myrmepropagandist
AI wasn't the "next big thing" ... the market has missed it. Maybe it's not obviously profitable enough. But, there are ideas out there growing that will change the world and only the people close to them even know what they are.
It's not totally possible to just decide that something is going to "change the world." You can fake it for a bit, but eventually reality sets in.
That doesn't mean that nothing is happing. You just might need to listen in different places.
digital luciferase
@futurebird at the risk of dorksplaining a pedantic nonpoint, we really need to stop calling these LLMs AI. I agree that they aren't the next big thing. They also aren't AI.
the larger point i have in mind is that we aren't done with the AI phase, just one of its preliminaries.
Rich Felker
@resonancewright @futurebird They absolutely are "AI".
For over half a century, "AI" has been a marketing term for parlor tricks to bamboozle folks without technical expertise to know better that the computer is doing something vastly more impressive than what it's actually doing.
That is THE meaning of "AI".
Rich Felker
@resonancewright @futurebird I'm not going to try to "reclaim" a name that has always meant scams. I want no association with it. I'm not ceding anything to them by this. I'm refusing to participate in the farce that "artificial intelligence" has any meaning beyond scams.
digital luciferase
if we are what marketers say we are, we're doomed.
I myself do not think they are the arbiters of reality, but I won't stop you if you do. It is certain that people constantly market 'smart' featured things as AI, when not even LLM features actual intelligence.
I have to say instead that I don't really GAF what they make of things, though, or what gullible people decide they must mean.
I think one can be two different things at the same time -- entirely skeptical of the current crop of businesses and products that are hyping the 'intelligence' of what they are offering for sale, and simultaneously cognizant that a true autonomic general intelligence will be a very different beast altogether.
Rich Felker
@resonancewright @futurebird Beyond that, pretending the name "artificial intelligence" has any connection to reality is premised on the concept that "intelligence" is a tangible thing you can measure and replicate and is a legitimate basis on which to evaluate/compare the worth of beings.
Rich Felker
@resonancewright @futurebird Um, no. No actual scientist "defines" "AI" as "passing a Turing test". Turing did not propose it as such and was very careful to describe the intricacies of this misconception. It sounds like you have a pop-scifi understanding of this stuff and want to pick fights with people who don't subscribe to your techno-futurist fantasies.
digital luciferase
...not at all. the most common definition scientists apply to an artificial intelligence is one that is hopelessly subjective -- passing a Turing test. No one is making an 'if than' argument to substantiate ai, they are making the 'walk and talk like a duck' argument. Most know it can't be objectively defined by us, or at least, we haven't figured out how to do it yet.
The rest of your post seems quite vehement, like you want to fight about something; best of luck with those sentiments, whatever they ultimately are, once rendered noncombative and speaking in normal tones. If they are meant to convey a profound distrust of the technologies that people try to use to replace workers, you are wise -- but that alone won't stop them, nor will insisting that it's all the same hype thing.
Rich Felker
@futurebird @resonancewright For things like creating a simulation of a human (or other animal) mind, I think the more interesting, less eugenicist-adjacent concept is "artificial life".
myrmepropagandist
I think the point was just that "AI" has been used as a marketing term for almost a century, thought this most recent round has been notably annoying.
What is "real AI" ? Creating a simulation of a human mind?
I think the point about intelligence not being well-defined are valid.
digital luciferase
no true scotsman, you say? :)
I would remind you that you were the one who chose to define AI as how it has been marketed to the masses. I don't think it's particularly useful, myself, but I won't quibble.
The Turing test remains the standard even though many people recognize the problem with it , not because it is flawless but precisely because there is no more objective workable definition in play.
That's the point.
For someone who insists that AI means what marketers say it means, you are now curiously and mysteriously deaf as to what they say about the Turing Test. But at least you're harking back to what AI researchers are actually saying now instead of Madison Avenue.
Rich Felker
@resonancewright @futurebird LLMs have nothing to do with human minds. They are distillations of past usage of *language* only, into a statistical model that can predict language usage likely to be misinterpreted as intelligent/meaningful by humans. You are deep in the singulaity/AGI delusion.
digital luciferase
I'm sure you saw the bit about Dawkins spending time with an LLM and announcing that he thinks it's actual intelligence somehow. To me, that illustrates the danger of approaching it as something that resembles the simulation of a human mind.
LLMs ARE the distillation of the human mind in a way. A rotgut, prohibition era, bathtub-gin type of distillation but there you have it. I think it's fundamentally more likely that we would decide non-intelligent actors are intelligent because they ape that model, when an actual generated intellect would be.... whatever it is.
We made gods in our image, too. We tend to understand our world via human metaphor.
There's a scenario for AI where no one recognizes what it is until it cannot be stopped, because people disregarded it at first as it didn't fit the image they had for AI.
I myself have the hopelessly subjective notion that if artificial general intelligence arises it will be under no particular urge to clue us in that it exists :)
Rich Felker
@resonancewright @futurebird The Fermi paradox isn't a paradox at all, just scifi sophistry. Space is just utterly gigantic. This does not tell us anything about likely outcomes of civilizations.
I didn't assign you a label "out of mental frustration". I assigned it because you're giving off al lthe signs.
digital luciferase
@dalias @futurebird
no, that's just the niche you've assigned me out of mental frustration.
I don't believe in a benevolent techno-bro post-singularity future; i think it's much more likely that we'll all be dead in fifty years than that we'll be holding hands all watched over by machines of loving grace.
One does not need to believe that nuclear weapons are a good thing to have understood that in the 1930s and 40s many governments were seeking them, that we have people still trying to build them today. I feel the same way about actual artificial intelligence. It may well come to kill us all and replace us atop the food chain.
You see, I believe that Fermi's Paradox points one way -- species likely do not survive their full technological sophistication. I think it's a lot more likely that the first alien intelligence anyone ever encounters will be fully synthetic.
No, i don't like it. But there it is.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.