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

AI should elevate your thinking, not replace it

koshyjohn.com

Post reply on HN
Print view View original

591–600 of 656 posts

Re: AI should elevate your thinking, not replace it

#591

Post by SkyBelow »

willhslade wrote: post #171
ua709 wrote: post #88

Earlier quoted context omitted.

I get your point, I just wonder how accurate it is. We basically never look at the output of the compiler, so I agree that tool allows one to operate at a higher level than assembly. But I always have to wade through the output from AI so I’m not sure I got to move to the next level of abstraction. But maybe that’s just me.

Are compilers deterministic?

Technically LLMs can be ran in deterministic mode as well, but I don't think that is enough. Even a deterministic LLM is too chaotic, small changes in prompts or the otherwise general context can result in vastly different outputs. This makes me think we need some other factor that is stronger than (or maybe orthogonal to) determinism. A notion of being well-behaved or some other non-chaotic term, so that slightly different inputs don't lead to vastly unexpected results.

Even that doesn't feel quite correct, because a compiler does seem quite chaotic. Forget a semi colon and an otherwise 99.99% code base results in a vastly different output. But it is still a very understandable output. Very predictable. So while treating both compilers and LLMs as functions that map massive input strings to massive output strings, there is some property I can't well define that compilers have that LLMs still lack (and the question is if they'll always lack it). I can't really define what it is, but it is something more than determinism.

Re: AI should elevate your thinking, not replace it

#592

Post by lokhura »

bartread wrote: post #588
nunez wrote: post #303

> If the job were mainly about producing syntactically valid code, then of course A.I. would be on a direct path to replacing large parts of the profession. But that was never the highest-value part of the work. The value was always in judgment. > The valuable engineer is the one who sees the hidden constraint before it causes an outage. The one who notices that the team is solving the wrong problem. The one who redu…

Does AI remove all struggle? I would credit my relative success (compared with many complaints I’ve read) at building software using AI as a helper to the fact that I’ve already solved the problem in my mind before I prompt the AI, and I tell it how I want the problem to be solved. Generally it’s possible to do that in English much more tersely than it is in code. So what I’ve done is the hard (in the sense of proble…

>Generally it’s possible to do that in English much more tersely than it is in code.

I really doubt this is the case for any decent amount of complexity. English is ambiguous. Programming languages exist because we don't want this ambiguity from natural languages. A sufficiently descriptive spec written in English is barely indistinguishable from just code. We've come full circle.

Re: AI should elevate your thinking, not replace it

#593

Post by at-fates-hands »

abustamam wrote: post #582
netcan wrote: post #318

Earlier quoted context omitted.

Myopic is inevitable, to some extent. It's very hard to project this stuff. Socrates wrote about what was being lost as philosophy was becoming written rather than oral...and he was right. We can't even understand what was lost. Many methods of learning and thinking became entirely lost. You could say they were redundant, and they were. But... writing largely replaced oral traditions. It didn't just augment them. He…

I think if manual coding becomes "outdated" then there will just be no demand for junior engineers to manually code. People will probably still learn to code manually, just as there are folks who will still build their own furniture. There may just not be a business demand for it. What that means 20-30 years from now when the seniors of today retire if there are no juniors right now is yet to be seen. People say that…

>> What that means 20-30 years from now when the seniors of today retire.

I fear that many won't retire and instead completely leave the industry which is already happening. Its anecdotal, but when I first started as a junior dev, I was working with many intermediate devs who had a few years on me.

I kept ties with a group of about two dozen devs. We all went through a lot of the same stuff. Last year I attended two local conferences. Out of the 24 or so, who were all seasoned senior devs now? Only 3 of us remained in the industry. Granted, I'm in accessibility and another moved more into a UI/UX design role but we were all that's left.

The majority of a discussion at lunch was about why they left and it was pretty universal. They were seeing AI creeping into everything they did and just walked away. The list was long of what they disliked about it and really didn't see the huge upsides that the industry was pushing. They had money, they had other opportunities they choose to pursue far and away from the tech industry.

