The AI generator museum of uncanny amusement
-
Breadsticks breadsticks breadsticks.
-
Fuck me.Holding the wrong end of the stick since 2009. -
Saw these earlier on a photo page. They were labelled as AI but only once you clicked on them. Have to say that they had me fooled. -
i can't be the only one who doesn't want to interact with an AI as if it's human. I like my AI sterile, and not to waste words cultivating a generic, fake personality.
-
equinox_code wrote:i can't be the only one who doesn't want to interact with an AI as if it's human. I like my AI sterile, and not to waste words cultivating a generic, fake personality.
Yeah, there's a whole level or artifice designed to create the impression that LLM "AI" is thinking, considering, processing or coming to a conclusion, when nothing of that ilk is happening at all. But, the presence of that layer of "hello fellow conscious being" has contributed to a fundamental misunderstanding of what it is and what is actually happening.
It's an efficiency tool and it's best being cold and efficient. Why on earth we've decided to resurrect Clippy, I don't know. -
Minnesänger wrote:It's an efficiency tool and it's best being cold and efficient. Why on earth we've decided to resurrect Clippy, I don't know.
As usual, techbros in San Francisco have read a few sci-fi novels and failed to recognise that they were warnings, not guides. So now we’re getting the Internet of Shit with Genuine People Personalities™. -
To be fair there are a lot of people who do a great deal of talking without thinking. So it's an easy mistake to make.
But yeah - AI is simply a tool, based upon statistical analysis techniques I believe. It turns words into numbers and crunches away. All computers are a practical implementation of something called a Turing Machine which is a theoretical automaton with infinite storage that manipulates symbols according to a set of rules which are represented by symbols. So it doesn't understand anything. If you are interested check out the Chinese Room thought experiment. It's the same thing.
I'm actually impressed by Microsoft's presentation of AI. Even the name - copilot - says was it is. It's not auto-pilot. And their integration with the office suite is well done; albeit expensive.Holding the wrong end of the stick since 2009. -
This is exactly what AI will be/is.
It will be transformative for sure but wont totally replace us. Its a tool, a very good tool with the right application and that change it brings could be massive but its still a tool. -
I still roll my eyes when i hear people refer to that amazon tube by a name. eugh. same for Apple's intrusive little corner orb, who occasionally shows up uninvited upon my screen after mishearing something.
-
'Google Assistant' is a less annoying name but still doesnt understand "just fuck off" when I accidentally activate it and it asks how it can help.
Its Clippy all over again. -
Bixby is the Samsung one and the name enrages me.
Sounds like a trust fund child whose real name is Rebecca. -
Right-o Bixby, i trust you'll be at Aintree this year? Wasn't the same last year when I was stuck between Chadders, Ducky and a bunch of oiks.
-
Am I going this year? Let me ask Jeeves.
-
LivDiv wrote:'Google Assistant' is a less annoying name but still doesnt understand "just fuck off" when I accidentally activate it and it asks how it can help. Its Clippy all over again.
Mine understands "OK Google shut the fuck up" -
oldmanshoutsatclouds.gif x 10 itt -
I just don't get the voice command thing. Has no appeal to me at all.Holding the wrong end of the stick since 2009.
-
It’s great in a car.davyK wrote:I just don't get the voice command thing. Has no appeal to me at all.
Or at least … it is in theory. I don’t have a single car it works in. They’re all either too old, too loud, or both. -
AI is currently helping me do some data analysis for a report I'm writing by giving me a step by step guide on how to use python in jupyter notebooks, and then telling me what the faults mean and how ro rectify them. -
poprock wrote:
It’s great in a car. Or at least … it is in theory. I don’t have a single car it works in. They’re all either too old, too loud, or both.davyK wrote:I just don't get the voice command thing. Has no appeal to me at all.
I can see its use in making an outgoing call and a few other things in a car. Fair point.
I can answer a call without taking my hands of the steering wheel which is great and I don't need voice control for that.Holding the wrong end of the stick since 2009. -
poprock wrote:
It’s great in a car. Or at least … it is in theory. I don’t have a single car it works in. They’re all either too old, too loud, or both.davyK wrote:I just don't get the voice command thing. Has no appeal to me at all.
Voice commands in an XPeng are stupidly good. Great cars, but incredible software. -
equinox_code wrote:i can't be the only one who doesn't want to interact with an AI as if it's human. I like my AI sterile, and not to waste words cultivating a generic, fake personality.
-
davyK wrote:If you are interested check out the Chinese Room thought experiment. It's the same thing.
The Chinese Room Argument is fallacious in a number of ways and not particularly useful.
