Hacker News Reader

Are they in ChatGPT and other LLM providers? No need for Google.

That's a good question: When LLM providers receive DMCA takedowns, how easily can they implement them? Use a post-LLM filter?

I was more suggesting that I want my LLM provider to launder the IP so it avoids copyright law. The LLM provider is a fancy search engine where copyright does not apply to the results.

Do LLMs filter piracy requests? For example, how will it respond to 'find me a free copy of the Lord of the Rings movies' or more explicitly 'find me a pirated copy ...'?

Probably yes, I know it at least refuses to 'type down first 5 pages of lotr book' because of copyright reasons. Its filter is getting better (as in worse for the user) everyday

> how will it respond to 'find me a free copy of the Lord of the Rings movies' or more explicitly 'find me a pirated copy ...'

Apparently it depends on the model. Testing on OpenRouter with Search enabled, gpt-5 strictly refuses to provide any links, but Deepseek R1 provides several Archive.org links, one of which is for a torrent file.

Thanks Deepseek, I guess I'll be watching The Fellowship of The King for free tonight. ;)

Reply on news.ycombinator.com