[HN Gopher] Thoughts on thinking
___________________________________________________________________
Thoughts on thinking
Author : bradgessler
Score : 209 points
Date : 2025-05-16 19:09 UTC (3 hours ago)
(HTM) web link (dcurt.is)
(TXT) w3m dump (dcurt.is)
| bradgessler wrote:
| I keep going back and forth on this feeling, but lately I find
| myself thinking "F it, I'm going to do what I'm going to do that
| interests me".
|
| Today I'm working on doing the unthinkable in an AI-world:
| putting together a video course that teaches developers how to
| use Phlex components in Rails projects and selling it for a few
| hundred bucks.
|
| One way of thinking about AI is that it puts so much new
| information in front of people that they're going to need help
| from people known to have experience to navigate it all and
| curate it. Maybe that will become more valuable?
|
| Who knows. That's the worst part at this moment in time--nobody
| really knows the depths or limits of it all. We'll see
| breakthroughs in some areas, and others not.
| don_neufeld wrote:
| Completely agree.
|
| From all of my observations, the impact of LLMs on human thought
| quality appears largely corrosive.
|
| I'm very glad my kid's school has _hardcore_ banned them. In some
| class they only allow students to turn in work that was done in
| class, under the direct observation of the teacher. There has
| also been a significant increase in "on paper" work vs work done
| on computer.
|
| Lest you wonder "what does this guy know anyways?", I'll share
| that I grew up in a household where both parents were professors
| of education.
|
| Understanding the effectiveness of different methods of learning
| (my dad literally taught Science Methods) were a frequent topic.
| Active learning (creating things using what you're learning
| about) is so much more effective than passive, reception oriented
| methods. I think LLMs largely are supporting the latter.
| zdragnar wrote:
| Anyone who has learned a second language can tell you that you
| aren't proficient just by memorizing vocabulary and grammar.
| Having a conversation and forming sentences on the fly just
| feels different- either as a different skill or using a
| different part of the brain.
|
| I also don't think the nature of LLMs being a negative crutch
| is new knowledge per se; when I was in school, calculus class
| required a graphing calculator but the higher end models (TI-92
| etc) that had symbolic equation solvers were also banned, for
| exactly the same reason. Having something that can give an
| answer for you fundamentally undermines the value of the
| exercise in the first place, and cripples your growth while you
| use it.
| skydhash wrote:
| Same with drawing which is easy to teach, but hard to master
| because of the coordination between eyes and hand. You can
| trace a photograph, but that just bypass the whole point and
| you don't exercise any of the knowledge.
| flysand7 wrote:
| Another case in point is that memorizing vocabularies and
| grammar, although could seem like an efficient way to learn a
| language, is incredibly unrewarding. I've been learning
| japanese from scratch, using only real speech to absorb new
| words, without using dictionaries and anything else much. The
| first feeling of reward came immediately when I learned that
| "arigatou" means thanks (although I terribly misheard how the
| word sounded, but hey, at least I heard it). Then after 6
| month, when I could catch and understand some simple phrases.
| After 6-7 years I can understand about 80% of any given
| speech, which is still far, but I gotta say it was a good
| experience.
|
| With LLM's giving you ready-made answers I feel like it's the
| same. It's not as rewarding because you haven't obtained the
| answer yourself. Although it did feel rewarding when I was
| interrogating an LLM about how CSRF works and it said I asked
| a great question when I asked whether it only applies to
| forms because it seems like fetch has a different kind of
| browser protection.
| avaika wrote:
| This reminds me how back in my school days I was not allowed to
| use the internet to prepare research on some random topics (e
| g. history essay). It was the late 90s when the internet
| started to spread. Anyway teachers forced us to use offline
| libraries only.
|
| Later in the university I was studying engineering. And we were
| forced to prepare all the technical drawings manually in the
| first year of study. Like literally with pencil and ruler. Even
| though computer graphics were widely used and we're de facto
| standard.
|
| Personally I don't believe hardcore ban will help with any sort
| of thing. It won't stop the progress either. It's much better
| to help people learn how to use things instead of forcing them
| to deal with "old school" stuff only.
| baxtr wrote:
| Yes agreed.
|
| Let's not teach people on mass how to grow potatoes just
| because we don't want them to eat fries.
| johnisgood wrote:
| You can learn a lot from LLMs though, same with, say,
| Wikipedia. You need curiosity. You need the desire to learn. If
| you do not have it, then of course you will get nowhere, LLMs
| or no LLMs.
| azinman2 wrote:
| Never underestimate laziness, or willingness to take
| something 80% as good for 1% of the work.
|
| So most are not curious. So what do you do for them?
| johnisgood wrote:
| [delayed]
| snackernews wrote:
| Can you learn a lot? Or do you get instant answers to every
| question without learning anything, as OP suggests?
| johnisgood wrote:
| [delayed]
| taylorallred wrote:
| Among the many ways that AI causes me existential angst, you've
| reminded me of another one. That is, the fact that AI pushes you
| towards the most average thoughts. It makes sense, given the
| technology. This scares me because creative thought happens at
| the very edge. When you get stuck on a problem, like you
| mentioned, you're on the cusp of something novel that will at the
| very least grow you as a person. The temptation to use AI could
| rob you of the novelty in favor what has already been done.
| worldsayshi wrote:
| I feel there's a glaring counter point to this. I have never
| felt more compelled to try out whatever coding idea that pops
| into my head. I can make Claude write a poc in seconds to make
| the idea more concrete. And I can write into a good enough tool
| in a few afternoons. Before this all those ideas would just
| never materialize.
|
| I mean I get the existential angst though. There's a lot of
| uncertainty about where all this is heading. But, and this is
| really a tangent, I feel that the direction of it all is in the
| intersection between politics, technology and human nature. I
| feel like "we the people" leave walkover to powerful actors if
| we do not use these new powerful tools in service of the
| people. For one - to enable new ways to coordinate and
| organise.
| kdamica wrote:
| Wow, one day after this post: https://blog.ayjay.org/two-
| quotations-on-the-brief-dream-of-...
| paulorlando wrote:
| I've noticed something like this as well. A suggestion is to
| write/build for no one but yourself. Really no one but yourself.
|
| Some of my best writing came during the time that I didn't try to
| publicize the content. I didn't even put my name on it. But doing
| that and staying interested enough to spend the hours to think
| and write and build takes a strange discipline. Easy for me to
| say as I don't know that I've had it myself.
|
| Another way to think about it: Does AI turn you into Garry
| Kasparov (who kept playing chess as AI beat him) or Lee Sedol
| (who, at least for now, has retired from Go)?
|
| If there's no way through this time, I'll just have to
| occasionally smooth out the crinkled digital copies of my past
| thoughts and sigh wistfully. But I don't think it's the end.
| ge96 wrote:
| Yeah there is the personal passion and then the points/likes
| driven which sucks the joy out
|
| I experienced this when I was younger with my rc planes, I
| joined some forum and I felt like everything I did had to be
| posted/liked to have value. I'd post designs/fantasy and get
| the likes then lose interest/not actually do it after I got the
| ego bump
| paulorlando wrote:
| I think that's why some writers refuse to talk about their
| work in progress. If they do, it saps the life out of it.
| Your ego bump example kind of fits into that.
| wjholden wrote:
| I just wrote a paper a few days ago arguing that "manual
| thinking" is going to become a rare and valuable skill in the
| future. When you look around you, everyone is finding ways to be
| better using AI, and they're all finding amazing successes - but
| we're also unsure about the downsides. I hedge that my advantage
| in ten years will be that I chose not to do what everyone else
| did. I might regret it, we will see.