It was pretty eye opening to say the least. We always imagined sitting around a table in our 60's recounting our experiences in tech and now we're not even into our 40's and the industry is losing amazing talent every year that IMHO cannot be replaced by an LLM prompt.

I don't have a good feeling about where this is headed.

Re: AI should elevate your thinking, not replace it

#594

Post by cindyllm »

ekeke wrote: post #568
TrackerFF wrote: post #352

Earlier quoted context omitted.

When GPT 3.5 was released, it could handle maybe a 500 LOC codebase. Experienced engineers were calling it cute, but zero threat to actual programmers. Then it became thousands. Now models can handle and operate on code bases with hundreds of thousands LOC, even low MLOC. So in just 3.5 years we've gone from LLMs being cute toys, to being powerful enough to actually replace junior engineers. Even if we hit a new AI w…

What damage lmao? Let’s see the llm producers raise the price to what is necessary to generate viable returns. BTW they need to make enough to finance reinvestment internally… so it’s a lot more than you think. When they raise the price firms will then have to do a deep dive analysis on what to do - for they cannot see operating expenses climb incrementally without seeing revenue and costs of operations go in a favou…

[dead]

Re: AI should elevate your thinking, not replace it

#596

Post by bartread »

lokhura wrote: post #592
bartread wrote: post #588

Earlier quoted context omitted.

Does AI remove all struggle? I would credit my relative success (compared with many complaints I’ve read) at building software using AI as a helper to the fact that I’ve already solved the problem in my mind before I prompt the AI, and I tell it how I want the problem to be solved. Generally it’s possible to do that in English much more tersely than it is in code. So what I’ve done is the hard (in the sense of proble…

>Generally it’s possible to do that in English much more tersely than it is in code. I really doubt this is the case for any decent amount of complexity. English is ambiguous. Programming languages exist because we don't want this ambiguity from natural languages. A sufficiently descriptive spec written in English is barely indistinguishable from just code. We've come full circle.

It’s the case because English is wildly more expressive than many commonly used programming languages.

And especially when you take into account the mess of front end web development and tooling.

Re: AI should elevate your thinking, not replace it

#598

Post by MarsIronPI »

bartread wrote: post #596
lokhura wrote: post #592

Earlier quoted context omitted.

>Generally it’s possible to do that in English much more tersely than it is in code. I really doubt this is the case for any decent amount of complexity. English is ambiguous. Programming languages exist because we don't want this ambiguity from natural languages. A sufficiently descriptive spec written in English is barely indistinguishable from just code. We've come full circle.

It’s the case because English is wildly more expressive than many commonly used programming languages. And especially when you take into account the mess of front end web development and tooling.

Paul Graham might say if that's the case then you need a more expressive language.

Re: AI should elevate your thinking, not replace it

#599

Post by luckystarr »

kubb wrote: post #434
luckystarr wrote: post #391

Earlier quoted context omitted.

When the end result has problems and needs to be reworked. You can't figure this out instantly except when you'd review everything the LLM produces, which I am not. So the round trip time is pretty long, but I can trace it back to the intent now because I commit every architecture decision in an ADRs, which I pour most of my energy into. These are part of the repo. Using these ADRs helped a lot because most of the as…

Got it. I imagine concurrency bugs will hit hard with this approach because they show up rarely and are hard to debug.

That's why I use Rust and an Actor architecture. :)

Re: AI should elevate your thinking, not replace it

#600

Post by manoDev »

buddhistdude wrote: post #541
manoDev wrote: post #521

"Social media should bring people together instead of being used to set them apart." "Economy should improve allocation of resources universally, instead of allowing a handful to hold what many are lacking." And so on... Society has advanced in technology but didn't progress the same way in philosophy.

Well at least you for yourself can be aware of the danger that AI could assimilate your thinking, I think that's the message of this post.

The risk is the next generation, who will not have lived in a pre-AI world, won't know the difference to compare.
Post reply on HN