Having no experience with word2vec, some reference performance numbers would be great. If I have one million PDF pages, how long is that going to take to encode? How long will it take to search? Is it CPU only or will I get a huge performance benefit if I have a GPU?
By and large I expect someone riding a bicycle to be fitter than the average bear. In contrast, someone using an escooter could have any possible composition: slim/obese, young/old, safe/reckless. The increasing popularity of scooters also means that the absolute number of people has likely risen from a decade ago.