| oytis wrote:
| If AI is going to be as economically efficient as promised,
| there is going to be no way to avoid using it altogether. So
| the trick will be to keep your thinking skills functional,
| while still using AI for speedup. Like focus in the age of
| Internet is a rare skill, but not using Internet is not an
| option either.
| fkfyshroglk wrote:
| Not all processes are the same, though. I strongly suspect
| any efficiency improvements will come in processes that
| didn't require much "thinking" to begin with. I use it daily,
| but mostly as a way to essentially type faster--I can read
| much faster than I can type, so I mostly validate and correct
| the autocomplete. All of my efforts to get it to produce
| trustworthy output beyond this seem to trail behind the
| efficiency of just searching the internet.
|
| Granted, I'm blessed to not have much busywork; if I need to
| produce corporate docs or listicles AI would be a massive
| boon. But I also suspect AI will be used to digest these
| things back into small bullet points.
| MarcelOlsz wrote:
| Have not resonated with an article for a long time until this
| one.
| paintboard3 wrote:
| I've been finding a lot of fulfillment in using AI to assist with
| things that are (for now) outside of the scope of one-shot AI.
| For example, when working on projects that require physical
| assembly or hands-on work, AI feels more like a superpower than a
| crutch, and it enables me to tackle projects that I wouldn't have
| touched otherwise. In my case, this was applied to physical
| building, electronics, and multimedia projects that rely on
| simple code that are outside of my domain of expertise.
|
| The core takeaway for me is that if you have the desire to
| stretch your scope as wide as possible, you can get things done
| in a fun way with reduced friction, and still feel like your
| physical being is what made the project happen. Often this means
| doing something that is either multidisciplinary or outside of
| the scope of just being behind a computer screen, which isn't
| everyone's desire and that's okay, too.
| sanderjd wrote:
| Yeah I haven't found the right language for this yet, but it's
| something like: I'm happy and optimistic about LLMs when I'm
| the one _doing_ something, and more anxious about them when I
| 'm _supporting_ someone else in doing something. Or: It makes
| me more excited to focus on _ends_ , and less excited to focus
| on _means_.
|
| Like, in the recent past, someone who wanted to achieve some
| goal with software would either need to learn a bunch of stuff
| about software development, or would need to hire someone like
| me to bring their idea to life. But now, they can get a lot
| further on their own, with the support of these new tools.
|
| I think that's _good_ , but it's also nerve-wracking from an
| employment perspective. But my ultimate conclusion is that I
| want to work closer to the ends rather than the means.
| apsurd wrote:
| Interesting, I just replied to this post recommending the
| exact opposite: to focus on means vs ends.
|
| The post laments how everything is useless when any
| conceivable "end state" a human can do will be inferior to
| what LLMs can do.
|
| So an honest attention toward the means of how something
| comes about--the process of the thinking vs the polished
| great thought--is what life is made of.
|
| Another comment talks about hand-made bread. People do it and
| enjoy it even though "making bread is a solved problem".
| sanderjd wrote:
| I saw that and thought it was an interesting dichotomy.
|
| I think a way to square the circle is to recognize that
| people have different goals at different times. As a person
| with a family who is not independently wealthy, I care a
| lot about being economically productive. But I also
| separately care about the joy of creation.
|
| If my goal in making a loaf of bread is economic
| productivity, I will be happy if I have a robot available
| that helps me do that quickly. But if my goal is to find
| joy in the act of creation, I will not use that robot
| because it would not achieve that goal.
|
| I do still find joy in the act of creating software, but
| that was already dwindling long before chatgpt launched,
| and mostly what I'm doing with computers is with the goal
| of economic productivity.
|
| But yeah I'll probably still create software just for the
| joy of it from time to time in the future, and I'm unlikely
| to use AIs for those projects!
|
| But at work, I'm gonna be directing my efforts toward
| taking advantage of the tools available to create useful
| things efficiently.
| apsurd wrote:
| ooh I like this take. We can change the framing. In the
| frame of one's livelihood we need to be concerned with
| economic productively, philosophy be damned.
| WillAdams wrote:
| This is only a problem if one is writing/thinking on things which
| have already been written about without creating a new/novel
| approach in one's writing.
|
| An AI is _not_ going to get awarded a PhD, since by definition,
| such are earned by extending the boundaries of human knowledge:
|
| https://matt.might.net/articles/phd-school-in-pictures/
|
| So rather than accept that an LLM has been trained on whatever it
| is you wish to write, write something which it will need to be
| trained on.
| noiv wrote:
| I agree, frustration steps in earlier than before because AIs
| tell you very eagerly that unique thought is already part of
| its training data. Sometimes I wish one could put an AI on
| drugs and filter out some hallucinations that'll become main
| stream next week.
| jonplackett wrote:
| I'm so so glad at the end it said "Written entirely by a human"
| because all the way through I was thinking - absolutely no way
| does AI come up with great ideas or insights and it definitely
| would not write this article- it holds together and flows too
| well (if this turns out to be the joke I'll admit we're all now
| fucked)
|
| Having said that I am very worried about kids growing up with AI
| and it stunting their critical thinking before it begins - but as
| of right this moment AI is extremely sub par at genuinely good
| ideas or writing.
|
| It's an amazing and useful tool I use all the time though and
| would struggle to be without.
| tutanosh wrote:
| I used to feel the same way about AI, but my perspective has
| completely changed.
|
| The key is to treat AI as a tool, not as a magic wand that will
| do everything for you.
|
| Even if AI could handle every task, leaning on it that way would
| mean surrendering control of your own life--and that's never
| healthy.
|
| What works for me is keeping responsibility for the big picture--
| what I want to achieve and how all the pieces fit together--while
| using AI for well-defined tasks. That way I stay fully in
| control, and it's a lot more fun this way too.
| bachittle wrote:
| The article mentions that spell and grammar checking AI was used
| to help form the article. I think there is a spectrum here, with
| spell and grammar checking on one end, and the fears the article
| mentions on the other end (AI replacing our necessity to think).
| If we had a dial to manually adjust what AI works on, this may
| help solve the problems mentioned here. The issue is that all the
| AI companies are trying too hard to achieve AGI, and thus making
| the interfaces general and without controls like this.
| perplex wrote:
| I don't think LLMs replace thinking, but rather elevate it. When
| I use an LLM, I'm still doing the intellectual work, but I'm
| freed from the mechanics of writing. It's similar to programming
| in C instead of assembly: I'm operating at a higher level of
| abstraction, focusing more on what I want to say than how to say
| it.
| cratermoon wrote:
| The writing _is_ the work, though. The words on paper (or
| wherever) are the end product, but they are not the point. See
| chapter 5 of Ahrens, Sonke. 2017. _How to take smart notes: one
| simple technique to boost writing, learning and thinking - for
| students, academics and nonfiction book writers._ , for advice
| on how writing the ideas in your own words is the primary task
| and improves not only writing, but all intellectual skills,
| including reading and thinking. C. Wright Mills in his 1952
| essay, ""On Intellectual Craftsmanship" says much the same
| thing. Stating the ideas _in your own words_ is thinking.
| i1856511 wrote:
| If you do not know how to say something, you don't know how to
| say it.
| FeteCommuniste wrote:
| When I microwave a frozen meal for dinner, I'm still a chef,
| but I'm freed from the mechanics of preparing and assembling
| ingredients to form a dish.
