Yeah but ChatGPT is using $700k per day in compute (or was in April). Someone's going to make an analog machine that uses mirrors and light interference or something to do self attention and become very, very wealthy.
How does the bit that determines if 2 articles are really about the same story work? How did you structure training data for that and get a training set?
Lockdowns or not, nobody could've avoided getting covid in the long term. It's just too contagious. The only way out is letting it run through the population and mutate. Flattening the curve might alleviate strain on the medical system but it also prolongs the time to benign mutations.