@@mattmarzulasomeone being faster than you has nothing to do with your creativity, and if we are going off the information given also has nothing on your originality.
@@mattmarzuladont down yourself convergent creativity does happen, theres even a couple musicians who accidently recreated the same notes as songs they had never heard, being original doesnt necessarily mean only you came up with it or you came up with it first, it just means you came to that result be your own devices, so good on you sir😊
The average human is not trained in creative thinking, while the AI is probably trained on every creative work ever produced, and it still won't beat the most creative human
It will never be, unless the whole program changes and we somehow figure our how to let it have a mind of it*s own. Since now the ai can only collect data, some of which are stolen without consequence and write answers from this piled up data. It can't be better than best humans since it does not have more information that us humans have given it.
You're comparing it to the absolute best we have to offer. If it's better than average it still leaves 4 billion people in the dust. AI that could win in chess was developed decades ago, but no AI could consistently win against grandmasters of the game Go until like 2019. It's not a matter of if. Only when.
Kind of reminded me of the Yellowstone park ranger that comments about the trash can locks, “there is a considerable overlap between the intelligence of the smartest bear and the dumbest human”
On what do you base that? How can you fake creativity? In any case they were not trained to be creative - the researchers were as surprised as anyone else when they discovered what LLMs could actually do.
Chatbots work based on statistical tools, but you're brain also can be modelled as statistics too... The rate neurons are firing, localized behaviour, and all other emergent properties are basically statistics too.
There's nothing in the definition of "AI" that says that it can't be predictive text. We've been calling plenty of less advanced systems "AI" for decades, there's no reason this can't be as well
I can go a bit more in depth, but when u want to model the global tendencies of a model, you use ststistics because that's basically what it was made for. You start with random events (or events impossible individually to measure so they might as well be random), make a couple of assumptions, like the firing rate of neurons tends to be localized and periodic, do more analysis like a topological map of neuron activations in your brain, and with statistics you can basically infer what the tendency of the model is.
This just shows that you haven't researched how they are actually trained. There are no "correct answers" to a creativity test like there are for say maths.
@@manuellayburr382 That is.. entirely irrelevant. People's answers are out on the internet somewhere, if the AI had access to that, it could simply use one of the highest rated ones
@@guymanyes4756 That is not how it works. You may not be aware that the LLMs have come to conclusions they were never trained on through inference. If you are unfamiliar with that check it out it's very interesting.
It doesn't have to be. Look at people on the internet, when it comes to creative thinking most can be easily beaten by the AI. Or by a squirrel. Or by a bucket of fish guts.
@@neverstopschweiking that's true, but you also need to remember that it didn't come up with anything unique, it built off of already known ideas. You won't find any sort of breakthroughs that way, just advancements which are great don't get me wrong
Creatives are *already* losing job opportunities to subpar AI results! We've been seeing it with book covers, exhibition posters, sometimes even newspaper illustrations. The problem is never that the AI is more creative or proficient, plus it also has an... ambiguous relationship with copyrighted work. The deciding factor is always money, unfortunately 😅
When they first tried to use machine learning for creative work they attempted it with music. Immediately shut down by the industry. Wish the same could happen for artist.
To be honest, graphic designers did get cocky, just like average IT "specialists". Thankfully AI is grounding them a bit. Those who learn how to work with AI will stay successful, though. The most important thing is to embrace the change.
@@BojanTomic any Ai picture is just an amalgamation of stolen artwork. Just because you can't tell the individual pieces apart anymore doesn't change that.
The way I understand the topic is: People invent. Chatbot use things already invented by humans and mix them up based on human imput. I think human over reliance on AI might be an issue where it can create a closed loop, but really special and creative people will always push the world forward.
AI systems require data of course, we could give it to them or make them collect the data themselves, that's just incidental. so far we usually provide the data for better chance of creating a productive tool quicker but having them gather data themselves would be the more organically scalable method in long term (all intelligences including humans go through phase of being untalented babies but at the moment the industry is moving too rapidly to tolerate delay). saying humans have some kind of monopoly in creativity or inventing is copium, the scientific method is basically a computer program already. read up on latent spaces, connecting disparate topics isn't that hard. in the past overly optimistic people have had disappointments in trying to create AI because the resources simply weren't there but falling back on assumptions based on that would be foolish, like Amara's law says "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run". even if it might still take more than our lifetimes to create truly independent self-improving general AI doesn't really matter in grand scheme of things, the seeds are germinating, it is happening.