| jebarker wrote:
| > nothing I make organically can compete with what AI already
| produces--or soon will.
|
| No LLM can ever express your unique human experience (or even
| speak from experience), so on that axis of competition you win by
| default.
|
| Regurgitating facts and the mean opinion on topics is no
| replacement for the thoughts of a unique human. The idea that
| you're competing with AI on some absolute scale of the quality of
| your thought is a sad way to live.
| steamrolled wrote:
| More generally, prior to LLMs, you were competing with 8
| billion people alive (plus all of our notable dead). Any novel
| you could write probably had some precedent. Any personal story
| you could tell probably happened to someone else too. Any skill
| you wanted to develop, there probably was another person more
| capable of doing the same.
|
| It was never a useful metric to begin with. If your life goal
| is to be #1 on the planet, the odds are not in your favor. And
| if you get there, it's almost certainly going to be
| unfulfilling. Who is the #1 Java programmer in the world? The
| #1 topologist? Do they get a lot of recognition and love?
| harrison_clarke wrote:
| a fun thing about having a high-dimensional fitness function
| is that it's pretty easy to not be strictly worse than anyone
| martin-t wrote:
| This resonates with me deeply.
|
| I used to write open source a lot but lately, I don't see the
| point. Not because I think LLMs can produce novel code as good
| code as me or will be able to in the near future. But because any
| time I come up with a new solution to something, it will be
| stolen and used without my permission, without giving me credit
| or without giving users the rights I give them. And it will be
| mangled just enough that I can't prove anything.
|
| Large corporations were so anal about copyright that people who
| ever saw Microsoft's code were forbidden from contributing to
| FOSS alternatives like wine. But only as long as copyright suited
| them. Now abolishing copyright promises the C-suite even bigger
| rewards by getting rid of those pesky expensive programmers, if
| only they could just steal enough code to mix and match it with
| enough plausible deniability.
|
| And so even though _in principle_ anybody using my AGPL code or
| anything that incorporates my AGPL code has the right to inspect
| and modify said code; yet now tine fractions of my AGPL code now
| have millions or potentially billions of users but nobody knows
| and nobody has the right to do anything about it.
|
| And those who benefit the most are those who already have more
| money than they can spend.
| Trasmatta wrote:
| Agreed. I keep seeing posts where people claim the output is all
| that really matters (particularly with code), and I think that's
| missing something deeply fundamental about being human.
| boznz wrote:
| Think Bigger. As the LLM's capability expands use it to push
| yourself outside your comfort zone every now and then. I have
| done some great fun projects recently I would have never thought
| of tackling before
| cess11 wrote:
| This person should probably read pre-Internet books, discover or
| rediscover that the bar for passable expression in text is very
| low compared to what it was.
|
| Most of that 'corpus' isn't even on the Internet so it is wholly
| unknown to our "AI" masters.
| tibbar wrote:
| AI is far better in style than in substance. That is, an AI-
| written solution will have all the trappings of expertise and
| deep thought, but frankly is usually at best mediocre. It's sort
| of like when you hear someone make a very eloquent point and your
| instinct is to think "wow, that's the final word on the subject,
| then!" ... and then someone who _actually understands_ the
| subject points out the flaw in the elegant argument, and it falls
| down like a house of cards. So, at least for now, don 't be
| fooled! Develop expertise, be the person who really understands
| stuff.
| uludag wrote:
| What happens to conversation in this case? When groups of people
| are trained to use LLMs as a crutch for thinking, what happens
| when people get together and need to discuss something. I feel
| like the averageness of the thinking would get compounded so that
| the conversation as a whole becomes nothing but a staging ground
| for a prompt. Would an hour long conversation about the
| intricacies of a system architecture become a ten minute chatter
| of what prompts people would want to ask? Does everyone then
| channel their LLM to the rest of the group? In the end, would the
| most probable response to which all LLMs being used agree with be
| the result of the conversation?
| FeteCommuniste wrote:
| Eventually the LLMs get plugged directly into our brains and do
| all our thinking and lip movements for us. We can all be Sam
| Altman's meatpuppets.
| mlboss wrote:
| What we need is mental gyms. In modern society there is no need
| for physical labor but we go to gyms just to keep ourselves
| healthy.
|
| Similarly in future we will not need mental "labor" but to keep
| ourselves sharp we need engage in mental exercises. I am thinking
| of picking up chess again just for this reason.
| incognito124 wrote:
| IMO chess is not the best mental gym. Personally, I've started
| exercising mental arithmetics
| spinach wrote:
| I've had the same experience but with drawing. What's the point
| when AI can generate perfect finished pieces in seconds? Why put
| all that effort in to learning and drawing something. It's always
| been hard for me but it used to feel worth it for the finished
| piece, but now you can bring a piece of computer art into being
| with a simple word prompt.
|
| I still create, I just use physical materials like clay and such,
| to make things that AI can't yet replicate.
| add-sub-mul-div wrote:
| AI won't create something perfect or even better. And to the
| extent we enjoy creating for its own sake that's still there.
| But it's true that real creators will have a harder time being
| seen by others when there's an ocean of slop being shoved down
| our throats by parties with endless budgets.
| regurgist wrote:
| You are in a maze of twisty little passages that are not the
| same. You have to figure out which ones are not. Try this. The
| world is infinitely complex. AI is very good at dealing with the
| things it knows and can know. It can write more accurately than I
| can, spell better. It just takes stuff I do and learns from my
| mistakes. I'm good with that. But here is something to ask AI:
|
| Name three things you cannot think about because of the language
| you use?
|
| Or "why do people cook curds when making cheese."
|
| Or how about this:
|
| "Name three things you cannot think about because of the language
| you use?"
|
| AI is at least to some extent a artificial regurgitarian. It can
| tell you about things that have been thought. Cool. But here is a
| question for you. Are there things that you can think about that
| have not been thought about before?
|
| The reason people cook curds is because the goal of cheese making
| was to preserve milk, not to make cheese.
| curl-up wrote:
| > The fun has been sucked out of the process of creation because
| nothing I make organically can compete with what AI already
| produces--or soon will.
|
| So the fun, all along, was not in the process of creation itself,
| but in the fact that the creator could somehow feel superior to
| others not being able to create? I find this to be a very
| unhealthy relationship to creativity.
|
| My mixer can mix dough better than I can, but I still enjoy
| kneading it by hand. The incredibly good artisanal bakery down
| the street did not reduce my enjoyment of baking, even though I
| cannot compete with them in quality by any measure. Modern slip
| casting can make superior pottery by many different quality
| measures, but potters enjoy throwing it on a wheel and producing
| unique pieces.
|
| But if your idea of fun is tied to the "no one else can do this
| but me", then you've been doing it wrong before AI existed.
| patcon wrote:
| Yeah, I think you're onto something. I'm not sure the
| performative motivation is necessarily bad, but def different
|
| Maybe AI is like Covid, where it will reveal that there were
| subtle differences in the underlying humans all along, but we
| just never realized it until something shattered the ability
| for ambiguity to persist.
|
| I'm inclined to so that this is a destabilising thing,
| regardless of my thoughts on the "right" way to think about
| creativity. Multiple ways could coexist before, and now one way
| no longer "works".
| quantumgarbage wrote:
| I think you are way past the argument the writer is making.
| ebiester wrote:
| Let's frame it more generously: The reward is based on being
| able to contribute something novel to the world - not because
| nobody else can but because it's another contribution to the
| world's knowledge.
