A: This Tweet is making the rounds: "Nearly every ambitious person I know who has dived into Al is working harder than ever, and longer hours than ever. Fascinating dynamic tbh. I have NEVER worked this hard, nor had this much fun with work." I'm in this Tweet.
C: Suggests a very, ah, particular attitude to pre LLM SE imho. You're not the first people to invent excitement or better mechanical advantage in SE. The Al crowd are culturally rooted in old languages, building not proving etc. Fine. But different interests would mean different tools excite
C: Many people who aren't your idea of "ambitious and in Al" find the idea of a stochastic slave mind the opposite of their ideal. No explicability and a surrender of the territory of formal nolalon. Al folk claim that all such folk are now obsolete due to unproductivity. Maybe.
V: You gotta give in to your times. A surrender of the territory of formal notation? Oh, yeah, because we were totally writing coq proofs for every CRUD API we wrote at our shitty jobs before the model came. Every 1/0 operation was optimal before the model. C'mon. The model can output formal notation.
V: My point being: you're talking about orientation. The model is orientation-agnostic. If you gave a shit about formally proving every line of code, you would use the model to build a tool that actually made that feasible.
C: The model & tooling is very clearly not orientation agnostic. You're getting the model to do a Coq proof of every shitty CRUD api you trash the climate to spit out? C'mon. Also: CRUD is a really great example. Fetishized, dumb, cried out to be simply automated away through higher level concepts.
C: Also: your shitty job not mine. My point is the insinuation here is that CS/SE pre LLM was some awful, stuck in the mud thing that no one could get excited about enough to work hard. Not true.
C: And re formal notation: the point of notation is not to output it… it is a tool for thought. The point about LLM generated stuff is that you ideally don't. Saying the LLM could output notn is missing the point (in a way that neatly shows mine about ditterent priorities)
V: Okay. So what you're saying is: none of these problems would exist if only everyone did their job in the very special smart way that I have always done mine. You wouldn't need the Ilm, and also you'd have the same productivity gains that the Ilm claims to give us. So why didn't that happen?
C: Nope missed again. I'm saying that LLMs excite some people that care about some outcomes, namely shipping shit. I'm not even saying that's a bad thing to care about. I'm saying that if you cared, say, about denotational semantics, you'd be totally nonplussed about the ability to autogen CRUD
C: And I'm also saying that if your balance of concerns were more heavily on explicability or PLT (pre or post code) or expressiveness, you'd also be less enthused about the genAl approach, because the whole idea is to get the human away from the code even at the cost of an erosion of insight.
C: And I'm saying OP basically suggests that SE/CS were just waiting to be liberated by Al and are only now staying up all night with intellectual excitement! And that is BS that derides both the passion of other people and the beauty of their ideas.
C: But again: different value systems. Oh and I am pointing out that there has been pretty deep social assortment of these people and the Al community, incredibly bright though they are, have surprise surprise not much cared for this stuff: why trouble themselves w/ it when they will force Als to!
C: And the indifference verging sometimes on contempt is of course often mutual. Someone who really wants to solve full abstraction isn't that excited by stealing a brain and putting it into silicon.
C: Of course only one side of this dichotomy routinely paints the other as dinosaurs and relishes the idea of putting them and pretty well every cultural or desk worker out of a job so the moral equivalency does rather strain after a while.
