Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

media.mit.edu

Post reply on HN
Print view View original

331–340 of 529 posts

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#331

Post by lazide »

smallerfish wrote: post #285
lazide wrote: post #271

Earlier quoted context omitted.

Most of the tests are BS too. And nobody is talking about verifying if the AI bubble sort is correct or not - but recognizing that if the AI is implementing it’s own bubble sort, you’re waaaay out in left field. Especially if it’s doing it inline somewhere. The underlying issue with AI slop, is that it’s harder to recognize unless you look closely, and then you realize the whole thing is bullshit.

> Most of the tests are BS too. Only if you don't constrain the tests. If you use agents adversarially in generating test cases, tests and review of results, you can get robust and tight test cases. Unless you're in research, most of what we do in our day jobs is boilerplate. Using these tools is not yet foolproof, but with some experience and experimentation you can get excellent results.

I don’t have to do boilerplate, generally.

And with all the stacking LLMs against each other, that just sounds like more work than just… writing the damn tests.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#332

Post by vidarh »

acedTrex wrote: post #326
vidarh wrote: post #167

Earlier quoted context omitted.

My "actual job" isn't to write code, but to solve problems. Writing code has just typically been how I've needed to solve those problems. That has increasingly shifted to "just" reviewing code and focusing on the architecture and domain models. I get to spend more time on my actual job.

Did you not read the comment you are replying too?

I did. Did you read the comment you replied to?

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#333

Post by iambateman »

wisty wrote: post #160
sdoering wrote: post #111

This reminds me of the recurring pattern with every new medium: Socrates worried writing would destroy memory, Gutenberg's critics feared for contemplation, novels were "brain softening," TV was the "idiot box." That said, I'm not sure "they've always been wrong before" proves they're wrong now. Where I'm skeptical of this study: - 54 participants, only 18 in the critical 4th session - 4 months is barely enough time…

Soapbox time. They were arguably right. Pre literate peole could memorise vast texts (Homer's work, Australian Aboriginal songlines). Pre Gutenberg, memorising reasonably large texts was common. See, e.g. the book Memory Craft. We're becoming increasingly like the Wall E people, too lazy and stupid to do anything without our machines doing it for us, as we offload increasing amounts onto them. And it's not even that…

You’re currently using the internet.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#334

Post by sersi »

doublerabbit wrote: post #199
CuriouslyC wrote: post #189

Earlier quoted context omitted.

Brains are adaptive. We're not getting dumber, we're just adapting to a new environment. Just because they're less fit for other environments doesn't make it worse. As for the productivity paradox, this discounts the reality that we wouldn't even be able to scale the institutions we're scaling without the tech. Whether that scaling is a good thing is debatable.

Brains are adaptive and as we adapt we are turning more cognitive unbalanced. We're absorbing potentially bias information at a faster rate. GPT can give you information of X in seconds. Have you thought about it? Is that information correct? Information can easily be adapted to sound real while masking the real as false. Launching a search engine and searching may spew incorrectness but it made you make judgement, t…

> Who opens an physical encyclopedia nowadays? I know plenty of people who binge wikipedia and learn new things through that. While Wikipedia is not always perfect, it's not like older printed encyclopaedia like Britannica were perfect either.

You have a point with trusting AI, but I'm starting to see people around me realising that LLMs tend to be overconfident even when wrong and verifying the source instead of just trusting. That's the way I use something like perplexity, I use it as an improved search engines and then tend to visit the sources it lists.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#335

Post by mcv »

jeffreygoesto wrote: post #238
mcv wrote: post #98

This seems to confirm my feeling when using AI too much. It's easy to get started, but I can feel my brain engaging less with the problem than I'm used to. It can form a barrier to real understanding, and keeps me out of my flow. I recently worked on something very complex I don't think I would have been able to tackle as quickly without AI; a hierarchical graph layout algorithm based on the Sugiyama framework, using…

Elk (Eclipse Layout Kernel) is a very good package solving that, you might want to check it's Javascript port https://github.com/kieler/elkjs

I'd briefly come across Elk, but couldn't tell how it was better than what I was using. The examples I could find all showed far simpler graphs than what we had, and nothing that seemed to address the problems we had, but maybe I should give it another look, because I've kinda lost faith that dagre is going to do what we need.

If I can explain briefly what our issue is: we've got a really complex graph, and need to show it in a way that makes it easy to understand. That by itself might be a lost cause already, but we need it fixed. The problem is that our graph has cycles, and dagre is designed for DAGs; directed acyclic graphs. Fortunately it has a step that removes cycles, but it does that fairly randomly, and that can sometimes dramatically change the shape of the graph by creating unintentional start or end nodes.

