Would the NYT have a weaker case if ChatGPT had not spit out a large chunk of verbatim content from the former's Wirecutter?

If so, that suggests that the problem isn't scraping the data but the failure to make it into something new. That is, ChatGPT failed to "generate" and simply "plagiarized". That seems eminently fixable by OpenAI.

If not, then this is a case about the legality of scraping, and everyone and his dog does that, not least Google.

If the problem is actually about scraping paywalled content, then that is yet a third thing, and I can see OpenAI having to negotiate hefty fees with the NYT, Reddit, and any other site that has high quality content.

Will be interesting to see the amicus briefs on this one.