Fundamentally, anything humans can do can be done by physical systems of some kind, (because humans are already such a system), so given enough time I’d bet that it would be eventually possible to make a machine do literally anything that can be done by a human. There might be some things that nobody ever does get an AI to replicate even if technically possible though, just because of not having a motivation to
The flavor of cinnamon toast crunch.
Was my exact thought. Lol
Since AI is trained by us, using the fruit of human labor as input, it’ll have to be something we can’t train it to do.
Something biological or instinctual… Like being in close proximity to an AI will never result in synchronized menstruation since an AI can’t and won’t ever menstruate.
So… That 👍
Synched Menstruation is supposed to be a myth now. I have experienced it many times, but I guess it’s mostly considered coincidence, which it could be, I’m not a mathematician.
What would you bet on AI not ever getting the ability to menstruate?
An exact 1:1 realtime copy of itself emulated within a simulated universe.
Pretty much everything else mentioned in this thread falls into the “never say never” category.
Also being able to analyze any program and guarantee it will stop
Probably still a never say never problem:
In their new paper, the five computer scientists prove that interrogating entangled provers makes it possible to verify answers to unsolvable problems, including the halting problem.
If you actually read the article, it doesn’t say anything about being able to solve the halting problem. It used the undecidability of the halting problem to prove equivalence of another class of problems to the halting problem.
Which is why I said it was still a “never say never” and not an already solved problem.
The halting problem is impossible for Turing machines, but if hypercomputation ends up possible, it isn’t impossible.
For example, an oracle machine as proposed by Turing, or a ‘real’ computer using actual real values.
The latter in particular may even end up a thing in the not too distant future assuming neural networks continue to move into photonics in such a way that networks run while internals are never directly measured. In that case the issue would be verifying the result - the very topic of the paper in question.
Effectively, while it is proven that we can never be able to directly measure a solution to the halting problem, I wouldn’t take a bet that within my lifetime we won’t have ended up being able to indirectly measure a solution to the problem and directly validate the result.
Computers will never consistently beat humans and humans will never consistently beat computers as snakes and ladders.
Or rock-paper-scissors, for that matter.
Didn’t some robotics lab build an unbeatable rock-paper-scissors robot a while back?
Calvinball! All hail Watterson lol
Pretty sure it won’t manage Ligma any time soon
They will when they perfect the bofa fill algorithm
There are already specialized robots just for that
Pooping, my guess is pooping.
TIL.
deez nuts
Organic intelligence? The qualifier never kind of removes a lot of answers when you also say “never”
A bit fallacious to add “organic” to intelligence. But then I’m sure we will be able to make organic computers at some point. I think there is research into this already.
Yeah there’s that one thing they were using rat cells or whatever and got computations off of it
Yes, yes, we all know about how OP struggled in high school mathematics
How humans think. AI “thinking” will always be different than human thinking. Because human brain is “that thing” that is impossibile to simulate in silico as is. We might be able to have good approximations, but as good as they can get, they’ll always diverge from the real thing
I guess a good part also comes from learned experiences. Having a body, growing up, feeling pain, being mortal.
And yes, the brain is an incredibly complex system not only of neurons, but also transmitters, receptors, a whole truckload of biochemistry.
But in the end, both are just matter in patterns, excitation in coordination. The effort to simulate is substantial, but I don’t see how that would NEVER succeed, if someone with the capabilities insisted on it. However, it might be fully sufficient for the task (whatever that is, probably porn) to simulate 95% or so, technically still not the real deal.
What makes you say that so definitely?
Funny enough I have the opposite opinion, human brains are the type of thinking we have most experience with - so we’ve devised our input methods around what we notice most, and so will be able to most easily train the AI.
I also believe that we’ll be abke to reduce the noise to a level lower than actual person variation fairly easily, cause an AI has the benefit of being able to scale to a populous size - no human even has that much experience with humans
I use to work on research on microscopic mechanisms of the brain, and I work in AI.
