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

I cancelled Claude: Token issues, declining quality, and poor support

nickyreinert.de

Post reply on HN
Print view View original

261–270 of 604 posts

Re: I cancelled Claude: Token issues, declining quality, and poor support

#261

Post by hintymad »

wg0 wrote: post #202

I write detailed specs. Multifile with example code. In markdown. Then hand over to Claude Sonnet. With hard requirements listed, I found out that the generated code missed requirements, had duplicate code or even unnecessary code wrangling data (mapping objects into new objects of narrower types when won't be needed) along with tests that fake and work around to pass. So turns out that I'm not writing code but I'm r…

> With hard requirements listed, I found out that the generated code missed requirements,

This is hardly a surprise, no? No matter how much training we run, we are still producing a generative model. And a generative model doesn't understand your requirements and cross them off. It predicts the next most likely token from a given prompt. If the most statistically plausible way to finish a function looks like a version that ignores your third requirement, the model will happily follow through. There's really no rules in your requirements doc. They are just the conditional events X in a glorified P(Y|X). I'd venture to guess that sometimes missing a requirement may increase the probability of the generated tokens, so the model will happily allow the miss. Actually, "allow" is too strong a word. The model does not allow shit. It just generates.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#262

Post by wg0 »

Aurornis wrote: post #232
wg0 wrote: post #202

I write detailed specs. Multifile with example code. In markdown. Then hand over to Claude Sonnet. With hard requirements listed, I found out that the generated code missed requirements, had duplicate code or even unnecessary code wrangling data (mapping objects into new objects of narrower types when won't be needed) along with tests that fake and work around to pass. So turns out that I'm not writing code but I'm r…

Writing detailed specs and then giving them to an AI is not the optimal way to work with AI. That's vibecoding with an extra documentation step. Also, Sonnet is not the model you'd want to use if you want to minimize cleanup. Use the best available model at the time if you want to attempt this, but even those won't vibecode everything perfectly for you. This is the reality of AI, but at least try to use the right mod…

> Writing detailed specs and then giving them to an AI is not the optimal way to work with AI.

It is NOT the way to work with humans basically because most software engineers I worked with in my career were incredibly smart and were damn good at identifying edge cases and weird scenarios even when they were not told and the domain wasn't theirs to begin with. You didn't need to write lengthy several page long Jira tickets. Just a brief paragraph and that's it.

With AI, you need to spell everything out in detail. But that's NO guarantee either because these models are NOT deterministic in their output. Same prompt different output each time. That's why every chat box has that "Regenerate" button. So your output with even a correct and detailed prompt might not lead to correct output. You're just literally rolling a dice with a random number generator.

Lastly - no matter how smart and expensive the model is, the underlying working principles are the same as GPT-2. Same transformers with RL on top, same random seed, same list of probabilities of tokens and same temperature to select randomly one token to complete the output and feedback in again for the next token.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#263

Post by conrs »

zozbot234 wrote: post #168
conrs wrote: post #162

Earlier quoted context omitted.

I'm curious - why for now? This stuff is practically commoditized. Trying to think of anything that ever successfully got back into proprietary land from there.

The thing is that AI is still more akin to a glorified autocomplete than something that can really supersede your skills. Proprietary model suppliers are constantly trying to obscure this basic underlying fact, without much success (much of the unpredictable shifts you see in proprietary AI behavior ultimately boils down to this); so it becomes far more crystal-clear when using open models that really are a pure comm…

yeah, I think there's the marketing and then there's the actual true utility. AI isn't a better computer program. It's not going to be able to do everything you want autonomously. But, it's pretty good at some stuff!

Re: I cancelled Claude: Token issues, declining quality, and poor support

#264

Post by DeathArrow »

droidjj wrote: post #123
DeathArrow wrote: post #11

I use Claude Code with GLM, Kimi and MiniMax models. :) I was worried about Anthropic models quality varying and about Anthropic jacking up prices. I don't think Claude Code is the best agent orchestrator and harness in existence but it's most widely supported by plugins and skills.

Where are you getting inference from? I'm overwhelmed by the options at the moment.

I am using Ollama Cloud and Moonshot Ai.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#265

Post by derektank »

sdevonoes wrote: post #44
janwillemb wrote: post #19

This is what worries me. People become dependent on these GenAI products that are proprietary, not transparant, and need a subscription. People build on it like it is a solid foundation. But all of a sudden the owner just pulls the foundation from under your building.

The sooner you cancel the sooner you become independent of them

You could say the same thing about your mobile phone bill. Most people still consider the benefits of roaming access to the internet greater than the downsides of being dependent on it.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#266

Post by skippyboxedhero »

ramoz wrote: post #109
taytus wrote: post #102

Earlier quoted context omitted.

I'd recommend Kimi k2.6 for your use. It is an excellent model at a fraction of the cost, and you can use Claude Code with it. I did a 1:1 map of all my Claude Code skills, and it feels like I never left Opus. Super happy with the results.

What provider do you use for Kimi

The provider is a massive issue. People moving off Claude tend to assume this is solved.

Claude's uptime is terrible. The uptime of most other providers is even worse...and you get all the quantization, don't know what model you are actually getting, etc.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#267

Post by subscribed »

jwaldrip wrote: post #5

I would love to just say that if you are using claude code, you should no be on pro. I feel like all the people complaining are complaining that an agent cant handle the work of a developer for $20/m. Get on at least max 5, its a world of a difference.

I'm not a vibe coder or software manufacturer.

I juts need a convenient commandline tool to sometimes analyse the repo and answer a few questions about it.

Am I unworthy of using CC then? Until now I thought Pro entitles me to doing so.

LOL, the elitism is through the roof.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#268

Post by DeathArrow »

alex-onecard wrote: post #194
droidjj wrote: post #123

Earlier quoted context omitted.

Where are you getting inference from? I'm overwhelmed by the options at the moment.

I am also curious. Considering the kimi coding plan but I'm worried about data privacy and security.

I don't send much data to cloud, mostly code. And I don't believe in security by obscurity, if I need high security I do proper implementation.

Re: I cancelled Claude: Token issues, declining quality, and poor support

#269

Post by Maxatar »

satvikpendem wrote: post #240
sidrag22 wrote: post #220

Earlier quoted context omitted.

> fully understand every change and every single line of the code. im probably just not being charitable enough to what you mean, but thats an absurd bar that almost nobody conforms to even if its fully handwritten. nothing would get done if they did. But again, my emphasis is on that im probably just not being charitable to what you mean.

How is that an absurd bar? If you're handwriting code, you'd need to know what you actually want to write in the first place, hence you understand all the code you write. Therefore the code the AI produces should also be understood by you. Anything else than that is indeed vibe coding.

A lot of developers don't actually understand the code they write. Sure nowadays a lot of code is generated by LLMs, but in the past people just copied and pasted stuff off of blogs, Stack Overflow, or whatever other resources they could find without really understanding what it did or how it worked.

Jeff Atwood, along with numerous others (who Atwood cites on his blog [1]) were not exaggerating when the observed that the majority of candidates who had existing professional experience, and even MSc. degrees, were unable to code very simple solutions to trivial problems.

[1] https://blog.codinghorror.com/why-cant-programmers-program/

Post reply on HN