Hacker News Reader

Comment by berkle4455 | Hacker News Reader

You can blame this all you want on evil social media corporations, but the reality is AI companies scraping public conversations to feed LLMs are the current reason for walls being erected around every single garden. Facebook, Instagram, Twitter, Reddit, LinkedIn.. all heading toward full walled garden mode to prevent scrapers from repurposing data and profiting off their systems.

Federated systems are a nice idea, but they're not funded and will crumble under the same pressure until they too go into private mode. It's simply not a financially sound decision to run an open node that is continually harvested by corporations seeking to profit off the conversations occurring on your platforms.

View on HN

AI companies is a stupid argument. If the AI company operates legitimately, then TOS prohibiting using the content for LLM training purposes would be enough. If the AI company doesn’t want to play ball then restricting public access won’t stop them, they’ll just register accounts en-masse and scrape that way.

TOS does not apply if the data is not behind a login

Twitter has been slow since Musk took over and fired all of the competent devops people and then shut down most of the data centers. It's reasonable to assume that he's either been lied to or is lying about the cause of this.

Assuming social media companies are profitable via their normal services, then how they are hurt by downstream AI companies squeezing out a bit of leftover value? Because it feels like me being mad because I find out Jiffy Lube is making money reselling the used oil from my oil changes without giving me a cut.

I think it's a mistake to block third party providers from profiting from their service. First of all the hypocrisy in that all of these large companies exist because of massively profiting off of mostly uncompensated user-created content. Second, this drive toward relentlessly monetizing every aspect of your company's business is how we get degraded services like Microsoft putting ads in their search bar. It's one thing if a downstream OEM does it, I can just use an alternative OEM that doesn't shovelware the crap out of their product. But when the primary provider does it, then their service is permanently borked and eventually becomes unusable. So if the idea of blocking scrapers is because the goal is to eventually provide their own shitty AI services, I think Twitter et al are just going to end up killing their own geese -- making their own services unusable out of greed.

Reply on news.ycombinator.com