I get the instinct to mock “ChatGPT-5 will be PhD smart” w self-deprecating jokes abt how ppl w PhDs aren’t actually smart.
But given the total contempt that AI Grifters hold for all forms of expertise, think the proper response is “fuck you, your parlor trick will have nothing like our expertise.”
I get he instinct, I do! I think most PhDs are instinctively self-deprecating, bc we live in a society that claims to be a meritocracy but is defensively socially derisive against those w actual academic expertise. (“Egghead,” “nerd,” “not a real doctor.”)
Don’t play into that, esp not now.
So much of both the AI push and the broader [waves hands at all the disinformation out there] relies on aggressively devaluating actual expertise.
The actual experts need to resist that game, not jokingly play along with it.
I mean, the point of a PhD (or any degree, or ideally any level of education ever) is to teach you how to see something NEW and UNUSUAL in what you’ve read.
That’s literally the opposite of the pattern-matching algo that is AI. It’s why almost none can even get simple math q’s right.
Just … no.
slight nitpick, not the point you were - but on simple math qs. not perfect obviously, but the closed source models do pretty consistently on a very wide range of math problems. arxiv.org/pdf/2405.12209
Yeah, I’d have been thrilled if AI could have answered my research questions for me, but it wasn’t just information or being “smart” that enabled me to do it myself.
