A user welcomes the perceived change, recalling questions about whether someone 'respects language.'
A user says the attitude of asking whether someone is a 'real NLP person' who 'respects language' seems to have disappeared at NLP conferences. They welcome its apparent disappearance: 'Good riddance.'
15.7K
13 Sources, first seen ago
One poster wrote that a familiar question at NLP conferences, whether someone was a “real NLP person” who “respects language,” seems to have disappeared. His verdict on the perceived change was brief: “Good riddance.”
A respondent argued that the larger shift was in machine learning’s status hierarchy. In his recollection, 12 to 15 years ago, core machine learning occupied its own position and computer vision was considered more serious than natural-language processing; now, he wrote, “everything is NLP”. The original poster remembered the communities differently, saying NLP seemed more elitist internally than vision, though he acknowledged that his perceptions might not have been representative.
Another respondent suggested that the gatekeeping may have shifted rather than disappeared. He said the old test had partially become whether someone is a “real frontier person” who runs large-scale experiments, a standard he called equally bad.
The thread also raised a separate concern about language quality. One commenter argued that LLMs produce unnatural language, then on product teams rather than researchers while calling for influential people who care enough about language to improve it.
—
Not ranked yet
—
Not ranked yet
@andrewgwilsOne attitude that seems to have disappeared at NLP conferences: “are you a *real NLP person* who *respects language*”? Good riddance.
@_onionesque@andrewgwils The real change is something else though. 12-15 years ago there was this hierarchy in ML. Core ML was at one place. Vision was always considered more serious, but NLP wasn't really. In a fitting revenge, now everything is NLP.
@natolambert@andrewgwils im a fake nlp person@andrewgwilsOne attitude that seems to have disappeared at NLP conferences: “are you a *real NLP person* who *respects language*”? Good riddance.@_onionesque@andrewgwils The real change is something else though. 12-15 years ago there was this hierarchy in ML. Core ML was at one place. Vision was always considered more serious, but NLP wasn't really. In a fitting revenge, now everything is NLP.@natolambert@andrewgwils im a fake nlp person@andrewgwilsOne attitude that seems to have disappeared at NLP conferences: “are you a *real NLP person* who *respects language*”? Good riddance.@_onionesque@andrewgwils The real change is something else though. 12-15 years ago there was this hierarchy in ML. Core ML was at one place. Vision was always considered more serious, but NLP wasn't really. In a fitting revenge, now everything is NLP.@natolambert@andrewgwils im a fake nlp person@andrewgwilsOne attitude that seems to have disappeared at NLP conferences: “are you a *real NLP person* who *respects language*”? Good riddance.@_onionesque@andrewgwils The real change is something else though. 12-15 years ago there was this hierarchy in ML. Core ML was at one place. Vision was always considered more serious, but NLP wasn't really. In a fitting revenge, now everything is NLP.@natolambert@andrewgwils im a fake nlp person
