Totally agree but I think it's only part of it, other problems probably take par in this:
- spammers are getting better and Google is less able and willing to moderate
- the use of AI without much design around it. It's speculation since who knows how their algorithm works, but from what I've read and the general mindset in ML these days, it's very possible that they just use a recommendation AI with a single target (I've heard watch time, but again, who knows) with little to not design around it. This just does not work, especially if the AI is good at his job. It's a similar problem to decision makers blindly following KPIs, knowing if you did well and choosing criterias can be as hard as taking the decision itself, and an AI can't do that, you can't avoid designing your product.
On the up side, a lot of things are much cheaper or free (medical fees, internet, life in general except maybe the housing, depending where in the US and in France).
But yes, for your income bracket, you're likely to be worst off.
Keep looking, because salaries in tech are wildly variable, literally 25k to 100k+.
Keep in mind, high salaries are likely to be in Paris, but life their is expensive, especially housing.
And finally, it's your choice. Money is not everything in life, and since you'll be confortable either way, you can choose to gain less for an experience you want to live. Only you can answer if it's worth it or not ;)
Not sure it's true that we train that many engineers or that it's because of Napoleon, but most prestigious schools (in particular in Engineering) where created during his reign.
It's a term we often hear, that implies there is "good" and "bad" data.
A dataset can have errors in labeling, be very small, be unbalanced, but all that can be managed with the proper methods.
THE biggest problem is when you training data does not correspond to the production use-case.
It's not that the dataset is "bad", it's just that the problem you're solving with your ML algorithm trained on that data does not correspond to the problem you're trying to solve.
The most "perfect" ML algorithm trained on the most "perfect" dataset for self-driving cars for example (for detection, segmentation of objects or whatever) made the US will have problems when the cars drive in an other country. Your MNIST-trained NN will have problems in a country where numbers are written slightly differently. Some people will put pictures of cats in your car model classification software. Pictures taken on a smartphone by your users will be different than your dataset scrapped on the web.
There is no bad data, just badly used data. And most of the work (and the most interesting part IMO) in ML is to identify, quantify and neutralize biases in models and differences between the data you have and the data the production system will work with.
Data from the Copernicus program has always been fully available, served with a nice web UI, API for both near real time data and archives.
It's the best source of open satellite data by far.
As for the licensing, I never actually looked it up, so maybe you're right.