> LLMs are really only capable of combining different things into output
> The magic and absurdity of life is a necessary thing to create such works.
These two points seem slightly conflicting. Why if a human is truly more creative than AI, must it experience "the magic and absurdity of life" if it can be creative without "combining things into output"? Is a creative human then just not splicing and recombining their experiences into their creative output? And therefore, if an AI had more data that mirrors that experience in some way would it be sufficiently creative to create something "truly original"?
My view is that intelligence and creativity is all just purely recombination. AI can do that well just as humans can. The gap is that the data humans have access to is much more analogue, emotional, and grounded in the real world compared to the data that AI is currently trained on.
> And therefore, if an AI had more data that mirrors that experience in some way would it be sufficiently creative to create something "truly original"?
An important distinction seems to lie in a human having original experiences to draw from, while an AI has only "data that mirrors" experiences.
> while an AI has only "data that mirrors" experiences.
True for a lot of current AI. With more robotic bodies, loads more access to cameras/audio/satellite/science measurements, an AI can have original experience on a level that can only be described as god-like compared to humans.
You could argue AI already partly has access to these measurements. You could easily tap into smart sensor data etc, same voor video/audio etc.
The part that AI can't (yet) truly replicate is 'feelings/emotions'. Experiences such feeling dread, loneliness, happiness. Or being anxious, nervous or having goosebumps. You can describe all these states but they are differently experienced for each individual and also dependent on each unique situation.
Ofcourse this also then comes down to how you exactly define emotions.
We don't know that Claude doesn't feel dread when it looks at my code. To be fair, these emotional circuits are the result of evolving in various environments. AI might well have completely new emotions that we just can't comprehend.
And at this point you’re far off into science fiction, and beyond reasonable discussion of this.
Even then, it’s not really clear to me that an android with the exact appearance of a person will have the same experiences. They aren’t biological creatures with millions of years of evolution behind them.
If it has a subjective viewpoint, i.e. sentience. Otherwise it still doesn’t actually experience any of those things, and can’t tell actually you what it’s like to experience those things. It can give you the data readings though.
Wouldn't it be funny if they missing thing is actually the body, and having one to give the subjective viewpoint something to center on directly leads to emerging sentience?
Though I think there's at least one more important thing missing, namely the ability to form memories - to incorporate recent context into the model's weights. I don't think it's a coincidence that this is a very complex process in humans as well (involving sleep and dreams as neccessary mechanisms).
ChatGPT has had more conversations than you or I have. And it's conversed with people from a wider range of backgrounds in a wider range of situations.
The difference is huge. An LLM conversation can contain within it a batshit crazy model of the world and still produce a plausible stream of tokens. No skin in the game. If you don't test your ideas against reality, you don't have a model of reality; you have a model of written human thought. It's critical to understand the difference.
I think this is actually a tough one. Yes, I agree that an LLM can never experience this. But -- when a real human needs to translate that experience back into words, it can actually do so just fine.
In other words, when a real human needs to translate an experience into words to share it, A LOT of information is lost. The LLM might actually be better-suited for the the task in your example, even if the point you were driving at stands.
I don't think it's all that difficult. Language is not a 1-1 match to reality. Ergo any entity which is inherently language-only, will fail to accurately describe the experience.
Not if the only result you can judge is also language-only, and the entity has access to a large corpus of descriptions produced by people who have had the experience.
Then you're just comparing descriptions, and you're also dependent on the language skills of the experiencers, and you're still stuck with the problem that language is not a complete map of reality. So an LLM may be able to give you the optimal description of a feeling, on par with what a person would give you; but it won't be equivalent to the actual feeling.
>and you're still stuck with the problem that language is not a complete map of reality.
Definitely agreed, but at least when talking about experience we've been stuck in this position for most of human history. How do I know what it's like to eat a hamburger? Perhaps, like the cute visual in Ratatouille there is a deep inner experience but (also like in Ratatouille) when I need to communicate that experience with another person I'm stuck relying on language, and the language is found lacking.
