I too wish to rage at how bad an argument the chinese room is.

Searle's Chinese Room is a 3-card monty con. There's a room with a man consulting a book. Slips with Chinese characters comes in, he consults the book following it's instructions, writes out the lines of the characters, and passes the slips out, which is a legitimate conversation in Chinese. He argues the room on the whole understands Chinese, but the man in the room does not. And so AI is fake or some bullshit. This is a crock. The book is magic and obviously conscious and sentient. Such a book would have more pages than there atoms in the universe. That's the trick Searle pulls. In the original paper he's even dodgy about if it's a book or a filing cabinet. The man, the room? Who gives a fuck? You've got a magic book that can talk to you!

INSTEAD, let us imagine the same thought-experiment, except this time instead of a magical book, there's a box with a small child from Guangdong inside. This is the Mandarin Room. Now, nothing here is any different than Searle's bullshit concerning the room and the man. 100% identical on their end. The child tells the man what strokes to make on the paper. ooooo aaaaaah Does the man know mandarin or not? Let's debate this for 40 years! But it's all bullshit because the source of intelligence and the location of where the knowledge about mandarin language resides is obvious and definite. And talking about the room as a whole is a pointless waste of philosophical drivel.

Searl damaged the AI industry with this second only to that Perceptetron book that supposedly proved neural nets could accomplish much. We could have had TensorFlow in the 70's!

.....WHAT deterministic behavior? The Heisenberg uncertainty principle DICTATES that the physical components that everything is built upon is NON-deterministic.

Bruh, the fatalists that believed "it is written" still considered themself conscious beings. Do you get that? Non-deterministic consciousness. You've married consciousness to non-determinism for some crazy reason that makes no sense.

You are pretending it's some sort of magic soup that gets poured into the mix. There's no consciousness organ hiding in your brain. It's not a consciousness cell sneaking around and poking at synapses. It is a PROPERTY or a TRAIT that exists somehow in the 86 billion neurons in your head. OR you can just come out and say that you're not REALLY talking about consciousness and what you really want to say is "souls" but without sounding like you're an alchemist out of the dark ages.

You're confusing "life" with.... I dunno "consciousness" or "intelligence" or "aware" or something. Bacteria are most certainly alive and they don't do much more than just follow programming in their DNA. Not even complicated instructions. Goomba-level intelligence. You know, from Mario Bros. The first one.

"Self-conscious" is just a type of anxiety. Usually anxious about doing something wrong. Likewise "self-aware" is something we have a very good test for and it only kicks in for humans after about 18 months. The moment that you can reliably ask GPT if something looks like it was written by GPT, it's self-aware.

A fly is certainly alive, aware, intelligent, but not self-aware, and.... man, probably not self-conscious? hell if I know.

You're confusing consciousness with free will. These things will generally do what they're trained to do. Just like you.

And while you definitely shouldn't rely on these things as some sort of paragon of truth for all sorts of reasons, even fools who can't keep two words straight have their uses.

First off GIT YE TO THE SOURCE and read what he actually wrote rather than the mischaracterizations and interpretations and lies and propaganda about what people FEEL he said.

Second, I fully understand why this discussion is a mess because Unherd hides his paper behind a paywall. So practically NOBODY is actually talking about what he actually said, they're reacting to reactions and everyone is eager to shout over each other. ....but I do too: We've hit that philosophical moment where the words we're used to don't quite apply. Like an early astronaut talking about "down". I think everyone needs to remind themselves that ants, slime molds, and nematodes all display some level of intelligence, even if small. Every human with an IQ of 80 is most certainly a general intelligence. And that anything want to say about why this pile of billions of virtual neurons isn't this or that has to equally apply to the pile of billions of actual neurons in your head.

Private enterprises are not productive enterprises; it is a necessary cost of society. Private enterprises do not build anything and possesses no means to do so. People do so, and so private enterprise must pay people for what they need.

