holy shit lol; how do you even stumble upon shitposting of this magnitude. the best kind of shitpost is where the author does not recognize how much of a shitpost it is (i assume that the poor lesswrong fella thinks it's a collection of accurate analogies)

this post has convinced me to go to next year's SPP. thank you mr. bloom.

I wish the would-be wielders of Occam's razor returned to their dens unblooded more often

Honestly, to an extent that is a good thing. At least chatting to relatives is a bit easier nowadays if you work on language in any way

esp. re: recurrent appearance of the "it's all word games" view of academic theoretical linguistics. frankly, insofar as the funding bodies don't think so, 'tis all good: but who is to say that the wind is blowing in a good direction

at this point the field needs a rebranding of that argument. call it "dearth of sentences" or something. PoS (argument) attracts PoSes (people)

let alone that the poster seems to repeat the old "chomskyans are scared of statistics" bullcrap. i thought that enough nuanced opinion pieces re: LLMs and lingusitics appeared since that Piantadosi master ragebait; but oh well

i'm actually confused to what extent is this a criticism of formal grammar as an approach to science of linguistics vs. a criticism of formal grammar as an approach to NLP-related problems (afaik Bender mostly did the latter). both criticisms are wrong tho but for different reasons

the fact that there is some other (extremely stimulus rich!) route to similar performance is just a non sequitur

In a way, everything written about the "linguistic wars" period of generative grammar (Newmeyer from the "whig" side; Harris from the funny book side).

What I like about this as a case study that the failure of generative semantics was both substantive and purely sociological, somehow

hear, hear!

19. Niche linguistics content here, but: DO support does not arise due to anything like a Stray Affix Filter, and I will die on this hill.