@@snooks5607 You don't seem to understand. AI gets it's information from humans. AI will never do something beyond what humanity has already done. AI does not have the ability to think. None of its ideas or solutions are new, you can find then in the material the AI was trained on. A brain that actually can think will always win against a computer designed by those that can think. The only way AI could ever gather any information is in the way we design it to. It will never exceed the scientific capabilities of humans, it'll just be able to do more experiments quicker. For surveys or other such, the AI would be getting information *from humans*.
There’s more than one type of creativity. Are you a “pull new ideas out of thin air” type of creative person or a “put things together in novel ways” type of creative person?
Its the same thing, when u think of something "novel" u use preexisting knowledge and mess with it in a new way. Science fiction emerged with the rise of scientofic literacy, go figure.
Unique ideas are extremely rare even in the state of the art science labs with brilliant amongst us working there. Most things, insights or ideas usually fall in the 2nd part you've mentioned
That just means that a bot, trained using the entire creativity backlog of humanity, has beaten, in some cases, 250 random people. We're comparing robots trained with the best we've ever had, versus random average people. I'm honestly not scared right now
@@SyNcLifeNo lol... what a dumb suggestion, I didn't get that at all reading his comment. He's saying that the average of people, when combined, is smarter than AI because that's how AI works in the first place. If you took the best 1,000 painters in the world currently, they'd beat AI and make less mistakes. Same if you took the best engineers, and the best scientists. AI isn't some existential crisis that's becoming God.
I like how little you guys understand what they’re saying. They’re saying that These chatbots are taking information from some of the most creative humans out there so up against average humans they win. It’s not impressive because the chatbots are just stealing the creativity.
You forgot to mention one very important catch. The chatbot can invent endless answers to an assignment like that, but it has no understanding whatsoever of how objects function in the real world. I suspect that humans avoid some of the chatbot's answers because they understand that these things are not physically possible, or at least highly improbable.
The answers they produce require having some internal model of the world. They absolutely do have some understanding of how things work. These algorithms are way more advanced than some simple markov chain prediction algorithm.
Predicting the next word while considering and in accordance of the previous word, sentence, paragraph, page, or book worth of text, these Bots basically ended up learning language, grammar, meaning and semantics. (leaned patterns are shown to also encapsulate the meaning of the text) It works on the principle of the cellular automata,simple rules yielding complex results beyond comprehension, where now predicting the next word at these scales means a whole lot and some sort of understanding is involved to do it.
About the "physical sense and understanding of the surrounding" part, these models are getting integrated into robots that has been given abilities to sense, see and hear (look at the deepmind research) I mean, it's gonna come sooner
@Tomyb15 AI does not have any sort of internal model. Just enough training, it mostly avoids dumb answers. It's just a lot of multiplication and addition.
@udaykadam5455 But that's not where it's at now. Now it's just a good Googler. Who's only sense of the real word is being told that one idea was dumb so let's not do that and do this slightly more probable one. It just has training. No other probability detector
Creativity is kind of a vague thing, and the test used is actually a case where LLMs can perform 'well'. Not because they have anything resembling understanding of what they're doing but because 'take thing, do something atypical' is at the core of how deep learning functions. Other commenters have also mentioned that the LLMs in question might well have been trained on data that included similar problems. TL;DR: No, AI has not surpassed human creativity, and I strongly doubt they are ever going to as they exist because they are based in remixing what we have already made. That kind of AI can never truly be original, even if we struggle in that regard too.
