a thoughtful web.
Good ideas and conversation. No ads, no tracking.   Login or Take a Tour!
comment by veen
veen  ·  479 days ago  ·  link  ·    ·  parent  ·  post: I feel like we need to talk about AI language models

One of the lenses through which to view socially impactful technology is that they are technologies that move too fast for our society to adapt to it - the boomers can't handle misinformation bots, the millenials are all addicted to their phones, yadda yadda. It's not a perfect metaphor but it is a useful way to think about this, because these AI models feel like a large new virus when our society hasn't yet fully developed the antibodies to withstand the previous changes. I'm still having a hard time teaching my parents to distinguish phishing mails from regular ones, and now there's the chance that someone writes a fully personal email automatically?! It might just make emails 10x less useful in the process.

It's not that I need this, it's that it is becoming stupidly easy to wrap these models around any nefarious text-based thing (which is most of the internet) and I feel like we do not have the tools to deal with this soon enough. So even if this does end up screwing most of social media and the internet over, if the patient is dead is the operation still succesful?





kleinbl00  ·  479 days ago  ·  link  ·  

Sure. Model T introduced in 1908; seat belts mandatory for all passenger vehicles 1968.

Social innovations will always kill off the bold and the slow. In general, those innovations that are judged to be a net negative are legislated against faster (lawn darts, MDMA, vaping) than those of mixed use (gasoline, VoIP).

I think if you said "hey I have invented gasoline" in the social media environment of 2022 it would be a controlled substance within a week. But I see that as a legislation problem, not an innovation problem... and our legislation around the internet has been stupidly piecemeal for a stupidly long time.