Thoughts on thinking • Hacker News Board

Re: Thoughts on thinking

#321

Post by jstummbillig » Sat, May 17, 2025, 11:22 AM UTC

guyfhuo wrote: post #225

jstummbillig wrote: post #121

Earlier quoted context omitted.

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.

> Maybe you are not that great at using the most current LLMs or you don't want to be? I’m tired of this argument. I’ll even grant you: both sides of it. It seems as though we prepared our selves to respond to llms in this manner with people memeing, or simply recognizing, that there was a “way” to ask questions to get better results early on when ranked search broadened the appeal of search engines. The reality is t…

The problem is that, knowing the public internet, what would absolutely happen, is people arguing the ways in which

a) the code is bad b) the problem is beneath what they consider non-trivial

The way that OP structured the response, I frankly got a similar impression (although the follow up feels much different). I just don't see the point in engaging in that here, but I take your criticism: Why engage at all. I should probably not, then.

Re: Thoughts on thinking

#322

Post by vaylian » Sat, May 17, 2025, 11:35 AM UTC

rollinDyno wrote: post #85

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…

This is spot-on. LLMs only cover the things that are already documented/debated. Research requires looking at the current landscape of facts and theories and noticing the gaps that can be filled.

Re: Thoughts on thinking

#323

Post by bonoboTP » Sat, May 17, 2025, 11:35 AM UTC

throwaway2037 wrote: post #320

bonoboTP wrote: post #311

Earlier quoted context omitted.

Is it better to be introduced to tons of painters vs fewer but in more detail? Or being told about a painter by someone in person vs by an algorithm? In the 90s you only had certain songs if you knew someone who had it on cassette and you borrowed it and put it on your mixtape. Throughout the interaction, you also got initiated deeper into the culture of that thing in person. I also notice that families rarely sit to…

> In the 90s you only had certain songs if you knew someone who had it on cassette and you borrowed it and put it on your mixtape. I knew lots of people who recorded 120 Minutes on MTV and listened to college radio.

I meant the niche long tail stuff, since the commenter mentioned "tons of painters I would not have been otherwise". The equivalent in music would not be on MTV.

Re: Thoughts on thinking

#324

Post by satisfice » Sat, May 17, 2025, 11:41 AM UTC

I want to shake this guy. I want to say “you are a choosing to be a fucking idiot because you falsely believe you are a vacuous moron.” But I know it won’t help.

Here’s the problem: he thinks that what LLMs produce are well-reasoned, coherent thoughts. Here’s a healthier alternative: what LLMs produce is shallow and banal text, designed to camouflage its true nature.

Now here’s a heuristic: treat anything written by an LLM, about any conceptual matter as wrong by definition (because it was not a product of human experience and insight, and because we are humans). If it LOOKS right, look closer. Look more carefully. Take that insight to the next level.

Second heuristic: anything written by an LLM that you cannot falsify is, by definition, banal. Ho hum. Who cares? Does an LLM have an opinion about how to find happiness? How cute… but not worth believing.

Third heuristic: that which an LLM writes which you can neither falsify nor dismiss as banal, you may assume that the LLM itself does not understand. It’s babbling. But perhaps you can understand it, and take it farther.

Define YOURSELF as that which lies beyond these models, and write from that sensibility.

Re: Thoughts on thinking

#325

Post by johnisgood » Sat, May 17, 2025, 11:45 AM UTC

Swizec wrote: post #162

johnisgood wrote: post #130

Earlier quoted context omitted.

You have to somehow figure out the root cause of the laziness, or if it really is laziness, and not something else, e.g. a mental health issue. Plus, many kids fail school not because of laziness, but because of their toxic environment.

> if it really is laziness, and not something else, e.g. a mental health issue. Kids optimize. When I was in high school I was fully capable of getting straight F's in a class I didn't care about and straight A's in a class I enjoyed. Why bother learning chemistry when you could instead spend that time coding cool plugins and websites in PHP that thousands of internet strangers are using? I really did build one of th…

What you just described is irrelevant to what we are discussing.

As for what you said, yeah, I got 1s (Fs) because I was too busy coding and reading books on philosophy, as a 14 years old.

Re: Thoughts on thinking

#326

Post by mistercow » Sat, May 17, 2025, 11:52 AM UTC

JackFr wrote: post #136

zdragnar wrote: post #46

Earlier quoted context omitted.

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 calcula…

Well I can extract a square root by hand. We all had to learn it and got tested on it. No one to day learns that anymore. The vast, vast majority have no idea and I don’t think people are dumber because of it. That is to say, I think it’s not cut-and-dried. I agree you need to learn something, but something’s it’s okay use a tool.