Well I know this argument has been made to oblivion, but your brain also "rehashes" your previous experiences and stimuli. Sure they might be your UNIQUE experiences, but the sheer quantity of knowledge LLMs have can get you a run for your money
@@superfeel1275except LLMs don't have any frame of reference for what they're spitting out. They're pattern replicators, and we are pattern seekers, so they're very good at initially seeming similar to human output. But until they have senses and ability to reality test, no model will be able to compete fully with any single human. And when they do acquire the ability to gather unique stimuli, the ethics of seeing them as nothing more than tools will come to a head. At this point they're an elaborate parlor trick. Once they're able to actually have novel experiences and test their own spat out nonsense, they'll have crossed over into an intelligence that may be unethical to use without consent to its labor
I saw a ChatGpt demo where they asked it to make a painting of Moses competing against the Pharoah's wizards. It showed Moses and the wizards and lighting and was interesting. Then they showed a painting made in the 1700s and it was immediately obvious that it was a masterpiece of composition,lighting and technical skill. While computers can render competent illustrations in seconds,they only compete against your average human scribbler but against true human talent, I think it can only imitate not create.
Not to say anything bad about your post.. it just occurred to me that since art is subjective.. you are saying that the image the Ai generated was subjectively worse than one created by a human talent. Since it is subjective.. it may apply to you but might not to everyone else. Abstract art, for instance.. to me is horrible art. I much prefer realistic portrayals even if they are of fantasy subjects. This is my subjective opinion, obviously. The point I am making though is that since art is subjective it is impossible, really, to gauge how 'good' an AI is at creating art. The only real metric which could reasonably be applied is to see how much someone would be willing to pay for the art.. and even that is a bad method considering what the majority of art that is highly praised in these days sold for back when they were originally created.
@@uoabigaillevey so you are saying that a Picasso is equal to an eight year old's refrigerator drawing? A layman can possibly look at 2 pieces of art and not understand why one is a masterpiece whole the other is not. To make "Art" is much more than just making a picture. There is composition, balance ,harmony ,proportion...an entire universe of things that go into making art. It was obvious to me which was a better image because the human used all these elements and the computer had trouble rendering hands ( despite having billions of examples to choose from). If you truly cannot tell the difference between a computer pretending and human's work,that just makes me sad for you.
@@rayramos8435 It is not a matter of classifying a work of art by it's use of composition or any of the other things you put forth. If it sucks to look at it.. it sucks to look at it. Picasso is horrible as art in my opinion. I would rather watch paint peel than look at it. It is my subjective opinion.. which by definition cannot be wrong as it applies to only me. That is the point I am making.. you are stating that an 'objective' opinion is to be applied to art.. when art is 'subjective' I am stating that, objectively, you are wrong.
@@uoabigaillevey I think the point is that the human moses painting was incredible, it felt immense and powerful, it had a lot of feeling in it, while the AI painting was just kind of the characters standing there, it had no soul because it was made by a machine.
They also did a study on A.I. art where if you keep feeding A.I. art that has been made by A.I. you consistently get worse art. (In other words the A.I. art bots can inbreed and create bad data that we humans still have to sort out and find and create good data for A.I. bots to use.)
This has been a problem with humans in any creative space for ages. It's called the commercial "creative" fields of art, music, movies, tv." The artworld is a world of thieves that clone, copy, and mimic on a daily basis with only true originality cropping up every so often when you get someone creating a novel work which will then be copied for another long period of time. AI is no different and it will easily surpass us on all fronts very shortly with only the current human experts in their holding out for a bit longer.
@@jesseeganpoirier376, I think the issue was that ChatGPT is aggressively average. It’s become very talented at finding enough writing that matches the topic to ‘write’ on its own. But it’s not actually ‘writing’ it. The program is averaging what has been written to give a standardized boring pat answer. But it can produce this standardized generic answer in minutes. When there’s so much AI content out there that its basic input is millions of standardised generic AI content, it’s folding in on itself and created worse content.
It’s not creative thinking if it’s just learning on that specific subset of questions. That’s just being better at memorizing than humans (shocker) and actually creative humans would always be better. Current level of AI has its uses for sure, but I am not understanding why people want to believe that it’s already better than humans so bad
Thats only if they are disgeniune with their results. But, the tendency for ai research is to be frank with your results as any breakthrough will be peer reviewed to oblivion. If researchers want true results, they make up questions that arent preexisting duhhh. Idk why pain a bunch of human subjects if ur going to fake it
The feild is in its infancy and has already surpassed humans in various domains from board games to language comprehension. We overestimate average person.
