Hacker News Reader

Comment by artdigital | Hacker News Reader

Same here, I did a few small taks with Claude Code after seeing this discussion here and is too expensive for me.

A small change to create a script file (20 LoC) was 10cts, a quick edit to a README was 7ct

Yes yes engineers make more than that blah blah but the cost would quickly jump out of control for bigger tasks. I’d easy burn through $10-20 upwards a day with this, or upwards $100-$300 a month. Unless you have a Silicon Valley salary, that’s too expensive.

I use other tools like Cody (the tool the author created) or Copilot because I pay $10 a month and that’s it. Yes I get rate limited almost daily but I don’t need to worry that my tool cost is going out of control suddenly.

I hope Anthropic introduces a new plan that bundles Claude Code into it, I’d be much more comfortable using that knowing it won’t suddenly be more than my $50/mo (or whatever)

View on HN

It's an interesting question. As a freelance consultant, theoretically a tool like this could allow me to massively scale up my income, assuming I could find enough clients.

I'm a bit nervous where I'd end up though - with code I'd "written" but wasn't familiar with, and with who knows what kinds of limitations or subtle bugs baked in.

I currently pay around $200-300 to a combination of Cursor + Anthropic through the API. I have both a full time job and freelance work. It pays for itself. I end up reviewing more than manual coding, to ensure the quality of the results. Funnily, the work I did through this method has received more praise than my usual work.

Did you outgrow the vase 500 searches that Cursor gives you per month and connect your API key for usage based pricing?

I’m having a hard time coming close to the 500 included in the monthly subscription and I use it like, a lot.

Just curious how you’re hitting that 200-300 mark unless you’re talking about paying Anthropic outside of cursor. Which I just now realized is probably the case.

I ran out of fast requests and using my own API key

> I'm a bit nervous where I'd end up though - with code I'd "written" but wasn't familiar with

This does seem like quite a big downside. It turns every new feature into “implement this in someone else’s code base”. I imagine you’d very quickly have complete dependency on the AI. Maybe that’s an inevitability in this new world?

>you’d very quickly have complete dependency on the AI

That's a feature, not a bug. At least from OAI/Anthropic's point of view.

It sounds fine as long as you can fully trust the AI to do good work right?

I don't think there's any current AI that is fully trustworthy this way though.

I wouldn't even put them at 50% trustworthy

I think we are going to see a cliff where they become 80% good, and every tiny bit of improvement past that point will be exponentially more difficult and expensive to achieve. I don't think we reach 100% reliable AI in any of our lifetimes

I think we are going to reach a cliff where a type of old school developers keep saying, "it just can't write code like I can" while at the same time wondering why they can't land a job.

Current AI is likely already beyond 50% trustworthiness, whatever that means.

> "it just can't write code like I can" while at the same time wondering why they can't land a job

People had this same prediction about offshore development

Those old school devs are able to find well paying work fixing broken software churned out by overseas code sweatshops

I predict if you can read and understand code without the help of AI models you will be in even higher demand to fix the endless broken software built by AI assisted coders who cannot function without AI help

>It turns every new feature into “implement this in someone else’s code base”.

It amazes me that more people aren't worried about this.

> Yes yes engineers make more than that blah blah but the cost would quickly jump out of control for bigger tasks.

Also (most) engineers don't hallucinate answers. Claude still does regularly. When it does it in chat mode via a flat rate Pro plan I can laugh it off and modify the prompt to give it the context it clearly didn't understand but if its costing me very real money for the LLM to over-eagerly over-engineer an incorrect implementation of the stated feature its a lot less funny.

Exactly! Especially agentic tools like Aider and Claude that are designed to pull in more files into their context automatically, based on what the LLM thinks it should read. That can very quickly go out of control and result in huge context windows.

Right now with Copilot or other fixed subscriptions I can also laugh it off and just create a new tab with fresh context. Or if I get rate-limited because of too much token use I can wait 1 day. But if these actions are linked to directly costing money on my card, then that's becoming a lot more scary.

