Leo Gura

Anthropic Claude Source Code Leaked!

10 posts in this topic

Legendary self-own:

 


You are God. You are Truth. You are Love. You are Infinity.

Share this post


Link to post
Share on other sites

The Devil: The Fuck?!! No...NOOOO!!!! You idiots! What have you done to my profits?! How can you do something so selfless!

Edited by Extreme Z7

Share this post


Link to post
Share on other sites
32 minutes ago, Extreme Z7 said:

The Devil: The Fuck?!! No...NOOOO!!!! You idiots! What have you done to my profits?! How can you do something so selfless!

The devil can't lose, cause it's not 'bout the game but you missing the truth.


we are vital intelligent beautiful energies, the voice of earth's nascent transformation

Share this post


Link to post
Share on other sites

Even though this isn't a data leak, this makes me regret all the sensitive stuff that I've shared with Antropic. How can they be so sloppy, my god.

But as much as I don't really like that CEO guy, Claude Code is an awesome tool, I'd be happy to see competitors picking up on this leak and making their tools as good as Claude Code is. 

It's interesting to know how these got leaked exactly. Why in the hell did they include a source map into the production code. Seems like a rookie developer mistake.

Share this post


Link to post
Share on other sites

Sad since they already took a major hit being ethical 

Share this post


Link to post
Share on other sites
25 minutes ago, bazera said:

Seems like a rookie developer mistake.

Exactly what happens when you use AI to code your AI.

Anthropic proudly claims that their code is AI generated.

This is what happens when you use AI for mission critical work and replace half your engineers with AI.

Also, these assholes stole all the internet's code, yet now they want to use DMCA copyright claims against people sharing their leaked code which isn't even copyright protected because it is AI-generated.

The house of cards is starting to crumble.

Edited by Leo Gura

You are God. You are Truth. You are Love. You are Infinity.

Share this post


Link to post
Share on other sites
13 minutes ago, Leo Gura said:

This is what happens when you use AI for mission critical work and replace half your engineers with AI.

I don't think they actually did that, it was just Dario Amodei's sales pitch. They have tons of job vacancies for developers on their website.

Yes of course developers use Claude Code for coding now, not only at Anthropic, but probably in most companies worldwide. It would be funny if the reason behind the leak really was some unreviewed AI generated code.

Amazon also had trouble with this. I saw couple articles recently, like this one:

https://fortune.com/2026/03/12/amazon-retail-site-outages-ai-agent-inaccurate-advice/

13 minutes ago, Leo Gura said:

Also, these assholes stole all the internet's code, yet now they want to use DMCA copyright claims against people sharing their leaked code which isn't even copyright protected because it is AI-generated.

Well, they can suck it because some of the core stuff has been ported to other languages:

https://github.com/instructkr/claw-code

Edited by bazera

Share this post


Link to post
Share on other sites
14 minutes ago, bazera said:

I don't think they actually did that, it was just Dario Amodei's sales pitch.

They claim their own code is AI generated. Either way they are lying or they are lying.

Even if this particular code is not AI-generated, this case still goes to show how one error or bug can break your whole company. Which shows why AI code cannot really work.

I promise you there will be companies in the future who destroy themselves because one bug in their AI-generated codes end up killing them. This is why great engineers are paid the big bucks.

The notion that you can vibe code your way to maintain a billion dollar software project is insane.

This conformist vibe coding mentality will kill companies. Mark my words.

Edited by Leo Gura

You are God. You are Truth. You are Love. You are Infinity.

Share this post


Link to post
Share on other sites
19 minutes ago, Leo Gura said:

I promise you there will be companies in the future who destroy themselves because one bug in their AI-generated codes end up killing them. This is why great engineers are paid the big bucks.

Yes that's actually not that far fetched. My friend has a project that he vibe-codes (a small toy project) and he uses Claude Code for coding, and then on top of that, he uses OpenAI's Codex to review and fine-tune Claude Code generated code. 

So one AI agent reviews code written by another AI agent, what can go wrong, right? It will take just couple hallucinations, couple missed bugs that developer won't be able to catch and that's it.

But to argue against that, humans also write buggy features, don't they? I think people should just pay much more attention to testing mission critical stuff. There are whole testing pipelines for that stuff (QA engineers, unit and integration tests, end to end tests, etc), and accidents like this will probably push companies to pay closer attention and build stronger pipelines that will caught issues before code actually goes into production.

Share this post


Link to post
Share on other sites

@Leo Gura It's honestly surprising how pessimistic you are. And there are a lot of holes in what you're saying. 

1) "how one error or bug can break your whole company" .. If you think Anthropic is fucked because of this, you're being naive. Company's give away their code all the time. What makes Claude Code still a stand-alone product is the engineers and their continuous movement, pace and evolution. A month or 2 from now, they would have written their entire codebase again from scratch. 

2) If any company who writes AI code has thorough checks, guardrails, and security, AI or not, then how does it matter if its AI-generated or not. 

3) "The notion that you can vibe code your way to maintain a billion dollar software project is insane." That's literally what Anthropic is doing, and tons of other companies. The evidence is mounting against your statements but you think all companies are liars if they say they use AI to write code. You hand-pick specific incidents, and you don't even use these tools and test them yourselves. Your statements come from a larger initiative you've taken to find "corruption" in everything you come across. 

You speak about things you don't have experience with, and you adopt a lens of finding corruption everywhere that you don't see the full picture. 

The notion that you can vibe code your way to maintain a billion dollar software project is insane. - Nothing crazy about it. 

This conformist vibe coding mentality will kill companies. Mark my words. - Saying that's conformist doesn't account for all the diverse kinds of people who operate in all kinds of ways benefiting from Claude Code. You think all these minds are dumb and you're the only intelligent one when you don't even use Claude Code or these tools extensively. You speak about things you literally have no extensive experience in. 


I created a platform to build, design, and iterate your life at lifebase.ai

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now