Fire in the hole!
It’s a little more than just the genAI part tho.
You can fuck the holograms. If real world genAI was fuckable, FuckAI would not be an anti-AI community; it would be a NSFW one.
Where does your jizz go?
Into the biofilters until some poor lower deck person empties it:
Into the hussy
Artificial intelligence in science fiction is intelligence that’s artificial.
Artificial intelligence in real life is more akin to what we’d call “artificial flavors” in some fruit drink that contains 0% fruit. Isn’t fruit, doesn’t taste like fruit, but it’s purple and sweet, so fuck it: “grape”. We’ve accomplished the “artificial” part, but not the bit that actually matters.
So, yeah, AI in sci-fi is unironically great. “AI” irl is shit. If the real version ever comes close to the sci-fi version, you’ll see a lot of the haters change our tune, but until then we’ll call a turd a turd.
OK, but how will we ever get to “great” AI if we don’t invest in less great AI to drive research. If the luddites on Lemmy calls the shots, Data will never see the light of day.
Well, rolling with the artificial flavors analogy, we can start by stopping the infuriating trend of adding artificial grape flavor to every fucking food item under the sun. No one asked for or wanted that shit, but now it’s become a chore to find so much as a bag of coffee beans that doesn’t have some blip on the packaging advertising “Now with grape!!” Not only does it taste bad, but it’s caused the cost of food to fucking double.
Second, we can start being honest about what it actually is. Again, it doesn’t taste like grape, so why the fuck are we calling it grape? With AI this is as much on us as it is with the companies marketing that garbage - for ages we’ve been calling so much as an NPC’s pathing algorithm in a videogame “AI” even though there’s objectively no intelligence at play. We’ve since tried to move the goal posts a bit by relabeling actual AI as “AGI” but I guarantee that’ll just become another term to muddy the waters with marketing bullshit. The shit we’re calling “AI” is just categorically not even on the same path as the the kind of AI we see in sci-fi. So if you want Data to see the light of day, we’ll need to stop wasting our time and money on what is effectively a spell checker that’s wrong occasionally.
Unless you think Data is just a chat bot with no consciousness, in which case, hooray, we’ve done it: AI. 🎉 …now can I please have a mug of coffee without Grape™ syrup in it? I know, such a Luddite, drinking coffee without grape! Guilty as charged. But adding that shit to my coffee is never going to get us a product that tastes and feels etc like a grape, so if that’s what you’re after, maybe keep it in R&D and off the fucking grocery shelves.
“research” doesn’t need billions of dollars in commercial data centres. The shitty “ai” we have is all about line must go up.
“Make an opponent capable of defeating Data”
ST AI: Have a supercharged Moriarty.
RL AI: deletes production database
If we got all the other socialist progressive benefits of the Star Trek universe, including no billionaires, then I think we’d be a lot less unhappy with generative AI as presented in that world.
How they built it, and what they plan to do with it is a pretty big part of the reason for our anger.
Exactly! I am not anti-ai. I am opposed to corporate controlled, massively polluting, unreliable slop engines that will be used to surveil us and take our jobs without benefitting us in any way.
Yeah, if we de-meme it and really want to define it, the difference is Star Trek is a post-scarcity world, so generative AI building things doesn’t take anything away from human artists - it doesn’t affect their livelihoods because their livelihoods are guaranteed.
As always…
Plus I think it helps that much of what the holodeck recreates is centuries old at that point. 21st century earth, long ago. It’s not anyone’s current creations, and then beyond that it’s using data that the federation personally got by exploration. Like if people only used AI to generate medieval drawings I don’t think anyone would be too upset really. Maybe weirded out but not upset. Aside from the environmental concerns I guess.
however generating holo characters that look like your coworkers and doing weird stuff to them is a entirely canon issue. And Star Trek never touched on it, but wow would it be easy to fake footage of historical events in a holodeck, especially to civilizations without advanced technology. I work in history as my career and the amount of YouTube videos telling fake stories and showing fake footage of 20th century history is genuinely terrifying. I don’t know what the solution is. Digitization was meant to preserve historical footage but now it just makes it look just like the fake stuff. If anyone digitized historical footage and the physical no longer exists, verifying anything will become a nightmare and sometimes it would be strictly possible.
Honestly the more I think about the holodeck the more I think it’s dangerous as fuck, literally and socially, and Star Trek shows usually only very innocence uses for it. Like you thought parasocial relationships are bad now?
The Star Trek universe also seems to be able to run their AI without boiling the oceans.
Still requires boiling oceans, but they’ve found a few uninhabited water planets that can safely be boiled without bothering anybody.
