Sure! Deciding which model to use a posteriori based on the shape of the data is definitely not ideal.

Price probably falls off faster than inverse square. The available space grows quadratically, but as you get further you lose more prestige, proximity with important landmarks etc.

It's been heavily tuned to avoid generating racist or sexist content, climate change denial and avoid a number of sensitive topics.

The plane of orbit is still relevant though.

There are many ways to read this data. One would be that the schools had to make space for URM for possibly very good reasons, but did so mainly by kicking out Asian, to avoid hurting whites as much as possible.

Doesn't matter. If they want to train the algo to detect early signs, they need longitudinal data (image of retina long enough before potential diagnosis). Collecting such data will take time, by definition, so they won't be able to extend to wider population soon. In any case, I'd be very careful with claims of AI diagnoses. There are many many ways to hide bs in these project.

My first thought as well. I'm too lazy to do my homework and check how solid the scientific background may be, but my prior is very low.

Her father had little faith in the future of the Greek monarchy; he didn't trust that Otto man.