> No one to day learns that anymore. The vast, vast majority have no idea and I don’t think people are dumber because of it.

Arguably, the kind of person who was helped by learning to do that by hand still learns to do it by hand, but because of curiosity rather than because a teacher told them to.

I remember being thirteen and trying to brute force methods for computing the square root. I didn’t have the tools yet to figure out how to do it in any systematic way, and the internet wasn’t at a point yet where it would have even occurred to me to just search online. Wikipedia wouldn’t exist for another two years.

I probably finally looked it up at some point in high school. I’m not sure exactly when, but I remember spending a lot of time practicing doing a few iterations in my head as a parlor trick (not that I ever had the opportunity to show it off).

If I were thirteen and curious about that now, I’d probably just ask ChatGPT. Then I’d have a whole follow up conversation about how it was derived. It would spit a lot of intimidating LaTeX at me, but unlike with Wikipedia, I’d be able to ask it to explain what those things meant.

This is the thing I don’t get when people talk about LLMs’ impact on education. Everybody focuses on cheating, like learning is inherently a chore that all students hate and must be carefully herded into doing despite themselves.

But that’s a problem with school, not learning. If your actual, self-motivated goal is to learn something, LLM’s are an incredible tool, not a hindrance.

Re: Thoughts on thinking

#327

Post by jstummbillig » Sat, May 17, 2025, 11:53 AM UTC

ay wrote: post #142

jstummbillig wrote: post #121

Earlier quoted context omitted.

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.

Could totally be the case, that, as I wrote in the very first sentence, I am holding it wrong. But I am not saying LLMs are impotent - the other week Claude happily churned me ~3500 lines of C code that allowed to implement a prototype capture facility for network packets with flexible filters and saving the contents into pcapng files. I had to fix a couple of bugs that it made, but overall it was certainly at least…

Reading this I am not sure I got the gist of your previous post. Re-reading the previous post, I still don't see how the two posts gel. I submit we might just have very different interpretations of the same observations. For example I have a hard imagining the described 3500 LOC program as 'simple'. Limited in scope, sure. But if you got it done 5-10x faster, then it can't be that simple?

Anyway: I found the writers perspective on this whole subject to be interesting, and agree on the merits — I definitely think they are correct on their analysis and outlook, and here the two of us apparently disagree – but I don't share their concluding feelings.

But I can see, how they got there.

Re: Thoughts on thinking

#328

Post by milliams » Sat, May 17, 2025, 11:54 AM UTC

mr_toad wrote: post #192

fennecbutt wrote: post #181

Earlier quoted context omitted.

99% if not 100% of human thought and general output is derivative. Everything we create or do is based on something we've experienced or seen. Try to think of an object that doesn't exist, and isn't based on anything you've ever seen before, a completely new object with no basis in our reality. It's impossible. Writers made elves by adding pointy ears to a human. That's it.

> Writers made elves by adding pointy ears to a human. Now that’s reductionist to the point of being diminutive.

Elves are wonderful. They provoke wonder. Elves are marvellous. They cause marvels. Elves are fantastic. They create fantasies. Elves are glamorous. They project glamour. Elves are enchanting. They weave enchantment. Elves are terrific. They beget terror. The thing about words is that meanings can twist just like a snake, and if you want to find snakes look for them behind words that have changed their meaning. No one ever said elves are nice. Elves are bad.

― Terry Pratchett, Lords and Ladies

Re: Thoughts on thinking

#329

Post by flysand7 » Sat, May 17, 2025, 12:07 PM UTC

layer8 wrote: post #164

flysand7 wrote: post #120

Earlier quoted context omitted.

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…

How much hours would you estimate did you watch (I assume it was video, not just audio) in those years? What kind of material? Just curious.

Mostly anime. Surprisingly, not that much, I think somewhere in the ballpark of 100 titles. In the beginning I was also watching some grammar tutorials on YouTube to get started with grammar quicker (Otherwise convergence on solution would be too slow).

Contrary to what I said I actually did use dictionaries, but the point I was trying to make is rather than memorizing phrases in advance, I used it to translate something I thought I heard.

Re: Thoughts on thinking

#330

Post by BlarfMcFlarf » Sat, May 17, 2025, 12:08 PM UTC

Cipater wrote: post #279

abakker wrote: post #193

Earlier quoted context omitted.

I for one, thing directing subordinates to do something I could be doing kinda … sucks? Like, I get that’s how you have to work with LLMs, but it isn’t a fun thing to do for me.

Do you do every single thing that you are capable of doing yourself?

Non hierarchical collaboration is the option you are excluding. Where you accept pushback and feedback because you know it comes with creative vision and perspective you lack. You can do creative things with other creative people.