Tweets – Page 227 – Trochaisms – Letters make words; sentences make paragraphs

@pyry more a dogmatic difference IMO, where they’r…

@pyry more a dogmatic difference IMO, where they’re so determined to learn structure from data that they won’t bootstrap from lx priors

@MJB_SF @fordm I see that my working definition is that of “Kinsley gaffe”, not @MerriamWebster’s definition of gaffe.

@MJB_SF @fordm It’s the “tactically unwise” part of the definition that’s shocking: should be okay, right? 2/2

@MJB_SF @fordm I can’t believe I’m well-actuallying here but gaffes are definitionally saying something true. 1/2

Aaaaaaugh this is such a pithy explanation of the impedance mismatch between NLP and ML people https://t.co/C8tNe69wsN

@KennyGibbsPhD a #toddlervore obligatory eats toddlers, AFAICT

oh, lord. tomorrow is 9/11 and trump is running for president. i’m thinking a media blackout is in order.

@sarahkendzior is that a blue British police box in figure 1? Seems apropos.

@hangingnoodles @JenLucPiquant “neural nets are really good because they learn the same things human learn” is tautological