News for hackers. Stuff that matters.
Re:U.S. appeals court rejects big tech’s right to regulate o... by programmarchy on Saturday September 17, 2022 @02:08PM (#32879758)
Yeah this makes the most sense. There’s probably enough local compute power to run these filters on the client. But that’s a huge power shift and the change is likely more political than technological.
Re:U.S. appeals court rejects big tech’s right to regulate o... by lanstin on Saturday September 17, 2022 @05:29PM (#32881772)
Doubtful. With the rise of natural language spam, you have to analyze it in terms of networks or social graphs of bots reinforcing each other. E.g. a lot of bots will comment positively on other bots spam posts, in a manner very similar to my agreeing with some human. It is data bounded, not compute bounded. Will Twitter be forced to open their full internal conversation graph, including using real IPs probably for accurate targeting, to companies writing spam modules? It seems weirdly implausible.
Re:U.S. appeals court rejects big tech’s right to regulate o... by hedora on Saturday September 17, 2022 @10:58PM (#32884014)
Yelp's not recommended reviews are a great example of why this is data bounded (assuming Yelp's filter is halfway decent).