• 3 Posts
  • 13 Comments
Joined 1 year ago
cake
Cake day: June 1st, 2023

help-circle












  • It wants to keep control of how people get access to its data. The recent massive surge of interest in A.I.s means that there’s a lot of people looking for good quality datasets to train new models. Reddit is sitting on a goldmine, and it currently handing out gold nuggets for free.

    It wants to charge these desperate users of its data through the nose for that access, and $12,000 per 50M API calls is the market rate it has determined (and it is clearly comfortable that existing commercial users of its data such as marketers will also pay those rates).

    The fact that this will kill third party clients is just the icing on the cake. If reddit wanted to kill such clients it would just turn off voting and comments in the API.




  • It’s welcoming but confusing. I think there’s two reasons for the latter:

    1- Many of us forget how basic Reddit was when we first started using it, and the features we all know and love got added over time and repeatedly refined based on use.

    2- Most of us here are because we have been users of incredibly well designed apps crafted by developers with a passion for great UI. If I try using the (new) Reddit site or their default app, I find myself equally confused.

    There are still so many changes happening in Lemmy functionality, and as we’ve seen with Mastodon, we will hopefully soon be overwhelmed with great apps.

    In the meantime there’s the great community already here and growing. I saw a comment that you can estimate that Reddit has 90% lurkers, 9% commenters, 0.9% posters, and 0.1% “community builders” I think it’s those latter groups who are leading the exodus, which is great news for us and terrible news for whoever ends up owning Reddit.


  • At least he was honest enough to say that the sudden massive turnaround in their attitude to API pricing was their realisation they had the AI bros over a barrel.

    Given how much of the content used in the existing models has been shown to have been scraped in violation of usage licences and copyright, anyone who is serious about developing new models is going to be scrabbling desperately to get access to good data sets.

    The Reddit board now realise that charging through the nose for access to their API might generate them more money than all those annoying users ever have.