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

AI makes the easy part easier and the hard part harder

blundergoat.com

Post reply on HN
Print view View original

231–240 of 382 posts

Re: AI makes the easy part easier and the hard part harder

#231

Post by kavalg »

le-mark wrote: post #9

I vibe coded a retro emulator and assembler with tests. Prompts were minimal and I got really great results (Gemini 3). I tried vibe coding the tricky proprietary part of an app I worked on a few years ago; highly technical domain (yes vague don’t care to dox myself). Lots of prompting and didn’t get close. There are literally thousands of retro emulators on github. What I was trying to do had zero examples on GitHub…

Same experience too. Even in some cases the AI was harmful, leading me into rabbit holes that did not pay off, but lost a whole day trying out.

Re: AI makes the easy part easier and the hard part harder

#232

Post by Eliezer »

Every time somebody writes an article like this without any dates and without saying which model they used, my guess is that they've simply failed to internalize the idea that "AI" is a moving target; nor understood that they saw a capability level from a fleeting moment of time, rather than an Eternal Verity about the Forever Limits of AI.

Re: AI makes the easy part easier and the hard part harder

#233

Post by kavalg »

derf_ wrote: post #169
ThunderSizzle wrote: post #116

Earlier quoted context omitted.

I've seen many discussions stating patent hoarding has gone too far, and also that copyright for companies have gone way too far (even so much that Amazon can remove items from your purchase library if they lose their license to it). Then AI begins to offer a method around this over litigious system, and this becomes a core anti-AI argument. I do think it's silly to think public code (as in, code published to the pub…

> I've seen many discussions stating patent hoarding has gone too far... Vibe coding does not solve this problem. If anything, it makes it worse, since you no longer have any idea if an implementation might read on someone else's patent, since you did not write it. If your agent could go read all of the patents and then avoid them in its implementations and/or tell you where you might be infringing them ( without hal…

> If your agent could go read all of the patents and then avoid them in its implementations and/or tell you where you might be infringing them (without hallucinating), that would be valuable.

That would lead the whole society to a halt, because it feels impossible to do anything now without violating someone's patent. Patents quite often put small players at a disadvantage, because the whole process of issuing patents is slow, expensive and unpredictable. Also, I once heard a lawyer say that, in high-stake lawsuits the it is the pile (of patents) that matters.

Re: AI makes the easy part easier and the hard part harder

#234

Post by strogonoff »

tern wrote: post #212
xeyownt wrote: post #207

Earlier quoted context omitted.

Funny how people only looks at the easy part, but not the cost part. "I did it with AI" = "I did it with an army of CPU burning considerable resources and owned by a foreign company." Give me an AI agent that I own and operate 100%, and the comparison will be fair. Otherwise it's not progress, but rather a theft at planetary scale.

If you want to play that game, you need to offer a fair comparison against the cost of "operating" the equivalent human being(s), measured in caloric input, waste output, etc.

Minimising the cost of “operating” humans means getting rid of humans. When human beings are not slaves and operate in a fair system, humans doing things is part humans living. Believe it or not, they may actually enjoy doing things; taking away the ability to do things we enjoy, and especially for compensation, is in fact harmful.

Re: AI makes the easy part easier and the hard part harder

#235

Post by bayindirh »

thedevilslawyer wrote: post #228
bayindirh wrote: post #217

Earlier quoted context omitted.

While this is from 2022, here you go: https://x.com/docsparse/status/1581461734665367554 I'm sure if someone prompts correctly, they can do the same thing today. LLMs can't generate something they don't know.

That you had to look and find this from 2022 proves my point..

Nope. That was a handy bookmark. I keep a list of these incidents, and other things:

https://notes.bayindirh.io/notes/Lists/Discussions+about+Art...

I have another handful of links to add to this list. Had no time to update recently.

Re: AI makes the easy part easier and the hard part harder

#236

Post by reverius42 »

blackqueeriroh wrote: post #155
irishcoffee wrote: post #100

Earlier quoted context omitted.

The models need to get burned down and retrained with these considerations baked in.

No. We need to light all IP law on fire. You shouldn’t able to license or patent software.

What about novels? Nonfiction books? Scientific papers? Poems? Those things are all in the training data too.

Re: AI makes the easy part easier and the hard part harder

#237

Post by lijok »

nirui wrote: post #230

I'm feeling people are using AI in the wrong way. Current LLM is best used to generate a string of text that's most statically likely to form a sentence together, so from user's perspective, it's most useful as an alternative to manual search engine to allow user to find quick answers to a simple question, such as "how much soda is needed for baking X unit of Y bread", or "how to print 'Hello World' in a 10 times in…

I think you severely overestimate your understanding of how these systems work. We’ve been beating the dead horse of “next character approximation” for the last 5 years in these comments. Global maxima would have been reached long ago if that’s all there was to it.

Play around with some frontier models, you’ll be pleasantly surprised.

Re: AI makes the easy part easier and the hard part harder

#238

Post by IX-103 »

thedevilslawyer wrote: post #210
20k wrote: post #181

Earlier quoted context omitted.

This is why its astonishing to me that AI has passed any legal department. I regularly see AI output large chunks of code that are 100% plagiarised from a project - its often not hard to find the original source by just looking up snippets of it. 100s of lines of code just completely stolen Ai doesn't actually wash licenses, it literally can't. Companies are just assuming they're above the law

This is oft-repeated but never backed up by evidence. Can you share the snippet that was plagiarized?

It happens often enough that the company I work for has set up a presubmit to check all of the AI generated and AI assisted code for plagiarism (which they call "recitation"). I know they're checking the code for similarity to anything on GitHub, but they could also be checking against the model's their training corpus.

Re: AI makes the easy part easier and the hard part harder

#239

Post by josefrichter »

The OP’s example of AI writing 500 LOC, then deleting 400, and saying it didn’t… Last time I saw something like that was at least a year ago, or maybe form some weaker models. It seems to me the problem with articles like this is while they sometimes are true at the moment, they’re usually invalidated within weeks.

Re: AI makes the easy part easier and the hard part harder

#240

Post by aembleton »

nirui wrote: post #230

I'm feeling people are using AI in the wrong way. Current LLM is best used to generate a string of text that's most statically likely to form a sentence together, so from user's perspective, it's most useful as an alternative to manual search engine to allow user to find quick answers to a simple question, such as "how much soda is needed for baking X unit of Y bread", or "how to print 'Hello World' in a 10 times in…

Which is great because then I can use my domain expertise to add value, rather than writing REST boilerplate code.
Post reply on HN