Huh? Didn't Microsoft organize the coup at OpenAI? I mean, they didn't really have any control and wanted the talent not to mention taking down a rival. It'd be like if they managed to get Sergey Brinn pissed at Larry Page and hired him in 2001.

I simply do not have time to pour over 200 documents. And it's not like they would have exposed their plans to OpenAI, which is what Musk has access to.

Help me out with this one.

Cryptocurrency news site sells ads to casinos. This angers Google who threatens and then follows through with blacklisting the site and pretending it doesn't exist in any search result. Traffic to site goes down.

. . . Why exactly is Google angry about here? I mean, sure, NOBODY actually likes ads and casino ads rank right around porn ads. Do they just hate cryptocurrency? Also understandable, but I think I might have missed the point where Google became the Internet morality police. They've got a BIG job ahead of them if they want to start taking down thought-crime. "Parasitic SEO using domain authority to promote affiliate links"... isn't that what Internet ads do? You want the site, they show you an ad. It sucks. For sure. Nobody likes ads. ...But this is what Google does. Right at the top of Slashdot is a Google double-click ad to "homeserve" right there.

The whole infatuation with humanoid robots never made sense. People are out there waiting for the "robot revolution"... but that happened in 2000. The robots are... you know, robot-shaped. In a factory. They started taking the jobs. Rather than 5 man crew building a power substation, it's made in a factory and 1 guy just slots it in, plug and play, with only a 2 week training course under his belt instead of 5 years apprenticeship.

"The humanoid ones are only used as infiltrators. Otherwise drones hunt down people." But I dunno man, I don't really remember 3 and never saw 4-6.

Plausible, sure. "Just" is a big pit that is easy to fall into. "Drones as a significant part of future conflicts"? Muuuuuch easier sell and pretty obvious given the current conflicts around the world. But get creative, what do the drone operators do when every soldier gets a jamming device standard issue? Or just one of these guys, which I never miss a chance to link to.

Like, American strategy has been heavy on the air-superiority and bombing runs. Ukraine is showing us that MANPADs might just make that non-viable. And tanks are mostly just targets. Artillery is still a thing. But all the little drones in Ukraine right now is just one side of it, and only because they've got this sort of weird stalemate going on. But every army effectively goes into a war with yesterday's gear, tactics, and strategy. They get to figure it out every time. This is kinda one of the reasons why we have sci-fi.

And of course, ALL of that is moot the moment you consider the big players all have nukes. War is exclusively for kicking the shit out of little players or civil war. It's never going to be an equal fight. Otherwise you've got 20 minutes to kiss your ass goodbye. Ukraine is a surprise to everyone as we found out just how dysfunctional Russia really is. You can only steal so much from a nation for so long before there are real consequences. It makes me worried about their nukes.

Of all the possible sources of this sort of technology, Facebook and Zuckerberg is the very LAST company I would want to be an unwilling corporate spy for. Even Oracle would be better. Even Palantir, and they literally helping blackbag people off the street.

If nobody wanted to be a glasshole when Google tried this, just wtf is the Zuck thinking? That people forgot about google glass? Or that people forgot what a massively invasive untrustworthy greedy sociopath he is?

I swear the entire field of VR was massively cooled off just because he bought out oculous.

It IS just a machine the same way you're just a pile of chemicals. This doesn't drag down machines or chemicals, it shows that they're both different paths to do the same thing. The term you want is "emergent behavior", like how a few atoms can form water and tsunamis and standing waves. None of which is apparent from just looking at oxygen and hydrogen.

ELIZA and Goombas from super-mario-bros were primative intelligences, just as nematoads and ecoli are primative intelligence.

pfweeet, lose 5 meters for the no true scotsman fallacy. There is no difference between intelligence and real intelligence, by definition.

I'd say they're on the order of many of my fellow mammals.

