Hacker News Reader

Comment by thr0waway001 | Hacker News Reader

AI reminds of listening to any person who seems like an intellectual authority on multiple subjects on YouTube and is not afraid to wax confidently on any topic. They seem very intelligent and knowledgable until they actually talk about something you know.

In other words, I try to learn from it whenever it does something I can't do but when it does something I can do or something I'm really good at it I find myself wanting to correct it cause it doesn't do it that well.

It just seems like a really quick thinking and fast executing but, ultimately, mid skilled / novice person.

View on HN

I think the even worse problem is that by extension now *everyone* sounds like an expert, even if they aren’t. In the past when someone wrote an RFC they would need to study to formulate it well. Now anyone can create content sounding like an expert and it becomes difficult for the reader to differentiate real expertise and depth vs shallow fancy words.

I also discussed this a bit more here: https://www.dev-log.me/everyone_is_an_expert_now/

In the last few years I have come to realize that first impression of anything is extremely important if your first few uses were good and wowed you you will be positive about it. If it was not you will be negative about it the bias of the first encounter stays with us no matter what.

Hence beginner's luck.

Most long-term gamblers will tell you that the first games they played, they won. This is a real thing, yet we cannot apply it by making one bet and then stopping, because so are the probabilities being fair and un-biased.

What squares these two things is that most of the people who played and lost their first games, did not get addicted to gambling.

Dont know about that, I dont think I had any superb first experience with it but even if I had, I got more turned on to it when I started using it for toy program/code solutions I needed on a one-off basis occasionally. If it didnt give me the code I needed to get various things done on the fly, I would maybe be more agnostic.

On non-code stuff, I think its improved or there are better options for making it get to the point and be concise more and I find when I correct it, quite often we actually get somewhere. The answers I remember from my initial use of it ofbasically how to do anything or most subjects was practically a 10 pager with some weird action plan that you were never gonna go thru.

AI's mistakes are sometimes so subtle.

Just yesterday I asked Gemini Pro 3.0 this question:

> Find such colors A and B:

> A and B are both valid sRGB color.

> Interpolating between them in CIELAB space like this

> C_cielab = (A_cielab + B_cielab) / 2

> results in a color C that can't be represented in sRGB

It gave me a correct answer, great!

...and then it proceeded to tell me to use Oklab, claiming it doesn't have this problem because the sRGB gamut is convex in Oklab.

If I didn't know Oklab does have the exact same problem I would have been fooled. It just sounds too reasonable.

You can sometime run a quick second check by taking the AI's claim and asking it for an evaluation within a fresh context. It won't be misled by the surrounding text and its answer will at least be somewhat unbiased (though it might still be quite wrong).

It helps if you phrase the question openly, not obviously fishing for a yes-or-no answer. Or, if you have to ask for a yes-or-no question, make it sound like you're obviously expecting the answer that's actually less likely, so the AI will (1) either be more willing to argue against it, or (2) provide good arguments for it you might not have considered, because it "knows" the answer is unexpected and it wants to flatter your judgment.

> It helps if you phrase the question openly, not obviously fishing for a yes-or-no answer. Or, if you have to ask for a yes-or-no question, make it sound like you're obviously expecting the answer that's actually less likely,

I do this all the time and hate that I have to do it, with the additional "do not yes-man me, be critical."

Great, now I have two answers and still no clue which one is the right one.

Now get a third opinion, and marvel at all the thinking that you have accomplished

In my experience the last answer it gives is usually the right one

Ah, so the trick is to figure out which one will be the last answer. The halting problem....

It’s always the last place you look

Any difference that asking two humans?

Yeah, enormously. People will hedge depending on how sure they are about something. They might also have credentials in whatever you ask them, if you get legal advice from a lawyer, that can be judged to be more reliable than from a lay person.

Relationships with real people are pretty cool actually. If you talk to people that you have a longer relationship with, you might also be able to judge their areas of expertise and how prone to bullshitting they are.

And it starts showing impatience when its about to run out of context, more like someone who wants to get out of the office exactly at 5.

Not just when running out of context, it's always. Once it fixates on a goal, all hell breaks loose and there's nothing that it won't be sacrificed to get there. At least that's my experience with Claude Code, I am pressing the figurative breaks all the time.

"I'm Mr. Meeseeks, look at meeee!"

Claude, I need you to help me take two strokes off my golf game.

It was sort of funny when codex switched mid-session from "patch complete, I'll now automatically run the tests and verify results" to "patch complete, if you want to run the tests, just paste this instruction into a terminal somewhere: ..."

Apparently this was caused by the context window getting full!