It's because everyone who hasn't been studying this for years is stunned by what they are capable of, because until now it was purely academic. The only thing truly new about Chat GPT is that it has a public-facing UI - language models that are architecturally similar have been around for years now. Google searches are driven by LLMs, facebook uses LLMs as part of its text moderation strategy, home assistants like Alexa use them to process things we say to them.
The AI might've just read the problem somewhere before. Recalling information is one of machine learning strengths. ChatGPT isn't smart, it just knows which words are most likely going to appear together in what context. It doesn't actually know what it's saying.
that's blatantly wrong and ignores current state of the art research. ChatGPT remembers a lot, yes, but it also builds a world model that is able to generalize. You can show ChatGPT novel problems and will usually still perform better than average people. Building a general world model happens during training because after learning surface statistics it can only become better at predicting the next token when it actually understands what' it's talking about.
These languages models *can* memorize answers, but their memory isn't nearly big enough to remember all of the things people think they're memorizing. They definitely do better at the types of questions they've seen before, they need to see examples to learn how to solve something, but they are actually learning the solving process, not just the answer key.
@@admar1208 chatgpt works in such a way that it knows which words belong to which subjects and in which order they most likely appear. That's how it was trained. It doesn't know what a door is, just that it's definition is associated with certain words so it will use them when asked to define it. I get that it feels like you're talking to someone, but you're not. It's just very good at guessing what you expect to read as a reply.
As a truck driver, I keep getting told that AI will put me out of a job. AI can't load and unload a truck, especially when you're out in the middle of nowhere, are the only one there, and you need a key to get in. AI can't fill a truck with diesel. AI can't decipher a State Trooper's hand signals when there's an accident up ahead.
The history of AI is humans saying "An AI can't do X", then when it succeeds in doing X, they say, "Oh well, that's not real intelligence, I'll bet it can't do Y" When IBM's Deep Blue beat world chess champion Gary Kasparov back in 1998, doubters said, "Oh well it's just following a set of rules". Rules is how humans play chess! AI eventually beats every challenge we set it. It just takes time.
@@johnfitzpatrick3094 I'm 79. My father was born before the Wright Brothers first flew and lived to see space travel. A lot can happen in a lifetime. You may not be replaced but it wouldn't surprise me if the next generation is. Of course that's assuming we don't get hit by an asteroid or World War 3!
Yep, first it replaces an average jobs, which is a waste majority of people who can’t create anything significant in their field. But then, new models will learn on that generation, so it will probably out stand a far more intelligent individuals being able to merge several layers of expertise into one.
@@aperson9973 Ai chatbots are extremely intelligent, GPT 4 outperforms 80% of humans on not only iq tests but pretty much every test that measures intelligence. Your notion that chatbots “can’t think” is egocentrist because you only factor if it can think *like a human* not if it can think *at all*. Because to know if it can think at all we would have to first fully understand the programs, which we don’t. You also didn’t answer my question, which is: if not this, what would you consider AI? Like seriously, if nailing every intelligence test isn’t enough for you what is? Would the AI need to function exactly like a human brain with some sort of a simulation of billions of brain cells or something, I’m genuinely curious.
@@MisterDemonYT In order for a creature to be able to think, it needs to be sentient. AI is not sentient. AI simply is trained to see patterns and report to stimulus based on its training. IQ tests mean nothing to this, since AI only sees the pattern.
Fun fact: the Pentagon looked into such a scenario, that is loss of the power grid and most electrical devices in a deliberate Starfish Prime event, 98% of people would die.
@@AtlasReburdenedit can only figure out what we already know. That’s like me saying I know everything and everytime I’m asked a question I pull up google
The AI isn't actually 'thinking' of creative uses. It doesn't have concepts like that. It is pattern matching to all the different ways the pencil had been used in the entire corpus of human publication - not coming up with ideas, just repeating them from an unimaginable large database.
This is what I think will remain the key distinction between AI and Humans. Like, AI is only creative because it was trained on what other people have done... which include creative people. If you train AI on a group of self-admitted UNcreative people, then it's very unlikely that AI will ever do anything creative. Conversely, if you train AI on creative people, then of course it'll be "more creative" than the self-admitted UNcreative people. But that doesn't mean AI can BECOME creative, it just was taught to be creative. AI is and will always remain a labor saving device, but only that. People said the same stuff about how non-linear video editing and Photoshop would destroy jobs because "anyone can be a videographer/artist!"... but obviously those tools didn't replaced anyone.
