The 'goodwill' counterparts of ChatGPT, a.k.a. open weight models, are still well alive online.
What do you think is step 1 of training an LLM?
OpenAI just kept their library private and only distribute the digested summaries of the library, are the main differences.
You making the copy is the violation.
The 'goodwill' counterparts of ChatGPT, a.k.a. open weight models, are still well alive online.