| curl-up wrote:
| If the core idea that was intended to be broadcasted to the
| world was a "contribution", and LLM simply expanded on it,
| then I would view LLMs simply a component in that
| broadcasting operation (just as the internet infrastructure
| would be), and the author's contribution would still be
| intact, and so should his enjoyment.
|
| But his argument does not align with that. His argument is
| that he enjoys the act of writing itself. If he views his act
| of writing (regardless of the idea being transmitted) as his
| "contribution to world's knowledge", then I have to say I
| disagree - I don't think his writing is particularly
| interesting in and of itself. His ideas might be interesting
| (even if I disagree), but he obviously doesn't find the
| formation of ideas enjoyable enough.
| Viliam1234 wrote:
| Now you can contribute something novel to the world by
| pressing a button. Sounds like an improvement.
| drdeca wrote:
| If one merely presses a button (the same button, not
| choosing what button to push based on context), I don't see
| what it is that one has contributed? One of those tippy
| bird toys can press a button.
| lo_zamoyski wrote:
| The primary motivation should be wisdom. No one can become
| wise for you. You don't become any wiser yourself that way.
| And a machine isn't even capable of being wise.
|
| So while AI might remove the need for human beings to engage
| in certain practical activities, it cannot eliminate the
| theoretical, because by definition, theory is done for its
| own sake, to benefit the person theorizing by leading them to
| understanding something about the world. AI can perhaps find
| a beneficial place here in the way books or teachers do, as
| guides. But in all these cases, you absolutely need to engage
| with the subject matter yourself to profit from it.
| mionhe wrote:
| It sounds as if the reward is primarily monetary in this
| case.
|
| As some others have commented, you can find rewards that
| aren't monetary to motivate you, and you can find ways to
| make your work so unique that people are willing to pay for
| it.
|
| Technology forces us to use the creative process to more
| creatively monetize our work.
| movpasd wrote:
| Sometimes the fun is in creating something useful, as a human,
| for humans. We want to feel useful to our tribe.
| Tuperoir wrote:
| Good point.
|
| Self-actualisation should be about doing the things that only
| you can. Not better than anyone else, but more like, the
| specific things that ony you, with the same of your experience,
| expertise, values and constraints can do.
| StefanBatory wrote:
| Knowing that what I do anyone can do, no matter how well I'll
| do it, is discouraging. Because then, what is my purpose? What
| can I say that I'm good at?
| curl-up wrote:
| Would you say that the chess players became "purposeless"
| with Deep Blue, or Go players with Alpha Go?
| rfw300 wrote:
| It's interesting that you name those examples, because Lee
| Sedol, the all-time great Go player, retired shortly after
| losing to Alpha Go, saying: "Even if I become the number
| one, there is an entity that cannot be defeated... losing
| to AI, in a sense, meant my entire world was collapsing...
| I could no longer enjoy the game. So I retired." [1, 2]
|
| So for some, yes. It is of course also true that many
| people derive self-worth and fulfillment from contributing
| positively to the world, and AI automating the productive
| work in which they specialize can undermine that.
|
| [1] https://en.yna.co.kr/view/AEN20191127004800315
|
| [2] https://www.nytimes.com/2024/07/10/world/asia/lee-
| saedol-go-...
| curl-up wrote:
| I am in no way disputing that some people would feel that
| way because of AI, just as some performing classical
| musicians felt that way in the advent of the audio
| recorder.
|
| What I am saying is that (1) I regard this as an
| unhealthy relationship to creativity (and I accept that
| this is subjective), and (2) that most people do not feel
| that way, as can be confirmed by the fact that chess, go,
| and live music performances are all still very much
| practiced.
| aprdm wrote:
| That's a deeper question that only you can answer. I can only
| say that your thinking based on how you phrased it doesn't
| really lead to happiness in general
| yapyap wrote:
| I mean yeah apparently so for the OP but I'm sure he did not
| mean for it to be that way intentionally
| jsemrau wrote:
| "The fun has been sucked out of the process of preparing food
| because nothing I make organically can compete with what
| restaurants/supermarkets already produces--or soon will."
| getpokedagain wrote:
| I don't think it's solely the rub in others faces that they
| can't create. You hope they learn to as well.
| kelseyfrog wrote:
| > So the fun, all along, was not in the process of creation
| itself, but in the fact that the creator could somehow feel
| superior to others not being able to create? I find this to be
| a very unhealthy relationship to creativity.
|
| People realize this at various points in their life, and some
| not at all.
|
| In terms the author might accept, the metaphor of the stoic
| archer comes to mind. Focusing on the action, not the target is
| what relieves one of the disappointment of outcome. In this
| cast, the action is writing while the target is having better
| thoughts.
|
| Much of our life is governed by the success at which we hit our
| targets, but why do that to oneself? We have a choice in how we
| approach the world, and setting our intentions toward action
| and away from targets is a subtle yet profound shift.
|
| A clearer example might be someone who wants to make a friend.
| Let's imagine they're at a party and they go in with the
| intention of making a friend, they're setting themselves up for
| failure. They have relatively little control over that outcome.
| However, if they go in with the intention of showing up
| authentically - something people tend to appreciate, and
| something they have full control over - the changes of them
| succeeding increase dramatically.
|
| Choosing one's goals - primarily grounded in action - is an
| under-appreciated perspective.
| garrettj wrote:
| Yeah, there's something this person needs to embrace about the
| process rather than being some kind of modern John Henry,
| comparing themselves to a machine. There's still value in the
| things a person creates despite what AI can derive from its
| training model of Reddit comments. Find peace in the process of
| making and you'll continue to love it.
| nthingtohide wrote:
| We need to start taking a leaf of advice from spiritual
| knowledge that "You are not the doer." You were never the doer.
| The doing happened on its own. You were merely a vessel, an
| instrument. A Witness. Observe your inner mechanisms of mind,
| and you will quickly come to this realisation.
| wcfrobert wrote:
| I think the article is getting at the fact that in a post-AGI
| world, human skill is a depreciating asset. This is terrifying
| because we exchange our physical and mental labor for money.
| Consider this: why would a company hire me if - with enough GPU
| and capital - they can copy-and-paste 1,000 of AI agents much
| smarter to do the work?
|
| With AGI, Knowledge workers will be worth less until they are
| worthless.
|
| While I'm genuinely excited about the scientific progress AGI
| will bring (e.g. curing all diseases), I really hope there's a
| place for me in the post-AGI world. Otherwise, like the potters
| and bakers who can't compete in the market with cold-hard
| industrial machines, I'll be selling my python code base on
| Etsy.
|
| No Set Gauge had an excellent blog post about this. Have a read
| if you want a dash of existential dread for the weekend:
| https://www.nosetgauge.com/p/capital-agi-and-human-ambition.
| senordevnyc wrote:
| This is only terrifying because of how we've structured
| society. There's a version of the trajectory we're on that
| leads to a post-scarcity society. I'm not sure we can pull
| that off as a species, but even if we can, it's going to be a
| bumpy road.
| GuinansEyebrows wrote:
| the barrier to that version of the trajectory is that "we"
| haven't structured society. what structure exists, exists
| as a result of capital extracting as much wealth from labor
| as labor will allow (often by dividing class interests
| among labor).
|
| agreed on the bumpy road - i don't see how we'll reach a
| post-scarcity society unless there is an intentional
| restructuring (which, many people think, would require a
| pretty violent paradigm shift).
