It's very possible, likely even, that enormous areas of physics still have plenty of room for major discoveries and subsequent applications to new discoveries.

It's very possible, likely even, that enormous areas of physics still have plenty of room for major discoveries and subsequent applications to new discoveries. Furthermore, tying even increments of these major possible discoveries to the existing field of knowledge and application would by itself lead to a great deal of new exploration. But what if we haven't advanced much in several decades because we now suffer from or soon will suffer from the Dog Problem? Basically: no matter how hard you try, you couldn't teach a dog how to perform engine mechanics (or any number of other higher cognition things), simply because its cranial biology isn't up to the task of fundamentally understanding such things. I have a sneaking (layman's) fear that we might be close to our own version of the same thing in certain fields. It would be nice to argue that AI augmentation of human-led scientific advancement will help us along considerably, but it still seems deeply hand wavy to say that. Sure, computers help enormously with vast amounts of technical number crunching, and enable the practical application of physics (among other fields) in ways that we could never achieve without them, but fundamental, self-directed discovery is still way beyond the scope of any AI I know of or foresee for the near future.