Right - I see the other topic has been locked / author hasn’t returned to follow community rules.
Here’s the original topic https://news.ycombinator.com/item?id=49003386
and a measured (IMHO) response to it. (BTW, do your self a favour and change tabs with that site open :)
https://マリウス.com/i-regret-migrating-to-codeberg/
It’s a tough spot, Codeberg has found themselves in and I wish them luck. But beyond that, this is (yet) another reminder that in 2026, if you don’t self host it, the cloud is just someone else computer
Not true, this does not occur frequently. This study and software LiCoEval from 2024 found 0.88% to 2.01% of code “strikingly similar to existing open-source implementations”. Afaik this is mostly textbook examples, snippets from stack overflow snippets or common github repositories, often replicated api examples and language boilerplate. How you prompt and refine also matters, and for generating novel code or business logic the LLM simply cannot use memorized snippets.
Presumably since then LLMs have worked to reduce that number of memorized code. Since LLMs cannot memorize all their training data, that number is limited. LiCoEval can find the often memorized examples and train to remove them, or suppress them, or they find other ways to reduce direct reproduction from memorization. For example it would be possible to do what malus.sh does with all the training data. Then it cannot memorize copyrighted code.
So for a model that came out 2026 this already small number might not be that relevant anymore.