@@minecraftminertime Yes but there is a human potential: sometimes creativity comes out of nowhere. Not all creativity came from someone else, and arguably every significant invention of humanity came from a spark of genius that came from nowhere. Even if every story today stems from something else before, if you go far back enough, eventually you'll reach a point where someone say "I just made it up". I argue that AI could NEVER create something from scratch, they could only iterate on what already ever exists (at least, whatever it was trained on). In other words: a human is inspired to make automobiles, AI could only ever make faster horses, because that was the only data it was trained on.
@@nickfifteen what is the biological process behind a "spark" of genius? I think that it is still based on other experiences synthesized together. An AI trained to synthesize other experiences may come up with the same idea. For example, if an AI learns all about horses and physics, and it is inputted to come up with ways to improve humanity, it may come up with a car if it is trained with the objective of improving humanity and its architecture makes it synthesize experiences.
@peez_nuts (I saw your username and it's funny.) is that an insult or a joke? or just an observation? (i always hate the way I speak for some reason so I'm sorry if I'm annoying)
@@mr.jitterspam9552 I wasn't thinking u were talking to me or saying that I talk like planet of the Apes, I was just asking what you meant by your comment.
I’m not good at the “think of alternative uses for this object” task but I am good at the “quick, you need to improvise a tool from random objects on your person” task.
@@Audon4150Not only is the suggestion good, but you fail to realize that automated systems already do this and have been doing this for many years. The printing press made reading and writing way more accessible to the world and it also caused a ton of people to lose their jobs. But it obviously generated a lot of money. People use computers to generate money simply because they do a bunch of stuff FOR YOU in a very quick fashion. AI isn't much different, it just allows a lot more "average people" to enter certain fields and generate money more easily. All because your parents didnt fund an artistic prodigy doesn't mean you shouldn't be allowed to make art for a living. All because you were born with no hands and can't paint, doesn't mean you should be barred from the entire art, music, and book writing industries.
@@imveryangryitsnotbutter I simply asked the question because by text you cannot always detect irony. I love the idea and didn't say anything negative about it in my comment..
In my opinion, ai chatbots can be useful for basic storyboard prompts. Maybe if you want somewhere to start or if you have writers block, you can have somewhere to jump from. Of course, not a copy paste. Just a general outline of events you can go towards in a story.
No AI in creativity. As an artist and writer I have been told my whole life that creativity is not "necessary" and os a waste of time. But now that big corporations can lay off a lot of their workers and individuals can get free stuff or make money off it, suddenly now art it is worthwhile and a necessary human right? Be honest, y'all didn't give a flip about supporting the arts or writers until you realized you could get something from AI, and all of a sudden it's fair to steal everyone else's work for personal benefit, and we need to "democratize" that art that you always scoffed at. How convenient. Creativity, and art and writing in particular, are endeavors that make us human; what does it say about the species that we're willing to gut that in favor of laziness, shallow and hollow entertainment, cheapness, and quick but deeply stained profit? Humanity is completely ungrateful for all the creative toil that came before, and undeserving of all that is yet to come, if it even comes to pass because AI will smother it in it's sleep.
@@Tomyb15Lol correct... the 1 guy replaced 49. Without the 1 guy, no one gets replaced. AI exists and only generates when we generate it. So it's still the guy who's replacing other guys
@swickens930 and that's when you get creative with the fuse box.... no more AI... the smarter you think you are the simpler and dumber the solution is to stop you being a problem.. 😮
As an artist i can get excited by the prospect of what A.I. as a tool can open up😊 but when people start calling what an A I. produces autonomously 'art' i start to doubt people's tastes.
Chess is not a good example, because computers have more computational power that consumes tons of energy, human mind on the other hand is super efficient. Brains can calculate like what 2-3 expected moves ahead, computers can easily do dozens. Obviously it's going to beat humans.
