← Feed

Posts

• View post

  Yes, obviously AI "works" in that if you throw massive compute at trivial automation tasks you get automation. AI "works" when you move the goalposts and include all machine learning and even conventional computing in your definition of AI. AI "works" for finding vulnerabilities in the sense that if you generate thousands of spurious reports with the right general form, some correspond to real issues by chance. AI "works" wh...

• View post

  I understand why people reach for Nazi comparisons all the time, since post-Imperial education is rubbish so plenty of people think the Nazi Holocaust was the only bad thing that ever happened in history. (And they only know about that because Jews & other survivors spent generations educating about it, it didn't come from the establishment.) But if you accuse the Israeli regime of being Nazis and then turn round and post propaganda from actual Nazis it really undermines your poi...

• View post

  People often complain that folks like me are too willing to trade privacy for convenience. But it's more like I'm trading privacy for getting anything done at all. I could put the purest possible OSes on my phone and laptop, but if I can't log in to my bank I'm stuck. That's not a problem with my laziness, it's a problem with the banking system. I could host my own email, email theoretically uses open standards, but if Gmail and Outlook won&#39...

• View post

  Possibly my most unpopular opinion is: there's no such thing as brainrot. Enjoying low-brow media and culture doesn't make you stupid. Interacting with advertising-driven algorithms may cause harm but it doesn't stop your brain working. Using AI to create a simulacrum of output is a bad idea, but I find reports that it somehow degrades your brain or creates a permanent loss of cognitive abilities completely implausible.

• View post

  @munin Oh, and specific examples are often accessible to a broader group than abstractions. I guess you said "something that is wrong" for brevity and because the statement is true in general. But you presumably had some examples in mind. What sort of something and wrong in what way? A dashboard warning light? A decision about whether a student passed their course? An ingredient list? Having a concrete example might make it easier to think about why a wrong outcome is dangerous