(At least I assume that because it went back to the old behavior after I triggered a compaction)

> In other words, I try to learn from it whenever it does something I can't do...

So you know it can be full of sh1t on all kinds of topics, and you start learning from it the moment it's 'talking' about subjects you know you don't know about? To me that sounds like the moment to stop, not the moment to start. Or am I missing something?

Sounds like an extension of https://en.wiktionary.org/wiki/Gell-Mann_Amnesia_effect, but sub newspaper for large language model.

It's a good analogy to comfort yourself. But remember AI is now being deployed on the frontline of mathematics and coming up with new theories.

The reality is much more stark then your description. Yes, in MANY instances it fails at things you know and you're an expert at. But in MANY instances it also beats you at what you're good at.

People who say stuff like the parent poster are completely mischaracterizing the current situation. We are not in a place where AI is "good" but we are "better". No... we are approaching a place of we are good and AI is starting to beat us at our own game. That is the prominent topic that is what is trending and that is the impending reality.

Yet everywhere on HN I see stuff like, oh AI fails here, or AI fails there. Yeah AI failing is obvious. It's been failing for most of my life. What's unique about the last couple years is that it's starting to beat us. Why? Because your typical HNer holds programming as not just a tool, but an identity. Your skill in programming is also a status symbol and when AI attacks your identity, the first thing you do to defend your identity is to bend reality and try to cast to a different conclusion by looking at everything from a different angle.

Face Reality.

I think that's a category error. Current AI is not better or worse than us but fundamentally different. Its main strength and weakness is that it knows too much about everything. It usually knows more than the user about the topic at hand, but it doesn't know what is actually relevant in that particular situation.

If you nudge the AI in the right direction, it may surprise you with what it's capable of. But if you nudge it in a wrong direction or just don't give it sufficient context, it can be very confidently wrong.

Yeah this is another way to comfort yourself. Yes AI is different but again FACE reality.

Ai is different but it is also similar, for example it can speak language. So it is different and similar at the same time. I am obviously referring to AI beating us where we are similar to it. Best example: software engineering. This is obvious. The fact that I need to spell it out shows how deep the delusion goes.

The rest of your response is just regurgitating what the parent post said. Sure. But it doesn’t address the fact that while everything you said is true part of the time the other part of the time it beats us (that includes both nudging and no nudging it in the right direction).

Additionally all of this is also completely ignoring the fact that AI leap frogged in capability in the past year with my entire company now foregoing the use of text editing or using IDEs and having Claude write everything. If what you say is true only now and we see this much velocity in improvement then wait a couple more years and everything you said can be completely false if the trendline continues.

And here’s the craziest part. Everything im saying is obvious. I’m basically being captain obvious here. The question is why are so many people in total denial.

You already stated the reason you think why so many people are in total denial. If indeed the reason is indeed a sense of threat to what people take themselves to be then I would be very surprised if the response would have been any different. Whether this indeed is the case is what remains to be seen. I for one do believe there is something different going on here than yet another technological advancement, but again - time will tell

So you do not need a text editor anymore, the LLM writes everything? That is not my experience. My experience is that it is for sure much smarter than i am, but still i need to hand-hold it to ensure the code stays readible and the amount of code does not blow up in relation to the actual use case. It usually does not know enough about the actual problem space. It's also not so good at keeping the correct abstractions at the correct place of the code architecture. When not hand-holding, at a certain point the codebase becomes unmaintainable be human nor LLM.

That’s not just my experience. That’s the experience of everyone where I work. Nobody uses a text editor anymore at my company.

We still hand hold it a bit. If it makes a mistake we just tell it to fix the mistake or do it in a different way. It’s that good.

It's an interesting parallel to, especially right wingers, want project intelligence into 1 dimension so things all humans can be ordered from inferior to superior. That logic was already strained with humans, but with the introduction of AI the wheels are really coming off for that model.

I am pretty sure this comment, as many others of the same poster were LLM-generated with slight prompt tweak, which is against TOS. Check their history for proofs.

Have you been actively using paid versions of the flagship models from Ant / OpenAI? I’m just curious if the conclusion was made within the last 6 months or not.

I got that experience 3 hours ago.

Gell-Mann amnesia. The things it tells you about things you don't know are things that would make a knowledgeable person go "dude, wtf? That's totally wrong."

You can really only use AI for: things that are easy to verify; things that you already know how to do but want done faster; things you're learning to do and are just one step out of your reach (so it's still comprehensible to you); or, things that just plain don't matter.

That's a lot of stuff, but it also doesn't include a lot of the stuff people claim AI can do.

Reply on news.ycombinator.com