I think this picture misses a major point point. If the AI becomes better at research than us (especially theoretical mathematicians), then what’s our purpose? Research will become even more of a funding challenge because everyone will need tokens to produce anything
A professor texts his PhD student:
“Good news! Our paper got accepted to a conference in Hawaii. We’re flying next week!”
The PhD student texts his girlfriend:
“I’ll be away for a week. Conference.”
His girlfriend texts her advisor:
“I’ll finally have time to finish the experiments next week. No one will be distracting me.”
Her advisor emails the dean:
“The results are almost ready. We can submit that major grant proposal.”
The dean calls the professor:
“We need you for an urgent proposal, you can’t go to Hawaii. We need you here.”
The professor texts his PhD student:
“Trip canceled.”
The PhD student texts his girlfriend:
“I’m staying.”
His girlfriend texts her advisor:
“The experiments may take a little longer.”
Her advisor emails the dean:
“It’s too early to submit the grant.”
The dean calls the professor:
“The proposal is no longer urgent. You can go to the conference.”
The professor texts his PhD student:
“We’re going after all!”