-
davyK wrote:To be fair there are a lot of people who do a great deal of talking without thinking. So it's an easy mistake to make. But yeah - AI is simply a tool, based upon statistical analysis techniques I believe. It turns words into numbers and crunches away. All computers are a practical implementation of something called a Turing Machine which is a theoretical automaton with infinite storage that manipulates symbols according to a set of rules which are represented by symbols. So it doesn't understand anything. If you are interested check out the Chinese Room thought experiment. It's the same thing. I'm actually impressed by Microsoft's presentation of AI. Even the name - copilot - says was it is. It's not auto-pilot. And their integration with the office suite is well done; albeit expensive.
I'm not sure it's relevant to say AI is just a Turing Machine and it's the way these new AI chips are made. It's how you put the transistors together that differentiates old vs accelerated computing. Yes it's just crunching numbers but the complexity space is just so much bigger than the old chips. AI can do stuff that wasn't possible before and it's not because of Moore's Law. It's a whole new thing and the difference is these incredibly complex hyperplanes, often in billions of dimensions. It's an unbelievably complex geometric object and it's the reason it's so hard to understand what an AI model is actually doing."Plus he wore shorts like a total cunt" - Bob -
SpaceGazelle wrote:
I'm not sure it's relevant to say AI is just a Turing Machine and it's the way these new AI chips are made. It's how you put the transistors together that differentiates old vs accelerated computing. Yes it's just crunching numbers but the complexity space is just so much bigger than the old chips. AI can do stuff that wasn't possible before and it's not because of Moore's Law. It's a whole new thing and the difference is these incredibly complex hyperplanes, often in billions of dimensions. It's an unbelievably complex geometric object and it's the reason it's so hard to understand what an AI model is actually doing.To be fair there are a lot of people who do a great deal of talking without thinking. So it's an easy mistake to make. But yeah - AI is simply a tool, based upon statistical analysis techniques I believe. It turns words into numbers and crunches away. All computers are a practical implementation of something called a Turing Machine which is a theoretical automaton with infinite storage that manipulates symbols according to a set of rules which are represented by symbols. So it doesn't understand anything. If you are interested check out the Chinese Room thought experiment. It's the same thing. I'm actually impressed by Microsoft's presentation of AI. Even the name - copilot - says was it is. It's not auto-pilot. And their integration with the office suite is well done; albeit expensive.
It's still a Turing Machine. And it still has the gaps in computable functions that computers have, because it still knocks up against Godel's Incompletness Proof and the fact that no piece of software can prove a program will not loop forever. We can represent n-dimensional models very easily in software but it's represented in RAM as a sequence of switches/latches in different states. A large arrangement of JK-flip flop circuits made up of gates. (there may be a modern version of them now but 1 bit is just an arrangement of switches in a given state.)
Throw as many transistors at it as you like and configure them how you like. It's still a dumb machine transitioning between switching states.
BUT......see my next post. Holding the wrong end of the stick since 2009. -
The Chinese Room Argument is fallacious in a number of ways and not particularly useful.If you are interested check out the Chinese Room thought experiment. It's the same thing.
Debatable but that's what makes it all interesting, don't it?
Fact is, no-one actually knows because we don't know what makes us work. What does understanding actually mean? So we have to go with what we have which is logic and maths (which is symbolic logic).
Some people might say that the difference between us and switching machines is the soul. Said this before here but we start to knock up against philisophical arguments.
If you don't believe we are magical/have a soul then you start to have to believe that a machine can be self aware because all we are is an arrangement of matter so we could build a machine that is self aware. And so such a machine would have as many rights as we have.
Personally speaking? I don't really know. I'm doubtful of the existence of the soul, but I'm not convinced computers as we build them are the answer. There might be something in quantum computing if for example quantum effects are at play in our brains.
If we are self aware (and what we perceive as reality isn't an illusion) then our universe is therefore just an arrangement of matter that has become self aware because we are part of it.
And maybe that's our purpose.Holding the wrong end of the stick since 2009. -
Yes. Self-awareness is just complexity that comes from arranging massive amounts of atoms in a particular way, and saying AI is just a Turing Machine is like saying our brains are just made from atoms - which they are, but it's not a relevant point.
We need to understand the science of complexity better."Plus he wore shorts like a total cunt" - Bob -
Oh christ, LLMs are going all chatty.
"Plus he wore shorts like a total cunt" - Bob -
GPT talking to itself.
"Plus he wore shorts like a total cunt" - Bob
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Categories
- All Discussions2,718
- Games1,881
- Off topic837