Tweets – Page 612 – Trochaisms – Letters make words; sentences make paragraphs

RT @nsaphra: As a #NLP researcher, how can one cho…

As a #NLP researcher, how can one choose a research topic to ensure human obsolescence hits people with wealth and power, not just truckers?

From an ACL 2016 paper. Is 0.01% improvement on CoNLL 2003 data considered statistically significant these days? https://t.co/VuOkUF7DVx

@gfish that’d be one supervillain shark-tank there: superpressurized, superhot, “black smoker” ventilator, light-isolation.

BETEO, combined with stubborn refusal to be more specific, is the “Tyranny of Structurelessness” of the 21c left-bro https://t.co/VAWRKGVW9d

The best “tell” for identifying bad data scientists is claims of mastery of the whole discipline. That is functionally impossible.

@FakeUnicode @Pinboard I’m holding out for Thnad Combinator funding.

also, what a design tire-fire. I can barely read their meaningless text. https://t.co/wDpeLn2la4

Attention @FakeUnicode, have we ever had a Series Yuzz or Humpf? https://t.co/5no3ARH9dF