Hacker News Reader

Comment by urbandw311er | Hacker News Reader

I wonder how many of those contributors, if re-consulted, would sign up to having their contributions used to train a for-profit LLM though?

I certainly didn’t sweat it out helping people on SO to pay for Sam Altman’s fucking swimming pool.

View on HN

I mean they would just scrape it if there's no data dump. It just makes it harder for the small guys. They probably scraped and are scraping HackerNews.

Generative AI doesn't follow copyright or even explicit software licenses as we have seen in AI art with human signatures and Microsoft Copilot.

They definitely scraped HN: one of my favorite ways to change the style of an LLM is to ask it to rephrase in the style of a snarky HN comment.

They probably didn’t even have to scrape it…

Sam Altman is still chairman of the board of directors of Y Combinator.

There was always the possibility of some sort of aggregator/other front end sitting on top of the SO data. We just didn't know exactly what a successful one would look like until relatively recently. I always limited how much I contributed based on that as likely outcome. Discontinuing the data dump is a much bigger deal to me and completely changes the value proposition of their various sites.

For what it's worth, as someone who has put a lot of writing online, I'm not bothered by having my writing including in the training sets of these LLMs. I write because I want to share knowledge, and it isn't important whether people get the knowledge directly from me versus mediated by friends, LLMs, etc.

Reply on news.ycombinator.com