Human thoughts derive from extremely complex microscopic mechanisms, that do not “average out” when moving to the macroscopic world, but instead create very complex non-linear stochastic process that are thoughts.
Unless some scientific miracle happens, human thoughts will stay human.
But an AI does anything but average out, else we wouldn’t be any more advanced than the earliest mathematicians.
Its skill comes from being able to have millions to billions of parameters if required, and then contain data within them all.
It doesn’t seem entirely unreasonable that it could use those (riding off our suprisingly good math skills) and create a model that represents a human with low enough noise we wouldn’t even notice.
(but also I’m in a similar more chemically focused field, nanotechnology so I have experience with nanoscopic-microscopic structures, and what can we artificially build from them while not killing the biological side of things)
As you are in nanotechnologies, when I say average out I am talking in a statistical mechanics way, i.e. the macroscopic phenomenon arising from averaging over the multiple accessible microscopic configurations. Thoughts do not arise like this, they are the results of multiple complex non linear stochastic signals. They depend on a huge amount of single microscopic events, that are not replicabile in a computer as is, and likely not reproducible in a parametrized function. Nothing wrong with that, we might be able to approximate human thoughts, most likely not reproduce them.
What area of nanotechnology are you? Main problem of nanotechnologies is that they cannot reproduce the complexity of the biological counterparts. Take carbon nanotubes, we cannot reproduce the features of the simpler ion channels with them, let alone the more complex human ones.
We could build nice models, with interesting functionality, as we are doing with current AI. Machines that can do logic, take decisions, and so on. Even a machine that can predict human thoughts. But they’ll do it in their way, while the real human thoughts will most likely stay human, as the processes from which they arise are very human
nano engineering, and course were talking some years in the future, but if anything nano’s convinced me were all just math when you break it down - when just depends on how much math we can do.
Even a simple conversation can be broken down into tokenizable words recently and bam chatgpt, reasonably the rest of our ‘humanity’ could be modeled following a similar trend until the Turing test is useless
What I mean is different. A dog thinks as a dog, a human thinks as a human, an AI will think as an AI. It will likely be able to pretend to think as a human, but it won’t think as one.
It won’t have a Proust’s madalaine (sensorial experiences that trigger epiphanies), have the need to travel to some “sacred” location looking for spirituality, miss the hometown were it grew up, its thinking won’t be driven by fears of spiders, need of social recognition, pleasure to see naked women. It’s thoughts won’t be dependent on the daily diet, on the amount of sugar, fat, vitamins, stimulants intake.
These are simple examples, but in general it will think in a different way. Humans will tune it to pretend to be “as human as possible”, but humans will remain unique
Cracking my knuckles nervously before I’m about to give a presentation in front of the whole class.
Social anxiety
Well, I know my social anxiety is basically just a hallucination based on a bad data set.
I’d like to be proven wrong but Empathy
Giving everyone money for free from the rich people! Yeah, that’s right… wealth redistribution! AI won’t ever be able to do that.
Truly creative, decent Dad Jokes.
I don’t know, there are a couple pretty good ones here by chatgpt:
Of course! Here are some classic dad jokes for you:
- Why don’t skeletons fight each other? They don’t have the guts.
- Did you hear about the cheese factory that exploded? There was nothing left but de-brie.
- I used to play piano by ear, but now I use my hands.
- What do you call a fish with no eyes? Fsh.
- Why did the scarecrow win an award? Because he was outstanding in his field.
- What’s brown and sticky? A stick.
- How does a penguin build its house? Igloos it together.
- I’m reading a book on anti-gravity. It’s impossible to put down.
- Parallel lines have so much in common. It’s a shame they’ll never meet.
- Did you hear about the mathematician who’s afraid of negative numbers? He’ll stop at nothing to avoid them.
Most of those predate the internet
then it truly is human, cause shit if 99% of us nowadays aren’t just googling dad jokes when we need them
