4 ms·
Google is automating things to a scary degree, with no way to appeal or ever even talk to a human. Unless you are a big corporation and have an account manager
by kerng 8y ago
Google is automating things to a scary degree, with no way to appeal or ever even talk to a human. Unless you are a big corporation and have an account manager things can be tough. Similar with GCP, countless horror stories over the years. I wonder if Google ever plans to address these issues or if in grand money scheme of things it just doesnt matter.
- mark-r 8y agoThis has been the way Google has worked from the beginning. Stories like this need to be shouted from the rooftops every time they happen, just to warn people. Google got to where they are today by building absolutely everything with a mind to scale, and justice doesn't scale.
- nikofeyn 8y agoturns out "don't be evil" doesn't scale.
- teddyh 8y agoNeither does “be evil”. Don’t anthropomorphize the paperclip maximizer.
- jessaustin 8y agoWas the greatest trick that Page, Brin, and Schmidt ever pulled, convincing the world that they don't exist? No, that's ridiculous. There are real people making real fortunes here. If they're not ethical, it doesn't take a conspiracy theorist to point that out.
- nikofeyn 8y agoi have no idea what that means.
- dcbadacd 8y agoThis game explains the paperclip maximizer: http://www.decisionproblem.com/paperclips/ http://www.decisionproblem.com/paperclips/
- teddyh 8y agoNo? Not as far as I can see. This is better: https://en.wikipedia.org/wiki/Instrumental_convergence#Paperclip_maximizer https://en.wikipedia.org/wiki/Instrumental_convergence#Paper...
- dcbadacd 8y agoYou haven't played it long enough to see how it becomes the paperclip maximizer, I posted the link to the game because everyone can find the Wikipedia link.
- nikofeyn 8y agoi can and did look it up, since i hadn't heard of it before. however, that doesn't answer the question as to what the commenter meant with their comment in the context of my reply. i meant "that" as in their whole comment in context.
- mark-r 8y agoI think the idea is that a goal maximizer can create evil actions even in the absence of evil intent.
- redler 8y agoOne might imagine a near future where some sort of automated deep learning system at Google makes these account termination decisions -- and not only can't they be appealed, but the reasoning behind the decision is unknowable, even to Google.
- spc476 8y agoI sometimes thing that's the case now.