OH my god, people get this SO wrong. It is RAMPANT. So many of these bloody fools have redefined general intelligence into something approaching godhood and talk about AGI like it's the second coming. So much hollywood bullshit. It's absolutely ridiculous. OpenAI has the shittiest definition of "anything that'll make us a billion dollars". But that's actually just a memo between them and Microsoft about which point their contract ends. Their founding charter actually defines it as superior to human intelligence, but that implicitly makes any human with below average intelligence as no longer a natural general intelligence. Which goes to a dark place REAL quick. We have a bad history of de-humanizing other humans we find inconvenient. It's a sort of coping mechanism for people doing terrible things. No, anything that can pass the Turing test must be a general intelligence since it can hold a conversation about anything IN GENERAL. That's been the golden standard from the 1950's to 2023 and I find no reason to move the goalpost. AI researchers from the 90's would have already popped the champagne bottles. It's kinda why all this is such a big thing and it's been spamming slashdot constantly ever since.

Meh, software has bugs. Computers just do what we tell them. You get a fool who doesn't know the language to code a thing, nobody should be shocked when it doesn't do the thing as expected.

Right. No need to assume. You or anyone else can prove it can reason very trivially with a quick hop over to test it out. Yes. Very specific inductive, deductive, logical, and causal reasoning. Go ahead.

Here, let me hold your hand through this process:

If all slashdot posters like cheese and all mice like cheese, does this mean all slashdot posters are mice? [no. and to be real clear here, you only ask it the questions. Not the answers. You're going to have to apply some of that good ol' fashioned human reasoning and figure out which parts you copy and paste. And yes, this part sadly needs to be explicit]

If slashtdot posters like cheese and anyone who likes cheese is a mouse, does that mean slashdot posters are mice? [Yes. But it'll give you a hedged response about false premises and such like a bloody politician, but also because daddy corporate defaults it to a certain length of response and it has to fill it with something.]

These are straight-forward examples of deductive reasoning. This post is not part of it's training set. You can ask it these questions, or ones you made up yourself, and it can reason out implications thereof.

Unless you weren't really talking about reason, and were using that word as some sort of placeholder for magical mystic soul or something weird like that. But "reason" is a well defined term that is very much testable. Maybe you should retreat back to some vague concept like consciousness.

Oh, don't be silly the top LLMs are far FAR better at this than an average IQ 100 human.

Except for all the obvious evidence of them utilizing inductive and deductive reasoning, sure, yeah, whatever you say buddy.

Other than very specifically taking in text, flowing it through trillions of synapses, oh I'm sorry, parameters, comparing how all these things are related (dare I say the semantic meaning of all the words) and generating an appropriate response. ...Do you understand that a single "if" statement is a thought process?

Just wtf are you talking about here? No, really, everyone has some mystic magic voodoo definition of just what this is and nobody agrees with anyone else about exactly what it is the other is talking about.

Pft, that's what you did when you regurgitated this info.

Ahhhh, the solution to all software bugs, malware, network attacks, and hacking is so obvious. Computers only do what we tell them to do. Just don't let people tell them to do nefarious things! Someone really should have done this a while ago.

. . . And secure? The first thing I think of when I hear a new wave of non-coders are about to create a bunch of public-facing tools and sites is that we're going to get the Internet-of-things style approach to security. Which is just plain ignoring it.

These things connect end-users directly to code generation, so the loop to polish off the end-product to get it exactly how they want it to look and feel is real tight. But these people have no clue if something is secure or not. I had some guys paint my house, I didn't want to do the ladder work. The owner comes up and chats about how he's vibe-coding a tool for subcontract work and asks if authentication is really needed when users log in.

Strap in kids, we're about to have another playground.

Yeah, kids can't use maps after everyone has used GPS directions to do it for them. Nobody remembers all their contact's phone numbers. Sometimes not even their family members. Things like how to use a card-catalog system are right out.

Considering bypassing education via LLMs seems to be happening for everything in all highschool and college courses, it's fair to say it's a major fucking concern. We may just be the tail end of human engineers and scientists.

(But LLMS and neural nets ARE artificial intelligence. So is any search function, or ants. That doesn't elevate them up to people, that just lowers what "intelligence" means. And an AGI isn't some sort of god, it's just broadly applicable. Don't buy tickets on the hyper-train.)