a megathread for developments on Reddit and with third-party Reddit apps
just so this doesn't overwhelm our front page too much, i think now's a good time to start consolidating discussions. existing threads will be kept up, but unless a big update comes let's try to keep what's happening in this thread instead of across 10.
I think this reply by spez has been badly overlooked:
“the LLM explosion put all Reddit data use at the forefront”.
What he means here is that earlier this year the board realised they were sitting on a massive gold mine, and their single focus right now is to exploit that as ruthlessly as possible. Jacking up the prices to access Reddit data to eye-watering levels is intended to fleece desperate AI bros, and this may well be the only revenue stream Reddit cares about in the future.
The fact that they have put no thought or care into managing the damage that this does to third party apps and to their own reputation with the Reddit user base tells me something else too. Why bother being a good custodian of a community website that has never made a profit, when you could live off selling access to one of the largest bodies of good quality human-generated text-based content out there?
Do they even care if Reddit goes to shit in the future? Maybe not, especially now we are beginning to realise how easy it is for careful bots to poison the conversations with AI-generated replies.
@myk@alyaza "Why bother being a good custodian of a community website that has never made a profit, when you could live off selling access to one of the largest bodies of good quality human-generated text-based content out there?"
Goes to show how important it is we use FOSS and decentralized tools for real community communications.
fleece desperate AI bros, and this may well be the only revenue stream Reddit cares about in the future.
Isn't it a bit late for that?
I mean, GPT is on its fourth iteration, they've been working on it for years, I don't know about Bing Chat but MS surely didn't start develop it only yesterday.
How can Reddit be so sure "AI bros" haven't already got the data they needed to train their models?
There’s going to be lots of other challengers out there: I’m sure every ML postgrad with any nous has spent the last couple of months contacting every funder they can track down to explain how their model is going to knock the socks off the old fashioned models used by these lumbering corporations.
And even the established models have been shown to contain content obtained in violation of user licences and copyright laws, leaving them open to all sorts of legal and political challenges. They will all be scrambling now to demonstrate that they’ve got clean hands in future models.
It will be like the NFT gold rush all over again—the only sure way to get rich is to sell the shovels.