You're viewing a single thread.
Real science is trying random stuff until you get slightly better performance out of your model and then creating contrived explanations for why you think it worked
btw yann lecun is the head of meta ai so this is just a couple of rich dickheads having a slap fight
Lecun has massively contributed to the past decades of progress in machine learning with his fundamental research. He may have sold his soul to Meta, but his work is definitely very respected, he's not just another rich guy.
Sorry, but no. Science can be either trying out random stuff because there is no preexisting knowledge you could apply but you have something in mind you want to see what it does, sure. But science is also using the knowledge that's already out there to make something new of it. Translation of knowledge from different approaches into something that you are interested in.
Unless you went for a real scotsmen kind of joke and I missed it...
A lot of machine learning work is empirical so it was more a joke at the expense of oversimplifying the situation.