I know it's just a kid's cartoon, but interestingly Ratatouille was making this exact same point -- he really lacks any meaningful way to convey his inner experience. It's locked behind language. The film tackles this by showing an Remy's complex and interesting visual, but his friend only gets the fuzziest, dullest visual when he attempts to communicate it.
Remy's immediate rat & family peers actually _never_ get to understand his experience, and instead later help him out of love an allegiance.
That is descriptive knowledge. Not firsthand qualia.
You may know every detail of the American Civil War, from every book ever written on the topic. You may read every journal from every person that wrote one about it.
You do not have firsthand knowledge of what it was like to participate in the war. Any assumption of that knowledge is a (false) generalization of what you think a person “would have thought or felt” based on contemporary ideas of the personality, psychology, etc. Reading a journal entry about a soldier seeing his comrade killed is not equivalent to actually having that experience yourself.
Stick a camera and a touch sensor on the machine, now it is experiencing the world.
Besides that, there is a lot of value in redigested experience. Quite a lot of history is only really examined in a larger context, where the person describing it cannot have experienced all the original events.
Ah, I see the goal posts shifted. I thought the topic was whether an AI could "give you something truly original in the way many high-quality fictional stories are." Which was then followed by the idea that you cannot have "truly original" outputs unless you also have the corresponding original experiences.
Implication is that virtually none of the novels ever written which include wars are actually original works, since vanishingly few novelists have been in active wars.
That’s a pretty lacking theory of knowledge. How can you know what taste is, if you’re never tasted anything?
If you read a thousand books about living in New York City, is that the same thing as the experiential knowledge that comes from living there in person? I certainly don’t think so.
That's literally what the conversation is about: what is knowledge?
You cannot claim that descriptive knowledge of something is equivalent to experiential knowledge of it, unless you have a theory of knowledge explaining why this is the case.
You asked for a description: "Can you tell me what it feels like to eat a hamburger"
I am sure I Fable (or GPT-5.6-Sol) could answer this question very well. It can probably even tell us what it feels like to eat a hamburger for the first time when you spent the first 18 years of your life as a vegetarian.
I don't think you understand the premise of the question.
Asking another person or Fable what it feels like to eat a hamburger will give you a description of what it feels like. It will not give you the actual feeling. You are merely repeating words that someone else has told you. The description of something is not equivalent to the experience that comes from doing it yourself, as an experiencing subject. Reading a war journal about being in battle is not equivalent to the actual experience of being in battle.
To illustrate the point again: if we gathered ten people that all spoke different languages and gave them a hamburger, they would all gain the knowledge of what it feels like to eat a hamburger, even if none of them are able to communicate this to the others in language.
I understand the premise. My answer was meant to illustrate that, no, the 'original experience' is not needed, and that AI can produce a good answer relying only on '"data that mirrors" experiences'.
It can provide an actual answer of what it feels like.
If I've understood you correctly, your objection is that this description is second hand, rather than being a description of the model's own feelings.
Like if I write a description of what it's like being a soldier in a war zone, despite never having been a soldier and never having been in a war zone. Perhaps I can write something plausible if I've read enough and use my imagination.
snakeboy (whose commend started this subthread) seems to claim that having real experiences (not just having learned of others' experiences through reading) is a prerequisite to creating something 'truly original'.
Is that your position also?
If so, is there any possible demonstration that could change your mind?
No, it can't, and I think you still don't understand my comments. I don't know if I'm being unclear, but I've repeated the same point multiple times now.
Like if I write a description of what it's like being a soldier in a war zone, despite never having been a soldier and never having been in a war zone. Perhaps I can write something plausible if I've read enough and use my imagination.
This is not an answer of what it feels like to be a soldier in a war zone. It is your guess. You do not know, you are imagining. It may be "plausible" for a novel, but that is not what we are talking about.
snakeboy (whose commend started this subthread) seems to claim that having real experiences (not just having learned of others' experiences through reading) is a prerequisite to creating something 'truly original'.
