You actually do not contradict the author that much. The thing is, the skills that are required are somewhat standard programming and graduate level maths, which narrows down the number of people to graduate level computer scientists. As he mentioned, the data science is NOT JUST application of someone's algorithms to the data, you need to have skills to preprocess the data, be able to use distributed systems for big data, actually know what is going on under the hood, etc. Also, I doubt that the netflix and kaggle competition winners were anyones from anywhere, they probably already had quite a bit of experience with ML.
I disagree. The effect of illumination temperature does take place, but in the article the stress is on color grading. It is that color grading that makes faces orange under both daylight (Transformers example) and tungsten light (iron man example).
Yes, you can describe tracking and Kalman filter using bayesian statistics. Although, my prof said that originally it was not developed that way, but it is easier to describe it using normal distributions as priors and likelihoods of the position of the object being tracked.
In the comments he says that the restrictions were not THE problem, THE problem was the instability. List of restrictions is kind of reminder of what he had to cope with, and wasted money on, until it turned out that GAE doesn't work properly.
The press release is at: http://www.ucl.ac.uk/news/news-articles/0617/290617-Smart-de...