More on mass-surveillance > The Pentagon has a laser that can identify people from a distance—by their heartbeat https://t.co/8wFamZJvEE

If the advertising business model is what pushed services companies towards consumer profiling, mass tracking, and privacy violation through #AI, how is it possible that the same companies suddenly become privacy champions without changing that business model?

And speaking of #AI and mass surveillance, I have never seen this setting in Facebook mobile app control panel before. It was set to “yes” by default. I wonder for how long it has been there. https://t.co/br0Vzs2w4V

There are very few topics that I care about at a personal level more than the impact of #AI on society. That’s why a first edition of Nineteen Eighty-Four, my most beloved novel, is one of the most precious gifts I have ever received. https://t.co/mjtbNt43ez

Timely recommendation considering the new research on #AI just published: “AI uses camera footage to track pedestrians in dense crowds”
https://t.co/hLkYL3JSrH https://t.co/BaEGCJJjxp

Which are most parts. But we don’t talk about them. Cause it’s not code. And if it’s not code, it’s a shameful past we wish to forget. Yesterday it was fine, but today it’s not fine. It’s shameful. In fact, can we talk about something else? https://t.co/bSBP8UAwpO

2/ The only thing research/references on technology adoption tell you is what was the thought process that led to a choice. Not how effective the adopted technology is (which is what people ultimately want to know).

1/ Experience tells me that market research and customer references on technology adoption invariably leads to misleading conclusions and delusions of all sorts. Rather than “what we have done”, I’d be much more interested in “what we have done wrong”.

4/ The answer to every problem cannot be “you have to learn to code” or “let’s have devs be experts in every discipline, from agile to security to ML”. Knowing how to code doesn’t imply high quality output. And being an expert at everything implies being an expert at nothing.

3/ “collaboration is difficult because they need to explain their code to machine learning engineers so that it can be refactored it into something that is production ready…I think one thing severely lacking amongst data scientists is a technical maturity in programming skills”