A 500k line codebase for an agent CLI proves one thing: making a probabilistic LLM behave deterministically is a massive state-management nightmare. Right now, they're great for prompting simple sites/platforms but they break at large enterprise repos. If you don't have a rigid, external state machine governing the workflow, you have to brute-force reliability. That codebase bloat is likely 90% defensive programming;…
Claude Code Unpacked : A visual guide
191–200 of 464 posts
Re: Claude Code Unpacked : A visual guide
#192A 500k line codebase for an agent CLI proves one thing: making a probabilistic LLM behave deterministically is a massive state-management nightmare. Right now, they're great for prompting simple sites/platforms but they break at large enterprise repos. If you don't have a rigid, external state machine governing the workflow, you have to brute-force reliability. That codebase bloat is likely 90% defensive programming;…
> A 500k line codebase for an agent CLI proves one thing: making a probabilistic LLM behave deterministically is a massive state-management nightmare. Considering what the entire system ends up being capable of, 500k lines is about 0.001% of what I would have expected something like that to require 10 years ago. You can combine that with all the training and inference code, and at the end of the day, a system that li…
You really need to compare it to the model weights though. That’s the “code”.
Re: Claude Code Unpacked : A visual guide
#193A 500k line codebase for an agent CLI proves one thing: making a probabilistic LLM behave deterministically is a massive state-management nightmare. Right now, they're great for prompting simple sites/platforms but they break at large enterprise repos. If you don't have a rigid, external state machine governing the workflow, you have to brute-force reliability. That codebase bloat is likely 90% defensive programming;…
[flagged]
Re: Claude Code Unpacked : A visual guide
#194Earlier quoted context omitted.
It's a TUI API wrapper with a few commands bolted on. I doubt it needs to be more than 20-50kloc. You can create a full 3D game with a custom 3D engine in 500k lines. What the hell is Claude Code doing?
Software doesn’t end at the 20k loc proof of concept though. What every developer learns during their “psh i could build that” weekendware attempt is that there is infinite polish to be had, and that their 20k loc PoC was That said, doesn't TFA show you what they use their loc for?
Re: Claude Code Unpacked : A visual guide
#195Earlier quoted context omitted.
It's a TUI API wrapper with a few commands bolted on. I doubt it needs to be more than 20-50kloc. You can create a full 3D game with a custom 3D engine in 500k lines. What the hell is Claude Code doing?
Comments like these remind me of the football spectators that shout "Even I could have scored that one" when they see a failed attempt. Sure. You could have. But you're not the one playing football in the Champions League. There were many roads that could have gotten you to the Champions League. But now you're in no position to judge the people who got there in the end and how they did it. Or you can, but whatever.
Re: Claude Code Unpacked : A visual guide
#196A 500k line codebase for an agent CLI proves one thing: making a probabilistic LLM behave deterministically is a massive state-management nightmare. Right now, they're great for prompting simple sites/platforms but they break at large enterprise repos. If you don't have a rigid, external state machine governing the workflow, you have to brute-force reliability. That codebase bloat is likely 90% defensive programming;…
It's hard to tell how much it says about difficulty of harnessing vs how much it says about difficulty of maintaining a clean and not bloated codebase when coding with AI.
Re: Claude Code Unpacked : A visual guide
#197> 500k lines of code Isn't it a simple REPL with some tools and integrations, written in a very high level language? How the hell is it so big? Is it because it's vibecoded and LLMs strive for bloat, or is it meaningful complexity?
- Opencode (anomalyco/opencode) is about 670k LOC
- Codex (openai/codex) is about 720k LOC
- Gemini (google-gemini/gemini-cli) is about 570k LOC
Claude Code's 500k LOC doesn't seem out of the ordinary.
Re: Claude Code Unpacked : A visual guide
#198Earlier quoted context omitted.
The history of programming is also each generation writing far less performant code than the one before it. The history of programming is each generation bemoaning the abstractions, waste and lack of performance of the code of the next generation. It turns out that there is a tradeoff in code between velocity and quality that smart businesses consider relative to hardware cost/quality. The businesses that are outcomp…
You asked why the size of the code matters, I gave you the answer. If you want to ramble about the non technical aspects of software development talk to someone else, I'm not interested.
Re: Claude Code Unpacked : A visual guide
#199Earlier quoted context omitted.
What do you mean by "actually governing the agents at the system level", and how is it different from "herding cats"?
Herding cats is treating the LLM's context window as your state machine. You're constantly prompt-engineering it to remember the rules, hoping it doesn't hallucinate or silently drop constraints over a long session. System-level governance means the LLM is completely stripped of orchestration rights. It becomes a stateless, untrusted function. The state lives in a rigid, external database (like SQLite). The database…
Re: Claude Code Unpacked : A visual guide
#200Earlier quoted context omitted.
But somehow, according to HN, LLMs make you less productive, not more :)
The people who don’t know how to use an LLM to make them more productive, or are scared it’s going to take their job, are louder than the people who are making good use of them to make them more productive. That just seems to be human nature unfortunately - the complainers are always louder.
Those within well informed, technical circles will fall somewhere in between the for/against labels, myself included.
The GenAI hype cycle is finally starting to collapse as the general population starts to realize that these systems aren't the panacea for "everything" after all. They provide enormous utility in some domains like coding, but even then there are massive tradeoffs, footguns and the usual horse blinder ills that come with every hype cycle. I just hope we stop having to "learn the hard way" with respect to undisciplined use of current-gen LLM systems writ large, and cooler heads prevail sooner rather than later.