| 9dev wrote:
| That seems like a very narrow perspective. For one, it is
| neither clear we will end up with AGI at all--we could have
| reached or soon reach a plateau with the possibilities of the
| LLM technology--or whether it'll work like what you're
| describing; the energy requirements might not be feasible,
| for example, or usage is so expensive it's just not worth
| applying it to every mundane task under the sun, like writing
| CRUD apps in Python. We know how to build flying cars,
| technically, but it's just not economically sustainable to
| use them. And finally, you never know what niches are going
| to be freed up or created by the ominous AGI machines
| appearing on the stage.
|
| I wouldn't worry too much yet.
| Animats wrote:
| > With AGI, Knowledge workers will be worth less until they
| are worthless.
|
| "Knowledge workers" being in charge is a recent idea that is,
| perhaps, reaching end of life. Up until WWII or so, society
| had more smart people than it had roles for them. For most of
| history, being strong and healthy, with a good voice and a
| strong personality, counted for more than being smart. To a
| considerable extent, it still does.
|
| In the 1950s, C.P. Snow's "Two Cultures" became famous for
| pointing out that the smart people were on the way up.[1]
| They hadn't won yet; that was about two decades ahead. The
| triumph of the nerds took until the early 1990s.[2] The
| ultimate victory was, perhaps, the collapse of the Soviet
| Union in 1991. That was the last major power run by goons.
| That's celebrated in The End of History and the Last Man
| (1992).[3] Everything was going to be run by technocrats and
| experts from now on.
|
| But it didn't last. Government by goons is back. Don't need
| to elaborate on that.
|
| The glut of smart people will continue to grow. Over half of
| Americans with college educations work in jobs that don't
| require a college education. AI will accelerate that process.
| It doesn't require AI superintelligence to return smart
| people to the rabble. Just AI somewhat above the human
| average.
|
| [1] https://en.wikipedia.org/wiki/The_Two_Cultures
|
| [2] https://archive.org/details/triumph_of_the_nerds
|
| [3] https://en.wikipedia.org/wiki/The_End_of_History_and_the_
| Las...
| gibbitz wrote:
| I think the point is that part of the value of a work of art to
| this point is the effort or lack of effort involved in its
| creation. Evidence of effort has traditionally been a sign of
| the quality of thought put into a work as a product of time
| spent in its creation. LLMs short circuit this instinct in
| evaluation making some think works generated by AI are better
| than they are while simultaneously making those who create work
| see it as devaluation of work (which is the demotivator here).
|
| I'm curious why so any people see creators and intellectuals as
| competitive people trying to prove they're better than someone
| else. This isn't why people are driven to seek knowledge or
| create Art. I'm sure everyone has their reasons for this, but
| it feels like insecurity from the outside.
|
| Looking at debates about AI and Art outside of IP often brings
| out a lot of misunderstandings about what makes good Art and
| why Art is a thing man has been compelled to make since the
| beginning of the species. It takes a lifetime to select
| techniques and thought patterns that define a unique and
| authentic voice. A lifetime of working hard on creating things
| adds up to that voice. When you start to believe that work is
| in vain because the audience doesn't know the difference it
| certainly doesn't make it feel rewarding to do.
| woah wrote:
| > But now, when my brain spontaneously forms a tiny sliver of a
| potentially interesting concept or idea, I can just shove a few
| sloppy words into a prompt and almost instantly get a fully
| reasoned, researched, and completed thought. Minimal organic
| thinking required.
|
| No offense, but I've found that AI outputs very polished but very
| average work. If I am working on something more original, it is
| hard to get AI to output reasoning about it without heavy
| explanation and guidance. And even then, it will "revert to the
| mean" and stumble back into a rut of familiar concepts after a
| few prompts. Guiding it back onto the original idea repeatedly
| quickly uses up context.
|
| If an AI is able to take a sliver of an idea and output something
| very polished from it, then it probably wasn't that original in
| the first place.
| apsurd wrote:
| If you are in the camp of "ends justify the means" then it makes
| sense why all this is scary and nihilistic. What's the point
| doing anything if the outcome is infinitely better done in some
| other way by some other thing/being/agent?
|
| If the "means justify the ends" then doing anything is its own
| reason.
|
| And in the _end_, the cards will land where they may. ends-
| justify-means is really logical and alluring, until I realize why
| am I optimizing for END?
| ankit219 wrote:
| Familiar scenario. One thing that would help is to use AI as
| someone you are having a conversation with.
|
| (most AIs need to be explicitly told before you start this). You
| tell them not to agree with you, to ask more questions instead of
| providing the answers, to offer justifications and background as
| to why those questions are being asked. This helps you refine
| your ideas more, understand the blind spots, and explore
| different perspectives. Yes, an LLM can refine the idea for you,
| especially if something like that is already explored. It can
| also be the brainstorming accessory who helps you to think
| harder. Come up with new ideas. The key is to be intentional
| about which way you want it. I once made Claude roleplay as a
| busy exec who would not be interested in my offering until i
| refined it 7 times (and it kept offering reasons as to why an AI
| exec would or would not read it).
| regurgist wrote:
| You are in a maze of twisty little passages that are not the
| same. You have to figure out which ones are not.
|
| Try this. The world is infinitely complex. AI is very good at
| dealing with the things it knows and can know. It can write more
| accurately than I can, spell better. It just takes stuff I do and
| learns from my mistakes. I'm good with that. But here is
| something to ask AI:
|
| "Name three things you cannot think about because of the language
| you use?"
|
| Or "why do people cook curds when making cheese."
|
| Or how about this:
|
| "Name three things you cannot think about because of the language
| you use?"
|
| AI is at least to some degree an artificial regurgitarian. It can
| tell you about things that have been thought. Cool. But here is a
| question for you. Are there things that you can think about that
| have not been thought about before or that have been thought
| about incorrectly?
|
| The reason people cook curds is because the goal of cheese making
| was (in the past) to preserve milk, not to make cheese.
| dave1999x wrote:
| I really don't understand your point. This in nonsense to me.
| agotterer wrote:
| I appreciate your thoughts and definitely share some of your
| sentiments. But don't underestimate the value and importance of
| continuing to be a creative.
|
| Something I've been thinking about lately is the idea of creative
| stagnation due to AI. If AI's creativity relies entirely on
| training from existing writing, anrchitecture, art, music,
| movies, etc., then future AI might end up being trained only on
| derivatives of today's work. If we stop generating original ideas
| or developing new styles of art, music, etc., how long before
| society gets stuck endlessly recycling the same sounds, concepts,
| and designs?
| MinimalAction wrote:
| > My thinking systems have atrophied, and I can feel it.
|
| I do understand where the author is coming from. Most of the
| times, it is easy to _read_ an answer---regardless of whether it
| is right /wrong, relevant or not---than _think_ of an answer. So
| AI does take that friction of thinking away.
|
| But, I am still disappointed of all this doom because of AI. I am
| inclined to throw my hands and say "just don't use it then". The
| process of thinking is where the fun lies, not in showing the
| world I am better or always right than so and so.
| williamcotton wrote:
| > The fun has been sucked out of the process of creation because
| nothing I make organically can compete with what AI already
| produces--or soon will.
|
| I find immense joy and satisfaction when I write poetry. It's
| like crafting a puzzle made of words and emotions. While I do
| enjoy the output, if there is any goal it is to tap into and be
| absorbed by the process itself.
|
| Meanwhile, code? At least for me, and to speak nothing of those
| that approach the craft differently, it is (almost) nothing but a
| means to an ends! I do enjoy the little projects I work on. Hmm,
| maybe for me software is about adding another tool to the belt
| that will help with the ongoing journey. Who knows. It definitely
| feels very different to outsource coding than to outsource my
| artistic endeavors.
