joele wrote: ↑Fri Jan 31, 2025 9:32 am
This whole thing is driving me a little mad, we get article over article about how china is winning the AI race.
Don't forget, these articles were written by jounalists, not scientists. All Deepseek did was basic science. Build on the work of others within whatever constraints you have. Ironically, the ban on chips might have forced them to focus on efficiency over brute force, thus building something that is better than what we had before.
The winners are us (at least those of us that like AI! who get free/cheap high quality "open source" models to use. And maybe the big western tech companies who might increase their focus on efficiency rather than wasting a trillion dollars on hardware.
Anyway, 'winning the AI race" is just media clickbait. Don't let them control you. Instead, go use Deepseek and let the Chinese control you, thus letting them win the AI race!
Re: AI Uses
Posted: Fri Feb 07, 2025 3:49 pm
by stylofone
Over on the dashcam thread the strangely compelling video of a very large number of people ruining their cars by driving through a famous floodway in a quaint corner of little Britain prompted me to search for some general knowledge on "hydrolocking", the mechanical problem caused by water getting into the bits of your car engine where no water should be.
The explanation on autochimps.com seemed pretty good, but it had a bland textual sheen to it which made me investigate it further, and yes, it is an AI slop site. The funny thing about it is that its creators have gone the extra parsec to make it look real, claiming that it is written by real people, fact-checked by other real people, and inviting you to look at their social media profiles and other sites supposedly verifying their realness. Unfortunately the "verification" pages also turn out to be AI slop sites.
This highlights one of the key traits of AI. In so many circumstances, its success depends to a very great extent on deception (or dishonesty, fraud, lying). At this early stage it is actually quite primitive and fairly easy to detect. But there are plenty of people who didn't bother to pay attention in Year 11 English class who would be fooled even by these flimsy deceptions. Also, rather than admit that they had been fooled, people are inclined to blame the fact -checker and insist that the site was genuine and the fact-checking is woke, arrogant politically correct etc. etc. etc.
This is why I see AI and LLMs as an extension of the tools of behavioural manipulation created by surveillance capitalism. social media and engagement algorithms. It's all downhill to dystopia from here, my fellow humans.
If you hate the nightmare world of big tech and its latest obsession, AI, and you don't mind flowery, long-winded writing, there is a fair bit in this article.
The fantasy and utility of AI, for the unconscionably wealthy and relentlessly wary masters of this space, converge in a high and lonesome abstraction—technology designed less to do every human thing for you than to replace all those human things with itself, and then sell that function back to you as a monthly subscription. This device will play with and talk to your child; this furry mouthless robot with enormous attentive eyes will replace your pet; your coffee is ready and your clothes for the day have been picked out for you.
Well, uses for AI? How about feeding all the raw data in from the department of education, including probably SS numbers, credit card details, addresses and other personal information, probably medical as well, to try and track down government fraud, waste and DEI initiatives:
who get free/cheap high quality "open source" models to use. And maybe the big western tech companies who might increase their focus on efficiency rather than wasting a trillion dollars on hardware.