A significant portion of content creators do it as a full time job. Ads is a stable revenue stream for these content creators. You as a consumer are free to move to any decentralized video platform you like, but the main bottleneck for any significant shift in market share depends on content creators, and ultimately, on advertisers.
> Roku alleges Google has asked it to favor YouTube music results from voice commands made on the Roku remote while the YouTube app is open, even if the user's music preference is set to default to another music app, like Pandora.
How else do you use voice search for a music video on Youtube? If I open youtube and do a voice search. I'm expecting the search to be constrained to the app.
> So if I just assign the majority label to all of the population of a given demographics group, I would get the same result right? i.e., predicting "left" for all minorities under 30. You would also get ~70% accuracy.
I meant that even if you control age, gender, ethnicity, a very trivial predictor (i.e., always predicting the majority label) could yield similar performance. What I meant to say was that their model may not perform as well as they made it sound.
My interpretation of control is fixing all other variables (the ones they mentioned) except for the one being measured (political orientation). If that's not what they did I'm happy to learn.
> "Accuracy remained high (69%) even when controlling for age, gender, and ethnicity."
So if I just assign the majority label to all of the population of a given demographics group, I would get the same result right? i.e., predicting "left" for all minorities under 30. You would also get ~70% accuracy.
> I have profound sympathy for the victims of authoritarian regimes, and I absolutely do not believe that the way to help them is to collaborate with the very system that oppresses them. Pretending otherwise sounds like Newspeak to me.
I have relatives in China. Censorship is the least of their worries when the current Baidu monopoly is allowing promoted results that are sometimes flat-out evil (https://en.wikipedia.org/wiki/Death_of_Wei_Zexi).
What you said is not entirely true. Daraprim (the drug you are referring to) costs $10 out of pocket for insured patients. For uninsured low income patients it's free of charge.
That's because your interview process is already biasing the result. To get a fairer comparison you should take a random sample from both groups (out of all the applicants) and see (although self selection bias is still in play here).
I have worked in healthcare related systems before that needs to be HIPAA compliant, even for those systems public disclosure of a vulnerability is not a requirement. No software is bug free, and many seemingly benign bugs are security vulnerabilities.
Try and name one company that reports all their bugs (security/non security) discovered internally.