|
| One thing that I know won't go away are the small pockets of
| poetry readings, singer-songwriters, and other artistic
| approaches that are decidedly more personal in both creation and
| audience. There are engaged audiences for art and there are
| passive consumers. I don't think this changes much with AI.
| Centigonal wrote:
| > But now, when my brain spontaneously forms a tiny sliver of a
| potentially interesting concept or idea, I can just shove a few
| sloppy words into a prompt and almost instantly get a fully
| reasoned, researched, and completed thought.
|
| I can't relate to this at all. The reason I write, debate, or
| think at all is to find out what I believe and discover my voice.
| Having an LLM write an essay based on one of my thoughts is about
| as "me" as reading a thinkpiece that's tangentially related to
| something I care about. I write because I want to get _my_
| thoughts out onto the page, in _my_ voice.
|
| I find LLMs useful for a lot of things, but using an LLM to
| shortcut personal writing is antithetical to what I see as the
| purpose of personal writing.
| sippndipp wrote:
| I understand the depression. I'm a developer (professional) and I
| make music (ambitious hobby). Both arts heavily in a
| transformational process.
|
| I'd like to challenge a few things. I rarely have a moment where
| an LLM provides me a creative spark. It's more that I don't
| forget anything from the mediocre galaxy of thoughts.
|
| See AI as a tool.
|
| A tool that helps you to automate repetitive cognitive work.
| hodder wrote:
| "AI could probably have written this post far more quickly,
| eloquently, and concisely. It's horrifying."
|
| ChatGPT write that post more eloquently:
|
| May 16, 2025 On Thinking
|
| I've been stuck.
|
| Every time I sit down to write a blog post, code a feature, or
| start a project, I hit the same wall: in the age of AI, it all
| feels pointless. It's unsettling. The joy of creation--the spark
| that once came from building something original--feels dimmed, if
| not extinguished. Because no matter what I make, AI can already
| do it better. Or soon will.
|
| What used to feel generative now feels futile. My thoughts seem
| like rough drafts of ideas that an LLM could polish and complete
| in seconds. And that's disorienting.
|
| I used to write constantly. I'd jot down ideas, work them over
| slowly, sculpting them into something worth sharing. I'd obsess
| over clarity, structure, and precision. That process didn't just
| create content--it created thinking. Because for me, writing has
| always been how I think. The act itself forced rigor. It refined
| my ideas, surfaced contradictions, and helped me arrive at
| something resembling truth. Thinking is compounding. The more you
| do it, the sharper it gets.
|
| But now, when a thought sparks, I can just toss it into a prompt.
| And instantly, I'm given a complete, reasoned, eloquent response.
| No uncertainty. No mental work. No growth.
|
| It feels like I'm thinking--but I'm not. The gears aren't
| turning. And over time, I can feel the difference. My intuition
| feels softer. My internal critic, quieter. My cleverness, duller.
|
| I believed I was using AI in a healthy, productive way--a bicycle
| for the mind, a tool to accelerate my intellectual progress. But
| LLMs are deceptive. They simulate the journey, but they skip the
| most important part. Developing a prompt feels like work. Reading
| the output feels like progress. But it's not. It's passive
| consumption dressed up as insight.
|
| Real thinking is messy. It involves false starts, blind alleys,
| and internal tension. It requires effort. Without that, you may
| still reach a conclusion--but it won't be yours. And without
| building the path yourself, you lose the cognitive infrastructure
| needed for real understanding.
|
| Ironically, I now know more than ever. But I feel dumber. AI
| delivers polished thoughts, neatly packaged and persuasive. But
| they aren't forged through struggle. And so, they don't really
| belong to me.
|
| AI feels like a superintelligence wired into my brain. But when I
| look at how I explore ideas now, it doesn't feel like
| augmentation. It feels like sedation.
|
| Still, here I am--writing this myself. Thinking it through. And
| maybe that matters. Maybe it's the only thing that does.
|
| Even if an AI could have written this faster. Even if it could
| have said it better. It didn't.
|
| I did.
|
| And that means something.
| keernan wrote:
| >>the process of creation >>my original thoughts >>I'd have ideas
|
| As far as I can tell, LLMs are incapable of any of the above.
|
| I'd love to hear from LLM experts how LLMs can ever have original
| ideas using the current type of algorithms.
| aweiher wrote:
| Have you tried writing this article on .. AI? *scnr
| bsimpson wrote:
| I admittedly don't use AI as much as many others here, but I have
| noticed that whenever I take the time to write a hopefully-
| inciteful response in a place like Reddit, I immediately get
| people trying to dunk on me by accusing me of being an AI bot.
|
| It makes me not want to participate in those communities
| (although to be honest, spending less time commenting online
| would probably be good for me).
| FeteCommuniste wrote:
| It'a vicious cycle. As AI gets better at writing, more people
| will rely on it to write for them, and as more AI content gains
| prominence, more people will tend to ape its style even when
| they do write something for themselves.
| quantadev wrote:
| There's quite a dichotomy going on in software development. With
| AI, we can all create much more than we ever could before, and
| make it much better and even in much less time, but what we've
| lost is the sense of pride that comes with the act of
| creating/coding, because nowadays:
|
| 1) If you wrote most of it yourself then you failed to adequately
| utilize AI Coding agents and yet...
|
| 2) If AI wrote most of it, then there's not exactly that much of
| a way to take pride in it.
|
| So the new thing we can "take pride in" is our ability to
| "control" the AI, and it's just not the same thing at all. So
| we're all going to be "changing jobs" whether we like it or not,
| because work will never be the same, regardless of whether you're
| a coder, an artist, a writer, or an AD agency fluff writer. Then
| again pride is a sin, so just GSD and stop thinking about
| yourself. :)
| hooverd wrote:
| Maybe, but will anyone understand what they are creating?
| quantadev wrote:
| Right. In the near term, I'm predicting a dramatic decline in
| software quality, because code was never understood and
| tested properly. In the future we will have better processes,
| and better ways of letting AI do it's own
| checking/validation, but that's lagging right now.
| hooverd wrote:
| It'll be interesting to see a society where there's a big
| negative incentive to actually sitting down and
| understanding how things work, in the name of efficiency.
| ThomPete wrote:
| My way out of this was to start thinking about what can't the
| LLMs of the world do and my realization was actually quite simple
| and quite satisfying.
|
| What LLMs can't replace is network effects. One LLM is good but
| 10 LLMs/agents working together creating shared history is not
| replaceable by any LLM no matter how smart it becomes.
|
| So it's simple. Build something that benefit from network effects
| and you will quickly find new ideas, at least it worked for me.
|
| So now I am exploring ex. synthetic predictions markets via
| https://www.getantelope.com or
|
| Rethinking myspace but for agents instead like:
| https://www.firstprinciple.co/misc/AlmostFamous.mp4
|
| AI want's to be social :)
| xivzgrev wrote:
| in life and with people I think about a car, knowing when to go
| and when to stop
|
| some people are all go and no stop. we call them impulsive.
|
| some people may LOOK all go but have wisdom (or luck) behind the
| scenes putting the brakes on. Example: Tom Cruise does his own
| stunts, and must have a good sense for how to make it safe enough
|
| What this author touches on is a chief concern with AI. In the
| name of removing life friction, it removes your brakes. Anything
| you want to do, just ask AI!
