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

LLMs are making me lose my savviness

pgaleone.eu

Post reply on HN
Print view View original

61–70 of 86 posts

Re: LLMs are making me lose my savviness

#61

Post by luke5441 »

dvt wrote: post #58
luke5441 wrote: post #57

Earlier quoted context omitted.

But if a LLM writes such a paper, how do you know there aren't any issues in it that cause wrong probabilities/distributions?

You don't, that's why you review it (just how you would handle a real human submitting a paper). The main issue with LLM papers is that they tend to be very dense/circuitous or—because LLMs don't truly understand what they're doing—have the wrong focus. An LLM might spend pages on a trivial result, but quickly gloss over a truly remarkable finding. Terry Tao talks about this challenge in some of his blog posts and in…

That was adamddev1s point. It's a unreliable tool. You cannot copy & paste its results into a paper as proof. You could do this with e.g. Lean. To review what it writes you need to understand the topic to such an extend that you could have written it yourself (at least that is how it is for me when I review code).

Yes, for inspiration or search they are great tools.

Re: LLMs are making me lose my savviness

#62

Post by Scubabear68 »

titularcomment wrote: post #24
Scubabear68 wrote: post #12

It is opposite for me than what the author describes. With AI I can explore new areas far more effectively. For example, I have never been great with front end web stuff, but asking LLMs to generate something and then looking at the code they built is pretty cool. Of course there are errors and gaffes and weird bits all over, but that helps keep me in the loop and understanding it better. I can now (more) boldly go w…

With all due respect, what you're describing doesn't quite spell learning in my opinion. Sure, its better to read the code of an application you envisioned to learn a programming language rather than an unrelated convulated program; but you're still offloading your thought process to the LLM. You're not quite seeing why it chose A instead of B, and why it uses that struct instead of the other. And honestly, the LLM d…

I am describing a problem and seeing how the LLM solves it (or how it goes off the rails). I am often exposed to techniques new to me. I call that learning.

To be clear, I am closing in on 60 years old. I’ve written a lot of code in a lot of languages. I have not bothered to check every fad that has come down in front end web and Javascript for the past umpteen years. Front end hasn’t been my thing. With current AI, I can leap frog a lot of bull crap I don’t care about and see what at least the AI thinks is good. I still retain a brain, I can still investigate what it did and form an opinion about it.

I treat the AI stuff as idiot savants that can make really nice stuff or go totally off in the weeds. But it is more than good enough to help me be a better me at what I do.

Re: LLMs are making me lose my savviness

#63

Post by 2muchcoffeeman »

titularcomment wrote: post #24
Scubabear68 wrote: post #12

It is opposite for me than what the author describes. With AI I can explore new areas far more effectively. For example, I have never been great with front end web stuff, but asking LLMs to generate something and then looking at the code they built is pretty cool. Of course there are errors and gaffes and weird bits all over, but that helps keep me in the loop and understanding it better. I can now (more) boldly go w…

With all due respect, what you're describing doesn't quite spell learning in my opinion. Sure, its better to read the code of an application you envisioned to learn a programming language rather than an unrelated convulated program; but you're still offloading your thought process to the LLM. You're not quite seeing why it chose A instead of B, and why it uses that struct instead of the other. And honestly, the LLM d…

> “ honestly, the LLM doesn't quite know why it chose that either.”

Uh, have you asked it?

These sorts of articles are funny. Like the author assumed that you have to use an LLM in a certain way.

“I can just ask the LLM and there it is! I didn’t learn anything! :(“

You can interrogate the LLM. You can ask it all sorts of things. Tell it to justify its decisions. Tell it to give you supporting evidence. Tell it to give you options. Tell it to explain things in detail.

I’ve been using it as a learning tool and it’s quite effective.

Granted the stuff I don’t care about really does get worse. If I have a config typo I just go “There’s a typo in blah. Find it and fix it.” Or setting up the boiler plate for a project. I’m getting quite useless at that.

But it doesn’t have to be any particular way. It’s a tool. Exploit it.

Re: LLMs are making me lose my savviness

#64

Post by skeledrew »

Begging the question here. Author is aware that they can choose not use LLMs and they... don't? It's like living at walking distance from town, being used to going in and around on foot, then electric scooters became available, so they start using them to get around while complaining that they aren't getting the exercise they want anymore.

Re: LLMs are making me lose my savviness

#66

Post by duxup »

Scubabear68 wrote: post #12

It is opposite for me than what the author describes. With AI I can explore new areas far more effectively. For example, I have never been great with front end web stuff, but asking LLMs to generate something and then looking at the code they built is pretty cool. Of course there are errors and gaffes and weird bits all over, but that helps keep me in the loop and understanding it better. I can now (more) boldly go w…

Being able to ask the dumb questions but then get focused on something more technical is so nice. Get examples on demand.

It helps me pickup and understand new things pretty fast.

Re: LLMs are making me lose my savviness

#68

Post by beej71 »

2muchcoffeeman wrote: post #63
titularcomment wrote: post #24

Earlier quoted context omitted.

With all due respect, what you're describing doesn't quite spell learning in my opinion. Sure, its better to read the code of an application you envisioned to learn a programming language rather than an unrelated convulated program; but you're still offloading your thought process to the LLM. You're not quite seeing why it chose A instead of B, and why it uses that struct instead of the other. And honestly, the LLM d…

> “ honestly, the LLM doesn't quite know why it chose that either.” Uh, have you asked it? These sorts of articles are funny. Like the author assumed that you have to use an LLM in a certain way. “I can just ask the LLM and there it is! I didn’t learn anything! :(“ You can interrogate the LLM. You can ask it all sorts of things. Tell it to justify its decisions. Tell it to give you supporting evidence. Tell it to giv…

I use it as a very effective learning tool, as well. To test my learning, I make sure I can code it myself after that.

Re: LLMs are making me lose my savviness

#69

Post by Scubabear68 »

grebc wrote: post #17
Scubabear68 wrote: post #12

It is opposite for me than what the author describes. With AI I can explore new areas far more effectively. For example, I have never been great with front end web stuff, but asking LLMs to generate something and then looking at the code they built is pretty cool. Of course there are errors and gaffes and weird bits all over, but that helps keep me in the loop and understanding it better. I can now (more) boldly go w…

Is reading example code uncool?

I often find that example code never covers my use cases, and I am left scratching my head over nuances not covered there.

With AI, I can describe my exact use case and see what it gives me. Sometimes it will just cough up a furball like a boilerplate example, but usually it gets pretty close, and I can see how my (maybe weird) edge case fits into the framework or API or whatever.

It may be that I approach problems in a weird way, and that is why examples are less helpful than with some, but AI generally does better here for me.

Re: LLMs are making me lose my savviness

#70

Post by anabis »

Lee Sedol, a top play of Go, quit playing professionally after AlphaGo. Others are still playing. I think this is a personal quest, and differs individually. There probably are similar stories at the start of industrialization.
Post reply on HN