Fatiguing big-time on the "AI" gold rush.
It's quick and can seem miraculous (and entertaining, like anything that's got to put on show because it's lacking in real thinking) but yet to show me it has more value than a regular search on Google.
Some self gathering/comparison/curation of on Google and comes no where near reading primary resources. it may take more time but I would bet money it produces a deeper more intrinsic and valuable understanding of a subject.
1
So decide what you want to be, a GPT prompter (in other words an Language modelling handler) or a person who understands the real thing. The thing GPT is designed to suck the marrow out of without giving any attribution or credit.
You can be both though, I guess. But still GPT cannot be sucking up so much of everything and presenting it as it's own magical thinking, anymore than a vacuum cleaner sucking up dust can claim to have alchemical God-like abilities to create dust from nothing.
2
Now let's get to the creative writing/art/lyrics/music/film/video BS that's supposed to be so mind-blowingly "REAL".
Nope. I MEAN NO FUCKING WAY is it any of those things and incorporate the depth, emotion, reality, truth, spirit of real person who's put time, sweat, tears into getting to a certain place with those things and finally achieves something fucking original and amazing that lifts everyone up to a better place, opens up new breathtaking possibilities....
3
....new worlds and adds to what's possible to share and communicate. It's a net gain. A net positive...a thing of the mind/heart/spirit made a real thing...
When I experience/hear/see/read any of that LLM stuff:
IT NEVER LOOKS/SOUNDS/FEELS/READS TRULY REAL. IN fact most of it is horrible and warped in someway that feels soulless and a laughably superficial facsimile that is bad. It's beyond bad. Most bad creative ventures can hit the most basic of the things needed.
"AI" isn't even that..
4
Or anywhere near that.
I'm yet to be fooled by any of it like this NY Times article is stupidly trying to insist is possible:
https://www.nytimes.com/2023/04/08/business/media/ai-generated-images.html
Yet, I guessed everything that was fake easily.
And they more I look at fake-generated stuff by this utterly over-hyped bs crop of "AI' tools, the more I can tell it's fake and the more grotesque and annoying it is....
Go figure, humans can quickly LEARN with their INTELLIGENCE to distinguish FALSE INTELLIGENCE ....
5
Artificial (non)intelligence AKA "Generative" (i.e., derivative, regressive, destructive) AI is to truth what a troll farm and psy-ops is to truth.
Meaning it's not the truth at all, it's a feint at truth, that's FRAMED as truth with all the superficial aspects of a true idea or true fact, but once you remove the frame, the context, the over-hyped, overwrought corrupt internal self-reference in service to an agenda, breaks the fuck apart into a thousand sad pieces that crushes the spirit..
6
It does not lift up, it's just something sad and fake that's hollow and diminishes us all.
The crop of AI services is a pathetic gold-rush. There's so many that depend on some gimmicky con-artistry.
I think it can probably do good things in crunching medical data and hard mathematical data and arriving to solutions that thus far were to labor intensive to get valauble data from, but what is that really???
That's more easily harnessed processing power, that's it and that's all.
6
So what we may have with LLM is an interface, an interpreter, somewhere between auto-complete and Google translate hooked up to vast data and super-computer processing power.
It's not a stand alone entity. To really work well and really shine in what it produces or have what it produces be truly valuable, people, like with any technology have to do that and pick up all the missing pieces.
Like any tool.
But ultimately It's an outgrowth of the Internet and the horrid dishonesty and abuse..
7
@TheAbbotTrithemius Hey but it's riven with bias and nearly-slave labor was used to review the source data for ptsd-inducing violence and sexual abuse.
Wait those aren't good arguments