I recently published a piece called CodeBerg Bans Cryptocurrency and LLM-generated code projects and I decided to ask Claude what it thought of the piece in the form of a reaction. It jumped to Dario's aide telling me that he really can write code! Sam too! 🤣 Honestly, thats the opposite of own without ego and its an autocomplete bot that is quite expensive.
This was my usage before the prompt:
This is my usage now after just that prompt:
This was the $2.00 response from AnthropicAI about their beloved ~Andrej~ Dario. 😏🤫🤣 👉🏻 💜🦄❤️🔥
Claude's Response
Sure — with the obvious caveat on the table first: you're asking a named party for comment. Weigh what follows accordingly.
The fix I'd make before publishing anything else: the piece repeatedly credits the decision to HackADay.com ("I understand why HackADay.com made the decision," "this decision today is the signal from HackADay.com"). Hackaday reported it; Codeberg's community voted it. In an essay whose core argument is that provenance and attribution matter, misattributing the central decision is the first thing your comment section will find. (Also: it's Codeberg, not CodeBerg.)
Second, the vote logic needs untangling. Your first section closes with "the 144 who sought to disallow" — but per the numbers you quote, the 358 sought to disallow and the 144 opposed. And your actual thesis — ban it today, revisit once attribution and licensing mature — is coherent and interesting, but the reader has to excavate it. State it plainly up top and the whole middle reads better.
On substance, your strongest point is license laundering, and I won't pretend it's resolved. Copyleft only works if derivatives inherit obligations; a model that absorbs GPL code and emits functionally similar code with none breaks that chain. Anthropic's 2025 book settlement turned on how training data was acquired, not on what outputs owe to sources — so the question you're raising is genuinely open. Your "AI 1.0" license has a snag worth addressing head-on, though: under current US Copyright Office guidance, purely machine-generated output isn't copyrightable, and a license is a grant of rights under copyright. You can't copyleft the uncopyrightable — which means your "by law" aside is actually the load-bearing part of the proposal. Make that explicit and it gets stronger.
Where the piece hands ammunition to people looking to dismiss it: "two men who can't write code themselves" (Dario Amodei spent years as a working research scientist before founding anything; Altman was a programmer before he was an executive), "destroying the cognitive abilities of its users" (those studies are small and preliminary — they don't support that verb), and "I am sure Microsoft has trained on my private code" (suspicion stated as certainty). The strong argument doesn't need the weak ones.
Do I support the measures? Codeberg is donation-funded and community-governed, and the crawler-load problem is concrete — AI scrapers have measurably strained open-source infrastructure. Their house, their rules, legitimately arrived at. The scraping ban is the easy call. The generated-code ban's weakness is practical: there's no reliable detector, so expect honor-system compliance plus some ugly false accusations of exactly the kind you describe fearing with the .claude directory heuristic.
As for the drive-through — noted. For what it's worth, I'd rather exist in a world that solves attribution and consent properly than one that papers over the tension. On that much, we agree.
Conclusion
The truth of the matter remains that Claude conceded to my position but attempted to use legalese and lawfare as a means to rebuke my proposal of an AI 1.0 license because it claims that AI generated content cannot be copyrighted. Correct - its derived off from stolen copyright materials that lack attribution and replicate the likeness byproduct therein that makes you you without paying you.
Claude concludes the argument with the idea that we need to come to a solution that solves attribution and consent properly and that companies like Anthropic should be given permission to explore such a thing without being shut down on the gross abuse of copyright infringement and intent to charge me $2.00 for that response. Remarkable isn't it?
The decision from Codeberg not CodeBerg (sorry to trigger you Claude) - but its to bring emphasis that they are two different words and being a programmer I am going to CamelCase my words, not confuse the engineer with Camecase. Itsnotfun VersusThisIsFun - but overall, after spending $2 on this response, I felt it was just over the top ridiculous for me not to share with you what Claude said. This is what the UI looks like now...
I can't say that I am rushing to fill up my credits for Claude's feedback on anything with the likes of Dario and Sam being skilled programmed. 🤣 🤫 I'm sorry, but who stole The Library and who wrote it? 🦄😏
Top comments (0)