I remember my college having a subject called sustainable engineering in which there's a section on finding alternative usecases for everyday objects and the alternative use case for pencil given in the notes was that it can be used as a weapon to stab 😂
It’ll all come down to who can direct the AI to (aid in) crate(ing) something: just like now, you can pretty much search for anything online, but how many people actually search for real thorough academic research vs something much more mundane online? In that sense, even if we end up utilizing AI to create a work of art, instead of picking up ‘traditional’ tools (create a painting, make a movie etc) it’ll still be the same in terms of requiring the same level of creativity to direct AI to do it.
It's hard to direct it to something truly creative. It keeps going back to the path already ran. It's in essence a statistics machine which predicts the next step based on the data it has, so the step will always lean towards the known path.
Eh it's only taking away jobs from nearly useless work like holly weird underwriters who haven't made a decent movie in years and people like accountants Blue collars aren't the ones to fear AI we support it
@@MimOzanTamamogullarhigher standards? If it takes away everyone's self worth and life goals how will that make any standards higher? Sure you'll be sitting around in your chair all day consuming stuff but you'll have no meaning in your life, nothing to strive for
@@someuser4166 That could be your response to the AI revolution. It wouldn't be mine. If the sole reason you live is to make rich people richer, and that's the only thing that gives your life meaning, then yeah, your life will be without meaning. Personally, however, the thought never even crossed my mind before. I just think of how it would free up time to do my hobbies, formally or informally educate myself on more subjects, be more politically active, spend more time with family and friends, meet new people, discuss hypothetical and practical subjects with more people, help people out, get into art, learn an instrument... The disestablishment of traditional capitalism by a hypothetical AI-led world order would give your life so much more meaning.
We dont pay doctors because we want to pay them but because we want treatments for our diseases. The ultimate point of all this is saving lives not saving jobs. If your job is taken by AI its completely an inevitable event, boo hoo go learn new skills and stop crying like a baby
They're not creative themselves (THEY'RE NOT AI, THEY'RE TEXT GENERATORS) - They take the creativity of millions of humans as input, and just spit that out as output. Why is this fascinating to people?
Yooo divergent thinking is one of my favorite ways that i confuse people in conversation -_- it's embarrassing and really hard to backpedal and explain why I'm not insane sometimes
The thing is that AI like ChatGPT uses creativity data created by humans for its output. I got the Bing chat bit that uses ChatGPT4 and Dall-E for image creation. It can make really good images from a single sentence. However there are limitations and it can create weird and funny images because it gets things wrong. 😂
It doesn't store the creative data, it learns and understands it from the ground up. The fact that it messes few things up, known as hallucination will prove to be its strength in the coming days, it's the thing that will separate it from its completely analytical and machine aspect. Also, in order to scale and manage resources, the bing version of both the chat and image creator is extremely nerfed, dampened version of the actually capable stuff.
@@somethinglikethat2176do you take everything you experience at face value and add it to your memory with the same weight as everything else you've ever experienced? Do you compress every bit of info down into word ratios and lower resolution images in your head? Conversely, if this is the same as a human, do you respect the machine's title as creator and owner of it's output? Or do you consider no human owner of their labor and creations?
If we asked an AI to generate images how many canvassed wouldn’t to have to waste before painting one which isn’t ass? The AI needs human input for both a prompt and the images it’s been fed, it’s not making things on it’s own. AI is a tool used by humans not an entity capable of doing this sort of thing on its own
I work in graphic design. I use AI to help give me inspiration. I use both Midjourney and ChatGPT. I've used it for a while, but the problem is that humans still have to input the right prompt for exactly what you're looking for.
@@AtlasReburdeneditself, ironically. GPT 1-4 were trained from text on the internet, which was almost exclusively written by humans since ChatGPT wasn't publicly available yet. But any future models will be trained on huge amounts of text that ChatGPT itself wrote, and this will be impossible to avoid. It will either diverge into word salad, or converge to the point that everything it writes is extremely similar (even more than now).
@@NastyCupid No, this has nothing to do with whatever political leanings ChatGPT might develop. That is a separate issue. It's that ChatGPT has no real consciousness, intelligence, or creativity, and everything it outputs is a copy of its training data, albeit an indirect one. It boils down to the old cliché that every time you make a copy of a copy, quality degrades.
