With regard to the dangers of surveillance, I've made a sustained argument about this in other talks. It boils down to the data being collected having great power to harm people if it is ever put to malicious use, and a lifespan that exceeds that of institutions we know how to run. My beef is not with the surveillance alone, but with the combination of surveillance and permanent storage.
On the regard of data talking into the wrong hands, I take issue to this argument because it's not a unique problem to personal data collection. Any data could be hacked - bank information, address, whatever. But that doesn't mean we don't use the internet for banking and etc. It means we try to make systems that are difficult to hack. It seems like you'd want data collection not to happen on websites like Facebook and Google, when hacking isn't a unique problem to those websites.
P.S. I really like your posts and your tweets are hilarious, please don't ever stop.
Also I don't necessarily understand Ceglowski's stance on why we shouldn't use deep learning and should avoid surveillance on the web. I don't take issue with becoming a datapoint in Facebook's web of people because nothing bad has happened or can happen from me giving Facebook my data. When most people speak out about the data that's being collected about Facebook and Google users they say they're "worried about what could happen" but then never list any bad things that they're actually afraid of. The speaker falls to this issue too. Ceglowski says:
>I worry about legitimizing a culture of universal surveillance.
But then never explains what bad could happen from legitimizing that culture. Maybe I'm completely missing the point of the talk? Please explain what I'm missing if I'm actually missing something.