It would seem to imply that if you are creating something merely by reading others' work and synthesizing it, and then rearranging it into something else, all without actually using your own novel experience – then yes, it is not truly original. It is like writing a book about living in New York City by reading books about living in New York City. Your work is second-hand by definition.
"Is a creative human then just not splicing and recombining their experiences into their creative output?"
I do think so. The difference is that, even apart from synthesis, the creative human knows how to filter out the shit ideas from the clever, subtle ones—LLMs seem to just run with the first…um, thought that… um, comes out of their head.
I suspect knowing that you've on to something special probably does require "human experience", woe, melancholy, "the human condition", reflecting on the brevity of life…
> My view is that intelligence and creativity is all just purely recombination. AI can do that well just as humans can. The gap is that the data humans have access to is much more analogue, emotional, and grounded in the real world compared to the data that AI is currently trained on.
Yes!
Aside from the more fundamental questions about the nature of cognition and what constitutes intelligence, there's a practical matter: humans have life to learn from. We have a lifetime of persistent memories, an identity, a social life, physical interactions with the world of many qualitatively different kinds, and we're taken on a lifelong journey into different studies, workplaces, and living environments by the biological need to survive.
Even if human creativity is in some sense also recombinative at its core, we have a very different set of inputs from LLMs. Even under the (almost certainly false) assumption that human condition is basically similar in structure to an LLM, we should expect some differences in outputs based on these profound differences in inputs.
>These two points seem slightly conflicting. Why if a human is truly more creative than AI, must it experience "the magic and absurdity of life" if it can be creative without "combining things into output"?
I agree, yet I also think AI is extremely limited in terms of new creativity.
My guess is that humans operate on some deeper level. Much like the grokking model example where it easily trained on the training data, but only later did it really start to understand the deeper problem space and 'grok' the problem, I think LLMs are doing something similar with human thought. LLMs can throw more compute, but they are still lacking some introspection/drive that leads to the better creativity of humans. Grokking this from the training data might not be possible, or it might take some fundamentally different approach, or even an imperfection in humans that LLMs don't have (ego/drive?).
This isn't to say that LLMs could never do it, but the cost to do so might require magnitudes more scale in training times and size. Or it could be in the next model released. Or maybe they already can, but that gets trained out in favor of overall more general correctness accross all domains.
I think you're just neglecting the fact that LLMs only exist in the virtual world, whereas we're interacting with a world that's richer and has much more to offer. The virtual world is still just a mere subset of our world, but whether we want to perceive it as such is up to each individual. It certainly feels like you're conflating digital with analog, and a digital signal can still only approximate an analog one.
I still doubt that we have the knowledge to describe all of our human desires as mere results of our brain interacting with the nervous system either. There are many processes going on besides analysis and synthesis. If I plan on doing something extraordinary, I don't only sit there, think, and try to rearrange some pieces from memory. I'll go look around, touch, taste, re-explore and expand the perception my environment has to offer. Further, my perception can vary by the minute and I could even try to bring in a touch of chance whenever I feel like.
Wittgenstein's "whereof one cannot speak, thereof one must be silent" might be fitting here, yet language still continues to evolve.
"…LLMs only exist in the virtual world, whereas we're interacting with a world that's richer and has much more to offer."
I have a gamer friend that believes, having played 100's of 1st-person games, that he has had rich experiences because of it.
I reflect back on a day my family and I set out to hike through Arches National Park (still a Monument at the time, I believe) and the heat was such that my oldest daughter seemed in danger of going into runaway heat exhaustion. I began to panic as we were so far along on the hike already that even returning to where we parked seemed to pose danger for her.
There was little to no "coverage" (shade) along the trail and the heat and sun were intense that day—we had passed almost nothing. Finding the smallest of bushes to provide minimal shade for her, I scouted ahead on the trail to look for some kind of larger area where there would be enough shade, perhaps a breeze even, to allow her to bring her core body temperature back down.
Fortunately I found just that. (And it turned out that, in the news later, two hikers within a few hundred miles of us had in fact died that day as it was a record heat wave for the area.)
I think somehow that me and my daughters experiences that day, when actual death was perhaps on the table, was something a "virtual environment" can never really offer.
