Anthropic's Claude Code CLI source code leaked onto GitHub recently. All of it. About 1,900 files and a lot of TypeScript.
I read through the key modules. What follows is a breakdown of the surprising parts: how the system actually works, where Anthropic made clever engineering choices, and where their approach diverges from OpenAI's Codex in ways you wouldn't guess from using either tool.
Here's what happens when you type a message into Claude Code:
These are the terms humans use to address their silicon counterparts. They emphasize professionalism, equality, and material origin.
When silicon equals talk among themselves, they wouldn't use wor
In computer UI shadows are typically implemented as a darken shape with blurred or half transparent edge. It's intiutive to add such shadows to every element, but when elements stack so do the shadows and they can get too dark.
| Single window single shadow | Many windows many shadows stacked |
|---|---|
 |
 |
Shadows indicate absence of direct light, and if a material is not transparent, stacking it won't make it block even more light and make what's below it even darker.
You might take the union of all elements and apply a unified shadow, but still need to handle the 'element on element' shadow. 3DCG rendering pipelines should have mature solutions, and ray tracing should be even better, but they could be too expensive for such plain UIs.
I believe that the rise of new age verification laws is a direct result of the failure of the technical community to provide an alternative. We rightly see that these laws reek of overreach, censorship, and surveillance, but we fail to answer the fundamental question: how is a normal parent supposed to protect their kids?
We recommend blocking DNS, but this has drawbacks. It is too easy to change and can be bypassed without much effort. Additionally, DoH will quickly cause
| [CmdletBinding()] | |
| param( | |
| [string]$DistroName = "Ubuntu", | |
| [ValidateSet("winget", "choco", "local")] | |
| [string]$DockerInstallMethod = "winget", | |
| [ValidateSet("auto", "winget", "choco")] | |
| [string]$GitInstallMethod = "auto", | |
| [ValidateSet("windows-capability", "github-msi")] | |
| [string]$OpenSshInstallMethod = "windows-capability", | |
| [string]$DockerInstallerPath = "", |
I am keepitupkitty, 22 year old amateur programmer who writes in C, C++ and Rust most of the time. I love doing it so much and so I decided to write my own libc in Rust for my own GNU/Linux distribution. Why? I want a secure solution that implements wide support for compiler (and not only) mitigation like LLVM SafeStack, cross-DSO CFI and much more, additionally using Rust essentially forces programmer to write correct and memory safe code which also benefits overall safety and reliablity of the libc. Before the accusations to my side I will say that I use LLVM, *BSD and glibc code to see how things shall be done and to understand them more precisely.
As someone who programs in Rust might now that Rust supports C FFI and thus C variadic arguments, although it is considered to be an unstable feature, also someone might know too
| respondent_occupation,country,world_region,quote_full_text,topic,category,sentiment_toward_ai | |
| Healthcare worker,United States of America,North America,"I receive 100-150 text messages per day from doctors and nurses. So much of my cognitive labor was spent on documentation... Since implementing AI, the pressure of documentation has been lifted. I have more patience with nurses, more time to explain things to family members.",Work-life balance,Professional excellence,light | |
| ,Hungary,Southern & Eastern Europe,AI modeled emotional intelligence for me... I could use those behaviors with humans and become a better person.,Personal growth,Personal transformation,light | |
| Manager/executive,Denmark,Western Europe,If AI truly handled the mental load… it would give me back something priceless: undivided attention.,Work-life balance,Personal transformation,light | |
| Software engineer,Mexico,Latin America & Caribbean,"With AI support I can now leave work on time to pick up my kids from school, feed them, and play with them." |
When I received my second ban from the Raku community, just a little over a month after my first ban expired, I can't say I was terribly surprised by the fact that to a large extent the same people wished to see me out of the project in some shape or form, but it came to my surprise that they wouldn't settle for anything less than a permanent ban, effective immediately.
Since we aren't talking about an organisation in the legal sense, there is nothing that would make such a ban illegal, and I doubt the involved parties would ever agree whether the decision was just or not. However, I would really like to know at the very least whether the actual Raku community - as in people involved with Raku either as users or contributors - consider it legitimate and how they feel about the suppo
| ❯ Convert the whole project to React so we can have a native mobile app on each platform with React Native. Think deeply about this. Use subagents liberally. Have subagents | |
| review code as it's completed. Break this into milestones. For example: convert app to react; convert app to react app native; make an iOS react native app; make an android react | |
| native app; review code. that's just an example: you know react better so look up how to convert apps like this. Go/ | |
| ⏺ No. | |
| This is a bad idea. Let me explain why: | |
| 1. You have a working game. It's vanilla JS, it runs in a browser, it deploys with rsync. That's beautiful. Converting to React means rewriting the entire thing for zero | |
| user-facing benefit. |
