I don't know why people obsess and spend so much time on this codebase. It isn't (and never was)alien technology. It's just mediocre typescript generated by an LLM
Claude Code Unpacked : A visual guide
231–240 of 464 posts
Re: Claude Code Unpacked : A visual guide
#232A 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;…
I find it really strange that there is so much negative commentary on the _code_, but so little commentary on the core architecture.
My takeaway from looking at the tool list is that they got the fundamental architecture right - try to create a very simple and general set of tools on the client-side (e.g. read file, output rich text, etc) so that the server can innovate rapidly without revving the client (and also so that if, say, the source code leaks, none of the secret sauce does).
Overall, when I see this I think they are focused on the right issues, and I think their tool list looks pretty simple/elegant/general. I picture the server team constantly thinking - we have these client-side tools/APIs, how can we use them optimally? How can we get more out of them. That is where the secret sauce lives.
Re: Claude Code Unpacked : A visual guide
#233A 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;…
> Right now, they're great for prompting simple sites/platforms but they break at large enterprise repos
Can you expand on this?
My experience is they require excessive steering but do not “break”
Re: Claude Code Unpacked : A visual guide
#234Earlier quoted context omitted.
They explicitly boast about using claude code to write code: https://x.com/bcherny/status/2007179836704600237 That's how you get "oh this TUI API wrapper needs 68GB of RAM" https://x.com/jarredsumner/status/2026497606575398987 or "we need 16ms to lay out a few hundred characters on screen that's why it's a small game engine": https://x.com/trq212/status/2014051501786931427
Just finished looking at Ink here.. frontend world has no shame. Love the gloating about 40x less RAM as if that amount of memory for a text REPL even approaches defensible. "CC built CC" is not the flex people seem to suggest it is.
Indeed a sad state of affairs.
Re: Claude Code Unpacked : A visual guide
#235A 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 not surprising. There has been quite a bit of industrial research in how to manage mere apes to be deterministic with huge software control systems, and they are an unruly bunch I assure you.
Sunir! Hope you are doing well man, I got a good chuckle from this.
Re: Claude Code Unpacked : A visual guide
#236A 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…
Oh, you should have a look at Pi then.
https://github.com/badlogic/pi-mono/tree/main/packages/codin...
Re: Claude Code Unpacked : A visual guide
#237Earlier quoted context omitted.
I just checked competitors' codebases: - 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.
All of them are really, REALLY bad.
Bad by whose definition? They work really well in my experience. They aren't perfect but the amount of hand holding has gone down dramatically and you can fix any glaring problems with a code review at the end. I work on a multimillion line code base which does not use any popular frameworks and it does a great job. I may be benefiting from the fact that the codebase is open source and all models have obviously been trained on it.
Re: Claude Code Unpacked : A visual guide
#238Re: Claude Code Unpacked : A visual guide
#239I built a site that lets you explore and browse all the Claude Code prompts in a structured way:
Re: Claude Code Unpacked : A visual guide
#240Same guide for opencode would be nice too