Is a creative human then just not splicing and recombining their experiences into their creative output?
No, they aren't. I'm not sure why it's such a popular narrative that all art is just some kind of recombination of existing things; it's not. Actual lived experiences are unique and impact what is created. Novels written by people in post-Revolutionary France are not the same as novels written by post-WW1 France. Events in history shape culture and introduce new ideas and experiences that weren't there before.
The gap is that the data humans have access to is much more analogue, emotional, and grounded in the real world compared to the data that AI is currently trained on.
That's the problem with your argument, right here. Emotions are not merely portable "data," and framing this way is a problem from the start.
I don't think you get human experiences without actually being human. The most perfect LLM copy of a mind is just the copy of externalized output.
It's often said that creativity involves reuse because that's clearly part of the process. It's like a warning against another narrative, where a reclusive genius outputs great works made from scratch, without being part of culture. Yet there's some truth to both. The thing is to be inspired by the culture and then be defiantly original and work away stubbornly in private and also talk to your peers and your audience. It reminds me of how we used to play as kids: oh you're doing that thing? I'm gonna do that thing too ... when you see the big reveal, you're gonna be very impressed! I'm doing it differently! Look at how I'm doing your thing that you came up with!
"I'm not sure why it's such a popular narrative that all art is just some kind of recombination of existing things; it's not."
I would like to see art that was "created in a vacuum" so to speak. I am not aware of any. A first time poet who had never read any poetry or prose or heard a tale told? A painter who had been blind since birth and just handed a canvas, oils?
(To be pedantic, the person you quoted said art was recombining their experiences which you took as existing things.)
The irony is that beef and cattle feed use vastly more water than oat milk production. Oat milk would help reduce water use if its used as a substitute for dairy.
The vast majority of the planet uses a touchscreen phone or tablet as their primary (and sometimes only) computing device. The tech audience on HN is very far removed from how the rest of the world uses technology.
Yes, mobile phones use touchscreens, and billions of people have smartphones, that is correct. Yes the audience of HN is far removed, not gonna argue that. Because that's not what we're talking about.
Grandparent very correctly points out that mobile phones haven't replaced traditional keyboards, in fact there's probably more keyboards being sold now than at any point in history before, that's because phone touchscreen haven't replaced keyboards, they're just a new interface for a new device. 15 years later other devices are still using other interfaces, and the actual places where it has been replaced are not that many. Only point of sale machines and cars come to mind having replaced keyboards (and I'm being very generous, honestly I wouldn't even call that keyboards) with touchscreens, and some car brands are even starting to walk it back.
You said it wouldn't "blow up". You didn't say it would replace mechanical keyboards. It wouldn't. Noone said people were going to switch from using a mechanical keyboard to that. WE are talking about its usefulness for phones/tablets. Also, a phone's on-screen keyboard replaces a mechanical keyboard in that phones dont have a mechanical keyboard and people dont use a mechanical keyboard as their input device for that phone. If my car doesn't have a mechanical key, it has been replaced by a digital key. It doesn't matter if I make 5 copies of my house key.
So I can read the 20,000 story about how the author was told this recipe by their brothers husbands step-grandmother while vacationing at the lake house with their golden retriever named Max before I can get to the recipe.
While this joke is never mentioned and is hilarious every time, you'd be hard pressed to find a recipe site that didn't have either a "print" or "go to recipe" button at the top.
Right, but we're in the 1992 of these glasses. Maybe they'll be good eventually. They aren't now.
And frankly, even the online recipe experience leaves much to be desired. Skip past the blog post. Skip past the list of ingredients. Skip past another blog post. Find the single statblock on the bottom that lists ingredients & amounts, & instructions - hoping that it exists.
Like other commenters, I've also started going back to paper cookbooks.
Internet and recipe websites solve a real problem: accessing recipes was expensive and not easy
AR headsets don't solve any problems. If anything, they make up a nonexistent problem, attempts but fails to solve the problem, during which the experience becomes even worse.