> Also (most) engineers don't hallucinate answers.

They absolutely do, where do you think bugs come from? The base rate is typically just lower than current AIs.

Bugs from engineers comes from a variety of reasons and most have nothing in common with an LLM hallucinating.

For exemple I can’t remember seing a PR with an API that seems plausible but never ever existed, or an interpretation of the specs so convoluted and edgy that you couldn’t even use sarcasm as a justification for that code.

Don’t take me wrong: some LLMs are capable of producing bugs that looks like humans ones, but the term hallucinate is something else’s and doesn’t fit with much humans bugs.

> For exemple I can’t remember seing a PR with an API that seems plausible but never ever existed

A PR is code that has already been tested and refined, which is not comparable to the output of an LLM. The output of an LLM is comparable to the first, untested code that you wrote based off of your sometimes vague memory of how some API works. It's not at all uncommon to forget some details of how an API works, what calls it supports, the details of the parameters, etc.

It’s kind of uncommon to be aware that you have only a vague recall of the API and not go check the documentation or code to refresh your memory. That self knowledge that you knew something and aren’t sure of the details is indeed the thing that these tools lack. So far.

Human programmers have continuous assistance on every keystroke - autocomplete, syntax highlighting, and ultimately, also the compilation/build step itself.

For an LLM-equivalent experience, go open notepad.exe and make substantial changes there, and then rebuild - and let the compiler tell you what's your base rate of hallucinating function names and such.

In the 1990s, that is closer to what making software was like. There, one had an even more heightened awareness of how confident one was in what one was typing. We would then go to the manual (physical in many cases) and look it up.

And we never made up APIs, as there just weren't that many APIs. We would open the .h file for the API we were targeting as we typed into the other window. And the LLMs have ingested all the documentation and .h files (or the modern equivalent) so they don't have a real excuse.

But I use the LLMs all the time for math, and they do profusely hallucinate in a way people do not. I think it's a bit disingenuous to say that LLMs don't have that failure mode that people don't really have.

But you don't submit that rough draft with the 110% conviction that it's correct, which is what an LLM will do when it hallucinates.

It won't say "I think it should look something like this but I might be wrong," it'll say "simple! here is how you do it."

Hence hallucination and not error. It thinks it's right.

I use Grok and it's free (even Grok3). I definitely don't hit limits unless it's a pretty heavy day and I do a lot of adjustments. Also, don't send entire codebases to it, just one-off files. What's quite amazing is how it doesn't matter that it doesn't have the source to dependent files, it figures it out and infers what each method does based on its name and context, frigging amazing if you ask me.

And it doesn't fight me like the OpenAI tooling does that logs me out randomly every day and I have to login and spend 4 minutes copying login codes from my email or answering their stupid Captcha test. And this is on their API playground where I pay for every single call - so not like I'm trying to scrape my free chat usage as an API.

You send your company's trade secrets directly to Elon Musk?

So? Most of your code is worthless to anyone except yourself.

> I use Grok and it's free (even Grok3). I definitely don't hit limits unless it's a pretty heavy day and I do a lot of adjustments

Okay maybe need to clarify: I hit those limits when I do agentic stuff, which is what Claude Code does: So let the LLM automatically pull in files into the context it thinks it needs, analyze my codebase, follow imports, add more code, etc. It can quickly balloon out of control when the LLM pulls in too many LoC and the context window gets too big.

Then do a few back and forth actions like "let's refine this plan, instead of X pls do Y", or "hmm I think maybe we should also look into file blah.ts" and you quickly hit 500k tokens.

If I use Cody only, which has some agentic capabilities but is much more "how can I implement Y in this file @src/file1.ts db models are in @src/models/foo.ts", then I rarely ever hit any rate limitations. That's more similar to what you describe of copying code back and forth, except it's in the editor and you can do it by writing @somefile.

Reply on news.ycombinator.com