this new thing uses less power and requires less impressive tech then, at the very least, it might be less of a drain on the world's resources.

Autodesk and Adobe already have plenty of FOSS alternatives innit?

I'm just grateful that DS got Sam Altman's knickers in a twist, the "open"AI business model became completely unsustainable overnight

gattonero2001 said:
I'm just grateful that DS got Sam Altman's knickers in a twist, the "open"AI business model became completely unsustainable overnight

also this too; bitta schadenfreude once in a while's good for the soul.

dba_afish said:
I ain't trusing no trumped-up transformer to do-- whatever hell the usecases these AIs are purported to have.

For now, it's use cases is a fancy search engine, a very fast but buggy generator of programming code, a fake image/video/art generator, and a generator of low-quality content to farm views, etc.

In the future, if the perhaps hyperbolic claims of AI-related content on my YouTube feed are to be believed, it will become smart enough to take over everybody's jobs (especially tech jobs, perhaps ironically including the jobs of the people making the thing) and drastically speed up medical research. That's why there's a "space race"/"Manhattan project" like race to be the first to make "AGI" (artificial general intelligence = human-like intelligence) or "ASI" (artificial super intelligence = smarter than humans), and be the one in control of such a powerful tool. How do we know if we've achieved AGI or ASI? Great question... I guess whenever the tech companies decide to declare their AIs passed some arbitrary AGI or ASI test that they made up...

kora_viridian said:
[...]

When the Windows re-install was done, I went to tell my boss. I also held up one of the "King of Programs" CDs and asked about it. He laughed and said, "Oh, yeah, I got those the last time I was back in $MIDDLE_EASTERN_COUNTRY. You can buy them at the market; I paid the equivalent of US$20 for 4 CDs full of stuff." "Are any of the programs legit?" "Ha, no, they're all cracked, so they don't need a license key or anything. The way I use them, if there's some software that costs $200 or $500 or $1000, I go see the King. If he's got the software, I install his version and use it for a couple of days. Sometimes I decide it doesn't do what I want, so I try some other software. If the King's version does do what I want, then I go to the real company and buy a real license."

Heh, same thing here in my country. "You want an expensive program but don't have the money to buy/subscribe? Go to that one computer guy in the corner, he'll download it for you for 1/100 the original price!". Difference is, if the pirated software you use works for you, you dont buy the original license, you just continue using the software until it randomly stops working.

Because remember: "It's always morally right to pirate Adobe products."

But no, what I want the Chinese geniuses to do is make their own software that is (slightly) better and release it for cheap or even free, like what they did with DeepSeek.

gattonero2001 said:

I'm just grateful that DS got Sam Altman's knickers in a twist, the "open"AI business model became completely unsustainable overnight

Only thing I really care about tbh
Don't care other than that. I don't do AI stuff

I do not seek deeply

Does anybody think this could be a pun on deep six? :\

Deepseek is Great! Trust Me, Bro
Deepseek might have gotten most of its chips through illegal smuggling
DeepSeek Is a Century’s Scam! Chinese Bloggers Unite to Expose

If it makes the AI bubble burst now that they forever shattered the "you need 27 4090ti to get the wrong instructions on how to open your PC," I'm for it.
Especially if it makes the ACTUALLY useful scientific, overly specific LLM models better and kills off the generative sludge that has no use