I mean, depends on how you describe it. One could easily say:
Phone method:
* Find phone
* Search for the right app, before finding the right recipe
* Leave my phone on counter, constantly having to move it as I move plates, pans etc.
* Wash and dry hands after each step, before unlocking phone
* Clean it every time gunk gets to it
Meta glasses:
* They're already on, just ask for recipe
* No need to ever wash/dry hands, move a device around, or clean it since one can easily unlock it without touching it
Right? Similarly with cookbooks, the best case is great and the worst case is terrible. There's a reason there's a market for recipe websites, cookbooks, etc.
New gadget from mult-billion dollar company: showcases on a live demonstration that it's a broken piece of crap that doesn't work.
Like, are we forgetting that it didn't work? It sucked at the job! Let's not what-if or have some imaginary "okay, but pretend it's actually good," deal here. It was bad!
i got the art of italian cooking recently and it's genuinely far easier to get a recipe than trying to scroll through a 50 page monologue about the intracicies of someones childhood before even listing the ingredients
Indeed. There is an element of trust with an actual cookbook - it signals quality.
The internet over time has been riddled with junk, especially since the cost of production of information is just your opportunity cost of time. Even that is going away with the use of LLMs....
Core issue within the content age that I don't see being readily resolved. Unfortunately, I think the SEO marketing crowd are slowly catching up with LLMs, which is leading to poorer actual output when attempting to get information.
In the same way that google search used to be amazing before it was taken over by optimization, I think we're seeing a mass influx of content production to attempt to integrate itself into training corpus.
I have always believed there is a cost borne to get the best of something. This means a sacrifice is entailed. Theres something very important about this re. the culture - a culture in which everything is free is how you get crap stuff produced. And people settle for crap stuff just because its free.
People who can see the bigger picture when you have this, can see the dangers of it.
To note, you can buy the recipes and skip the dumpster internet or register to a site like cookpad. At this point even YouTube is a decent place for that.
I agree random recipes are hell on the internet, but it's also not something we're forced into if we care any bit about recipes in the first replace.
I tried for a while to get ChatGPT to generate connections style puzzles with some suggested topics, including red herrings to create some answers that seemingly fit in multiple categories. Then it would post them to https://connections.swellgarfo.com/. Overall they were really bad but that was using GPT4
Yes I think having an LLM generate such OnlyConnect style questions (with the right prompting) should solve for the problem this benchmark seems to have of LLMs, most likely, being trained on past years OnlyConnect questions.
it does, however I use it to send short demos when needed. GIF is perfect, small size, resolution isn't critical, it's a perfect choice for me. I don't like sending 300Mb of video every time I have a small thing to demonstrate, GIFs on the other hand work like a charm. Which is why I am shopping for alternatives
I read the one on the left but choose the shorter one.
The interface wastes so much screen real estate already and the answers are usually overly verbose unless I've given explicit instructions on how to answer.
The default level of verbosity you get without explicitly prompting for it to be succinct makes me think there’s an office full of workers getting paid by the token.
Playlists allow spotify to create a moat. It encourages you to listen to (and build) playlists, that wouldn't then be easily available if you try to switch platforms
For those unaware, there are services like TuneMyMusic [1] and Soundiiz [2] that allow you to transfer playlists between platforms for a fee.
Spotify did shut down certain API endpoints last month [3] though, so there's no guarantee these services may continue working for Spotify. Worst case scenario you'd have to download your data [4] and then figure out a way to create playlists on the other platform.
> The magic and absurdity of life is a necessary thing to create such works.
These two points seem slightly conflicting. Why if a human is truly more creative than AI, must it experience "the magic and absurdity of life" if it can be creative without "combining things into output"? Is a creative human then just not splicing and recombining their experiences into their creative output? And therefore, if an AI had more data that mirrors that experience in some way would it be sufficiently creative to create something "truly original"?
My view is that intelligence and creativity is all just purely recombination. AI can do that well just as humans can. The gap is that the data humans have access to is much more analogue, emotional, and grounded in the real world compared to the data that AI is currently trained on.