I had a way to fix that, but even with that, it's still really hard to understand the graph. We need to cut it up into parts, group nodes together based on shared properties, and that's not something dagre does at all. I'm currently looking into cola with its constraints. But I'll take another look at elk.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#336

Post by jimmaswell »

carterschonwald wrote: post #86

idk, if anything I’m thinking more. The idea that I might be able to build everything I’ve ever planned out. At least the way I’m using them, it’s like the perfect assistive device for my flavor of ADHD — I get an interactive notebook I can talk through crazy stuff with. No panacea for sure, but I’m so much higher functioning it’s surreal. I’m not even using em in the volume many folks claim, more like pair programmi…

Same here re: ADHD. It's been invaluable. A big project that would have been personally intractible is now easy - even if the LLM gives slightly wrong answers 20% of the time, the important thing is that it collapses the search space for what concepts or tools I need to look into and gives an overall structure to iterate on. I tend to use ChatGPT for the big planning/architectural conversation, and I find it's also very good at sample code; for code writing/editing, Copilot has been fantastic too, lately mostly using the Opus agent in my case. It's so nice being able to delegate some bullshit gruntwork to it while I either do something else or work on architecture in another window for a few minutes.

It certainly hasn't inhibited learning either. The most recent example is shaders. I started by having it just generate entire shaders based on descriptions, without really understanding the pipeline fully, and asking how to apply them in Unity. I've been generally familiar with Unity for over a decade but never really touched materials or shaders. The generated shaders were shockingly good and did what I asked, but over time I wanted to really fine tune some of the behavior and wound up with multiple passes, compute shaders, and a bunch of other cool stuff - and understanding it all on a deeper level as a result.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#337

Post by hobofan »

gambiting wrote: post #252
foxes wrote: post #164

Earlier quoted context omitted.

Is this a copilot ad?

It does read like one.

It reads like an anti-ad for both. "I didn't use the Copilot IDE because I lack control over the context provided" and "I used Copilot 365 because it for sure doesn't have any context of anything because connecting things to it is hard/expensive".

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#338

Post by waterhouse »

mcv wrote: post #98

This seems to confirm my feeling when using AI too much. It's easy to get started, but I can feel my brain engaging less with the problem than I'm used to. It can form a barrier to real understanding, and keeps me out of my flow. I recently worked on something very complex I don't think I would have been able to tackle as quickly without AI; a hierarchical graph layout algorithm based on the Sugiyama framework, using…

Has anyone here switched roles from programmer to manager (of human programmers), and is it a similar feeling?

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#339

Post by rob »

vidarh wrote: post #332
acedTrex wrote: post #326

Earlier quoted context omitted.

Did you not read the comment you are replying too?

I did. Did you read the comment you replied to?

Just curious, are these articles on your "blog" written by AI or are you writing them all by hand?

https://hokstadconsulting.com/blog

Very impressive daily output! Keep up the good work.

Re: Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant

#340

Post by kranner »

vidarh wrote: post #320
lazide wrote: post #271

Earlier quoted context omitted.

Most of the tests are BS too. And nobody is talking about verifying if the AI bubble sort is correct or not - but recognizing that if the AI is implementing it’s own bubble sort, you’re waaaay out in left field. Especially if it’s doing it inline somewhere. The underlying issue with AI slop, is that it’s harder to recognize unless you look closely, and then you realize the whole thing is bullshit.

> Most of the tests are BS too. Why are you creating BS tests? > And nobody is talking about verifying if the AI bubble sort is correct or not - but recognizing that if the AI is implementing it’s own bubble sort, you’re waaaay out in left field. Verifying time and space complexity is part of what your tests should cover. But this is also a funny example - I'm willing to bet the average AI model today can write a far…

But a human developer, even a not so great one, might know something about the characteristics of the actual data a particular program is expected to encounter that is more efficient than this AI-coded hybrid sort for this particular application. This is assuming the AI can't deduce the characteristics of the expected data from the specs, even if a particular time and space complexity is mandated.

I encountered something like this recently. I had to replace an exact data comparison operation (using a simple memcmp) with a function that would compare data and allow differences within a specified tolerance. The AI generated beautiful code using chunking and all kinds of bit twiddling that I don't understand.

But what it couldn't know was that most of the time the two data ranges would match exactly, thus taking the slowest path through the comparison by comparing every chunk in the two ranges. I had to stick a memcmp early in the function to exit early for the most common case, because it only occurred to me during profiling that most of the time the data doesn't change. There was no way I could have figured this out early enough to put it in a spec for an AI.

Post reply on HN