drhead [he/him]

they’ll probably take information representing some aggregation of interactions for a user and make some scoring model that tries to learn from pairs of user data and outcomes (in terms of whether they successfully dated or whatever you do on these apps). 100% marketing bullshit, doing LLM inference for something like this would have costs spiral out of control FAST. But a scoring model is cheap, they have the data to make one, and it isn’t really all that innovative either.

You wouldn’t download an overclock.

Main difference is that blockchain never had an actual use case (speculation doesn’t count) beyond buying heroin and running ransomware. Machine learning had practical applications for years that nobody really thought much of at the time, and the marketers got a hold of it after it was fairly well established without them and right at the point of a massive wave of breakthroughs in the area.

That being said, there is a fucking massive AI bubble. A large portion of the things we’re seeing will survive when that pops, but boy are there a lot of very overconfident investors who are going to get burned hard on this.

Should also probably include some recommendations for less expensive routers that can use OpenWrt along with this? Mine cost me $300… I don’t regret it because I have found at least that much utility in it, from this and from finally being able to intercept my smart TV’s hardcoded DNS requests and blocking the ads, but I’m pretty sure that there are better deals to be had that don’t involve paying $300 for one of the radios to not work due to bad driver support.

Bias is important for credibility of a source, but not for the validity of the argument presented, and for the latter you actually have to understand and think about the argument presented.

The most important part of that page is its argument that all states wield authority and tend to tighten or relax the exercise of that authority in order to serve a given set of class interests. There’s nothing in this that relies on credibility, and dismissing it on account of bias makes as much sense as responding to someone in a debate by saying “you’re biased, so why should I believe you?”.