|
| But should you?
|
| I was out the other day, pondering what the word "respect" really
| means. It's more elusive than simply liking someone. Several
| times I was tempted to just google it or ask AI, but then how
| would I develop my own point of view? This kind of thing feels
| important to have your own point of view on. And it's that we're
| losing - the things we should think about in this life, we'll be
| tempted to not anymore. And come out worse for it.
|
| All go, no brakes
| asim wrote:
| True. Now compare that to the creation of the universe and feel
| your insignificance of never being able to match it's creation in
| any form. Try to create the code that creates blades of grass.
| Good luck creating life. AI for all it's worth is a toy. One that
| shows us our own limitations but replace the word AI with God and
| understand that we spend a lot of time ruminating over things
| that don't matter. It's an opinion, don't kill me over it. But
| Dustin has reached an interesting point. What's the value of our
| time and effort and where should we place it? If the tools we
| create can do the things we used to do for "work" and "fun" then
| we need to get back to doing what only humans were made for.
| Being human.
| mattsears wrote:
| Peak Hacker News material
| montebicyclelo wrote:
| My thoughts are that it's key that humans know they will _still
| get credit_ for their contributions.
|
| E.g. imagine it was the case that you could write a blog post,
| with some insight, in some niche field - but you know that
| traffic isn't going to get directed to your site. Instead, an LLM
| will ingest it, and use the material when people ask about the
| topic, without giving credit. _If you know that will happen, it
| 's not a good incentive to write the post in the first place._
| You might think, "what's the point".
|
| Related to this topic - computers have been superhuman at chess
| for 2 decades; yet good chess humans still get credit,
| recognition, and I would guess, satisfaction, from achieving the
| level they get to. Although, obviously the LLM situation is on a
| whole other level.
|
| I guess the main (valid) concern is that LLMs get so good at
| thought that humans just don't come up with ideas as good as
| them... And can't execute their ideas as well as them... And then
| what... (Although that doesn't seem to be the case currently.)
| datpuz wrote:
| > I guess the main (valid) concern is that LLMs get so good at
| thought
|
| I don't think that's a valid concern, because LLMs can't think.
| They are generating tokens one at a time. They're calculating
| the most likely token to appear based on the arrangements of
| tokens that were seen in their training data. There is no
| thinking, there is no reasoning. If they they seem like they're
| doing these things, it's because they are producing text that
| is based on unknown humans who actually did these things once.
| montebicyclelo wrote:
| > LLMs can't think. They are generating tokens one at a time
|
| Huh? They are generating tokens one at a time - sure that's
| true. But who's shown that predicting tokens one at a time
| precludes thinking?
|
| It's been shown that the models plan ahead, i.e. think more
| than just one token forward. [1]
|
| How do you explain the world models that have been detected
| in LLMs? E.g. OthelloGPT [2] is just given sequences of games
| to train on, but it has been shown that the model learns to
| have an internal representation of the game. Same with
| ChessGPT [3].
|
| For tasks like this, (and with words), _real thought is
| required to predict the next token well_ ; e.g. if you don't
| understand chess to the level of Magnus Carlsen, how are you
| going to predict Magnus Carlsen's next move...
|
| ...You wouldn't be able to, even just from looking at his
| previous games; you'd have to actually understand chess, and
| _think_ about what would be a good move, (and in his style).
|
| [1] https://www.anthropic.com/research/tracing-thoughts-
| language...
|
| [2] https://www.neelnanda.io/mechanistic-
| interpretability/othell...
|
| [3] https://adamkarvonen.github.io/machine_learning/2024/01/0
| 3/c...
| binary132 wrote:
| Ok, nice article. Now prove an AI didn't write it.
|
| As a matter of fact I'm starting to have my doubts about the
| other people writing glowing, longwinded comments on this
| discussion.
| zzzbra wrote:
| Reminds me of this:
|
| "There are no shortcuts to knowledge, especially knowledge gained
| from personal experience. Following conventional wisdom and
| relying on shortcuts can be worse than knowing nothing at all."
| -- Ben Horowitz
| armchairhacker wrote:
| I feel similar, except not because of AI but the internet. Almost
| all my knowledge and opinions have already been explained by
| other people who put in more effort than I could. Anything I'd
| create, art or computation, high-quality similar things already
| exist. Even this comment echoes similar writing.
|
| _Almost_. _Similar_. I still make things because sometimes what
| I find online (and what I can generate from AI) isn 't "good
| enough" and I think I can do better. Even when there's something
| similar that I can reuse, I still make things to develop my
| skills for further occasions when there isn't.
|
| For example, somebody always needs a slightly different
| JavaScript front-end or CRM, even though there must be hundreds
| (thousands? tens-of-thousands?) by now. There are many
| programming languages, UI libraries, operating systems, etc. and
| some have no real advantages, but many do and consequently have a
| small but dedicated user group. As a PhD student, I learn a lot
| about my field only to make a small contribution*, but chains of
| small contributions lead to breakthroughs.
|
| The outlook on creative works is even more optimistic, because
| there will probably never be enough due to desensitization.
| People watch new movies and listen to new songs not because
| they're _better_ but because they 're _different_. AI is
| especially bad at creative writing and artwork, probably because
| it fundamentally generates "average"; when AI art is good, it's
| because the human author gave it a creative prompt, and when AI
| art is _really_ good, it 's because the human author manually
| edited it post-generation. (I also suspect that when AI gets
| creative, people will become even more creative to compensate,
| like how I suspect today we have more movies that defy tropes and
| more video games with unique mechanics; but there's probably a
| limit, because something can only be so creative before it's
| random and/or uninteresting.)
|
| Maybe one day AI can automate production-quality software
| development, PhD-level research, and human-level creativity. But
| IME today's (at least publicly-facing) models really lack these
| abilities. I don't worry about when AI _is_ powerful enough to
| produce high-quality outputs (without specific high-quality
| prompts), because assuming it doesn 't lead to an apocalypse or
| dystopia, I believe the advantages are so great, the loss of
| human uniqueness won't matter anymore.
|
| * Described in https://matt.might.net/articles/phd-school-in-
| pictures/
| rollinDyno wrote:
| I recently finished my PhD studies in social sciences. Even
| though it did not lead me to career improvements as I initially
| expected, I am happy I had the opportunity to undertake an
| academic endeavor before LLMs became cheap and ubiquitous.
|
| I bring up my studies because what the author is talking about
| strikes me as not having been ambitious enough in his thinking.
| If you prompt current LLMs with your idea and find the generated
| arguments and reasoning satisfactory, then you aren't really
| being rigorous or you're not having big enough ideas.
|
| I say this confidently because my studies showed me not only the
| methods in finding and contrasting evidence around any given
| issue, but also how much more there is to learn about the
| universe. So, if you're being rigorous enough to look at
| implications of your theories, finding datapoints that speak to
| your conclusions and find that your question has been answered,
| then your idea is too small for what the state of knowledge is in
| 2025.
| analog31 wrote:
| Something I'm on the fence about, but just trying to figure out
| from observation, is whether the AI can decide what is
| _worthwhile_. It seems like most of the successes of AI that I
| 've seen are cases where someone is tasked with writing something
| that's not worth reading.
|
| Granted, that happened before AI. The vast majority of text in my
| in-box, I never read. I developed heuristics for deciding what to
| ignore. "Stuff that looks like it was probably generated" will
| probably be a new heuristic. It's subjective for now. One clue is
| if it seems more literate than the person who wrote it.