Divergent thinking is well discussed in a great talk called "Education Paradigms" by Ken Robinson. The odd thing about them is that younger children tend to score higher in these tests than older children/young adults. The education system tends to stop them having this creative thinking over time. So while these AI tools are scoring higher I'd be curious to see who they competed against and also if that merely shows they're less "educated" (whatever you want to considee that as).
Oh great another totally useless study anyone with a basic grasp could of answered.... Machine learning creativity is based on human creativity, it can only give out information we feed it in the first place. So it will always be limited short of a new breakthrough in AI. And please be precise, call it by its name, we are still far from an actual AI, machine learning is but one part, a subfield of artificial intelligence. This is like calling a dentist a doctor. It is not wrong but it leaves out information.... Unsubscribed, I judge my channels like my politicians: if they say BS about a subject I know, they probably say BS about subjects I don't too.
Remember when really good voice synthesis came out and everyone was freaking out saying stuff like "We'll never know if we're talking to a real human on the phone again!!!". So when's the last time a robocall fooled you?
Has anyone mentioned how a wooden pencil can be used as a leg on a wooden toy chair, or as the nose on a woodworked animal, or you can remove the lead and outside paint and use it as a straw (although not advisable)? If it is a mechanical pencil... Hoo buddy, the possibilites become ENDLESS!! (Top is ice cream holder for a Q-tip or mini marshmallow, middle is a straw, handle is used in a slingshot, eraser that's attached to the handle is TOTALLY gonna either be a weapon, toy pogo stick, or even one glued onto each of Ken's hands (after a paint job, of course) and he is now a clawed beast!!
I think this is also a type of question that really plays into the AIs hands. Because it has a habit of generalization. And it's easier for the AI to think of an items generic uses as opposed to create something nobody has thought of before.
AI will absolutely be part of the creative future. Imagine an artist hand-drawing a single frame of a movie, then getting AI to generate a 15 minute scene that matches the animation based on his storyboard. AI is a tool. Right now is it just being used by tools (the people kind) as a substitute for talent. When actual creative people with talent use AI, it will empower them more.
I wouldn't underestimate the potential of LLM. IMO, they are now what search engines were in the 90s. It's just matter of time before they get perfected.
I feel it's like an accelerator in a car. You occasionally need to tap the brakes to stay in control and not floor the accelerator every time you use it. Whether society - or certain Nations around the world - will develop that control is doubtful. But crucial.
The issue isn’t if it beats everyone, the issue is that it beats most, the layman, the average. If AI and automation can do that, be better than average, it will replace most. The Great Depression only had an unemployment rate of 20%, just something to think about.
It's consistent too. The most creative human can maybe brainstorm something in a few days, but AI will make 'lesser' creative ones every single second. Which we can then use to achieve better ideas, make AI better.
I saw a video once that talked about how chatgtp mainly adds a floor to society. Everyone who can write or think or do math beneath average can now function at average. Which is huge for them, but it has little effect on the leading minds in those fields. As a student of UBI I'm always happy to hear about floors added to society to keep people from falling out.
Add on to this those solutions are 'trained' Chatbots just collate and regurgitate information they already know. They didn't solve a problem, they simply remembered human input solutions.
I feel like i could predict exactly every single word of this video before you said it, I wouldn't be surprised if the bots are more creative than us, we're pretty bloody predictable.
People keep forgetting that this is only the beginning of the 2nd wave of AI, or 3rd, depending on how you're counting it. Just think about the computers and programs we had in 1970, compared to what we have today. That exact same trend of development is currently occurring in the filed of AI. Watershed papers are being published on nearly a weekly/monthly basis, and that pace is only accelerating. Like 2 minute papers says: Don't think about where we _are,_ think about where we _will_ be in the next 5-10 years. Hardly anyone is truly prepared for what's coming down the pipe.
Peace. Thank you! Have always looked at every day objects as they became obsolete or superfluous for normal uses. Kept two packing wheels for a couple of decades. Was finally able to use them as additional height supports for two 'legs' of a laundry tub. A clean, cardboard french fry tray was bottom of, 'boat' for a crafting project. And so forth... Peace.