There's a third too: Moore's law. Few researchers in the world had access to the compute resources to do this stuff in the 1980s and 1990s. Today a high-end desktop can train models large enough to do compelling things in a few hours and a phone or a small laptop can execute a model to render a prediction or do pattern classification in seconds.

For the really huge models we now have commodity cloud compute available when in the past you would have had to be among the elect with access to a supercomputer. You couldn't just go whip out a credit card and rent time on a Cray II or a Connection Machine and commodity hardware back then would have taken years to train something like GPT-3 if it even had enough storage.