|
| Stuff that's written for school falls into that category. It
| existed for some reason other than being read, such as the hope
| that the activity of writing conferred some educational benefit.
| That was a heuristic too -- a rule of thumb for how to teach,
| that has been broken by AI.
|
| Sure, AI can be used to teach a job skill, which is writing text
| that's not worth reading. Who wants to be the one who looks the
| kids in the eye and explain this to them?
|
| On the other hand, I do use Copilot now, where I would have used
| Stackoverflow in the past.
| largbae wrote:
| This feels off, as if thinking were like chess and the game is
| now solved and over.
|
| What if you could turn your attention to much bigger things than
| you ever imagined before? What if you could use this new
| superpower to think more not less, to find ways to amplify your
| will and contribute to fields that were previously out of your
| reach?
| sneak wrote:
| > _This post was written entirely by a human, with no assistance
| from AI. (Other than spell- and grammar-checking.)_
|
| That's not what "no assistance" means.
|
| I'm not nitpicking, however - I think this is an important point.
| The very concept of what "done completely by myself" means is
| shifting.
|
| The LLMs we have today are vastly better than the ones we had
| before. Soon, they will be even better. The complaint he makes
| about the intellectual journey being missing might be alleviated
| by an AI as intellectual sparring partner.
|
| I have a feeling this post basically just aliases to "they can
| think and act much faster than we can". Of course it's not as
| good, but 60-80% as good, 100x faster, might be net better.
| iamwil wrote:
| > But it doesn't really help me know anything new.
|
| Pursue that, since that's what LLMs haven't been helping you
| with. LLMs haven't really generated new knowledge, though there
| are hints of it--they have to be directed. There are two or three
| times when I felt the LLM output was really insightful without
| being directed.
|
| --
|
| At least for now, I find the stuff I have a lot of domain
| expertise in, the LLM's output just isn't quite up to snuff. I do
| a lot of work trying to get it to generate the right things with
| the right taste, and even using LLMs to generate prompts to feed
| into other LLMs to write code, and it's just not quite right.
| Their work just seems...junior.
|
| But for the stuff that I don't really have expertise in, I'm less
| discerning of the exact output. Even if it is junior, I'm
| learning from the synthesis of the topic. Since it's usually a
| means to an end to support the work that I do have expertise in,
| I don't mind that I didn't do that work.
| iamwil wrote:
| OP said something similar about writing blog posts when he found
| himself doing twitter a lot, back in 2013. So whatever he did to
| cope with tweeting, he can do the same with LLMs, since it seems
| like he's been writing a lot of blog posts since.
|
| > I've been thinking about this damn essay for about a year, but
| I haven't written it because Twitter is so much easier than
| writing, and I have been enormously tempted to just tweet it, so
| instead of not writing anything, I'm just going to write about
| what I would have written if Twitter didn't destroy my desire to
| write by making things so easy to share.
|
| and
|
| > But here's the worst thing about Twitter, and the thing that
| may have permanently destroyed my mind: I find myself walking
| down the street, and every fucking thing I think about, I also
| think, "How could I fit that into a tweet that lots of people
| would favorite or retweet?"
|
| https://dcurt.is/what-i-would-have-written
| ryankrage77 wrote:
| > All of my original thoughts feel like early drafts of better,
| more complete thoughts that simply haven't yet formed inside an
| LLM
|
| I would like access to whatever LLM the author is using, because
| I cannot relate to this at all. Nearly all LLM output I've ever
| generated has been average, middle-of-the-road predictable slop.
| Maybe back in the GPT-3 days before all LLMs were RLHF'd to
| death, they could sometimes come up with novel (to me) ideas, but
| nowadays often I don't even bother actually sending the prompt
| I've written, because I have a rough idea of what the output is
| going to be, and that's enough to hop to the next idea.
| ay wrote:
| Very strange. Either the author uses some magic AI, or I am
| holding it wrong. I used LLMs since a couple of years, as a nice
| tool.
|
| Besides that:
|
| I have tried using LLMs to create cartoon pictures. The first
| impression is "wow"; but after a bunch of pictures you see the
| evidently repetitive "style".
|
| Using LLMs to write poetry results is also quite cool at first,
| but after a few iterations you see the evidently repetitive
| "style", which is bland and lacks depth and substance.
|
| Using LLMs to render music is amazing at first, but after a while
| you can see the evidently repetitive style - for both rhymes and
| music.
|
| Using NotebookLM to create podcasts at first feels amazing, about
| to open the gates of knowledge; but then you notice that the flow
| is very repetitive, and that the "hosts" don't really show enough
| understanding to make it interesting. Interrupting them with
| questions somewhat dilutes this impression, though, so jury is
| out here.
|
| Again, with generating the texts, they get a distant metallic
| taste that is hard to ignore after a while.
|
| The search function is okay, but with a little bit of nudge one
| can influence the resulting answer by a lot, so I wary if blindly
| taking the "advice", and always recheck it, and try to make two
| competing where I would influence LLM into taking the competing
| viewpoints and learn from both.
|
| Using the AI to generate code - simple things are ok, but for
| non-trivial items it introduces pretty subtle bugs, which require
| me to ensure I understand every line. This bit is the most fun -
| the bug quest is actually entertaining, as it is often the same
| bugs humans would make.
|
| So, I don't see the same picture, but something close to the
| opposite of what the author sees.
|
| Having an easy outlet to bounce the quick ideas off and a source
| of relatively unbiased feedback brought me back to the fun of
| writing; so literally it's the opposite effect compared to the
| article author...
| jstummbillig wrote:
| Maybe you are not that great at using the most current LLMs or
| you don't want to be? I find that increasingly to be the most
| likely answer, whenever somebody makes sweeping claims about
| the impotence of LLMs.
|
| I get more use out of them every single day and certainly with
| every model release (mostly for generating absolutely not
| trivial code) and it's not subtle.
| deepsun wrote:
| > instantly get a fully reasoned, researched, and completed
| thought
|
| That's not my experience though. I tried several models, but
| usually get a confident half-baked hallucination, and tweaking my
| prompt takes more time than finding the information myself.
|
| My requests are typically programming tho.
| qudat wrote:
| Man people are using LLMs much differently from myself. I use it
| as an answer engine and that's about where I stop using it. It's
| a great tool for research and quick answers but I haven't found
| it great for much else.
| j7ake wrote:
| It's a faster Google search
| thorum wrote:
| If your goal is to have original and creative thoughts then you
| need to fully explore an idea _on your own_ before getting
| someone else's feedback: whether human or AI.
|
| Feedback from other humans is great but will tend to nudge you
| toward the current closest mainstream idea, whatever everyone
| else is doing or thinking about.
|
| Feedback from AI is just painfully bland. It's useful later on as
| a research tool and to spot errors in your thinking. But "Give
| the AI your half finished thought and let it finish it" is
| unlikely to produce anything that matters.
|
| Disconnect. Put your phone out of reach. Just you, your mind, and
| maybe a pencil and paper. See how far you can go with just your
| brain.
|
| Only after you've gone as far as you can should you bring in
| outside input. That's how you get to ideas that actually matter.
| lcsiterrate wrote:
| I wonder if there is any like the cs50.ai for other fields, which
| acts as a guide by not instantly giving any answer. The one I
| experiment to have like this experience is by using custom
| instructions (most of LLMs have this in the settings) wherein I
| set it to Socratic mode so the LLM will spit out questions to
| stir ideas in my head.
___________________________________________________________________
(page generated 2025-05-16 23:00 UTC)