(Also, our AI isn’t even that good. They’re not boiling the oceans. They’re boiling what previous little fresh water we have that could be used for drinking or irrigation. It would be a lot less of a problem if they just built their cooling systems to deal with salt water and used seawater instead.)
Depending on the rate of AI we will probably get something to the polity universe version than Star Trek.
Still not as bad but could be better
EXACTLY
If they didn’t invest our entire country and ecosystem into it because they thought it meant digital slaves that could replace the meat popsicles they abhor, then it might not have looked so fucking evil
Holodecks are discrete programs with subroutines.
Perhaps we give it a pass in a post-scarcity world? When capitalism isn’t in the picture, it’s not taking jobs or ruining the environment or anything like that.
Although, you kinda still have to “write” a holodeck program and it also seems to be slightly harder than prompting or they’d be doing it on the fly a bit more.
-
Holodecks are generative AI that actually works.
-
All LLMs are Generative AI, but not all Generative AIs are LLMs.
Holodecks are generative AI that actually works.
They work until the plot dictates that they don’t. Which happens a lot, actually.
Doesn’t happen a lot.
Only they don’t make episodes of the 99.99% of times they work perfectly.
If .01% of the time it creates a problem that kills someone or endangers the entire ship that seems like a lot to me for something that is used constantly by the entire crew.
True.
Fortunately the .01% of the time there’s an error, its incredibly minor.
-
How many times did a bad “prompt” nearly destroy the Enterprise? Kill a crew member? Commit sexual harassment, for lack of a better term?
Or create a fully sentient and conscious mind just because you asked it to? Jesus, holodecks are scary.
It’s part of the charm <3
Especially in “Schisms” (S6E5) where they build the operating table by trial and error… 100% feels like a modern ai imagegen workflow (only 3d and tangible).
Now this might have something to do with what they mean when they disable safeguards. Maybe there’s a layer of user-defined limits that are carefully defined and that’s part of the craft of a good holodeck program, but when safeguards are disabled, it just runs as a raw universe simulation, matter, photons and all.
Commander Maddox had a point. If Data was a box on wheels they would have shipped him off to be taken apart. But because he looked and sounded like a human everyone empathized with him.
They seemed to empathize with the exocomps after realizing they had sentience. They were a little more complex than box on wheels, but a far cry from humanoid.
Granted, that happened well after that landmark court case about Data’s sentience.
And only after Data advocated for them. The organics themselves wouldn’t have cared.
But they’re so cuuuuuute 😍 Look at them!
laughs in Peanut Hamper
It depends, what kind of tackle is that box on wheels working with?
I never got the impression that the holodeck (or Data) worked at all like an LLM.
It doesn’t. An LLM uses modeling to approximate the shape of a question and generate a novel answer within the parameters. It’s like a shortcut to actually knowing everything and providing authoritative answers.
Sort of like the Scrabble champion who won in Spanish and French despite not knowing those languages. He simply learned patterns of words he could build with letters, but he didn’t know the meaning of any of the words.
The holodeck ostensibly has access to all of federation knowledge and history, and when it has to generate novel information (like what would Einstein say to Newton while playing poker) then it extrapolates that informtation based on the sum total of existing information. The holodeck computer has a thorough understanding of the question, the parameters, and the answers (holograms) it creates. It isn’t abstracting the question or approximating the shape of the answer.
You are giving writers way too much credit. They wanted to write stories around human characters. The AI is as advanced as story needs it to be. Dimming lights, yes, the SotA flagship AI is allowed to do it. But split second combat tactics? Only if AI is inside the humanoid character. Shouldn’t overthink it, the narratives would be lame otherwise.
Who mentioned LLMs?
Replace LLM with any of the current types of generative AI
It, if it is imagined as using a similar underlying technology, would also be an example of that technology that is both far more mature and understood (at least until the writers want to use it as an excuse for whatever shenanigans of the day) than current AI tech, and more capable, and also not shoved into everything in an attempt to force people to use it so that it’s owners can extract more of everyone else’s shrinking supply of money, and also not part of an economic house of cards that people are told will eventually impoverish masses of people when it inevitability collapses.
It’s like saying that in the 21st century we have decorative plants and everyone loves those, to a dutch person from the 1630s who has complaints about the tulip industry.
Wrong, the holodeck follow instructions, the
AIneural simulation don’t.Unless you like to play in simulation created by a simulation, coded with an abstract language, run into a virtual machine.
At some point in time, humanity will learn.It’s not generative AI we hate, it’s LLMs.
LLMs are generative AI.
Yes but not all generative AI is LLMs. I hate wine, but saying “you know beer is alcohol too” is not something that should make me feel hypocritical about liking beer. It’s not alcohol I hate, it’s wine.
