Hacker News Reader

word2vec does not require a large amount of data

mnist might have required a large amount of data at its creation, but it became a staple dataset

There was a lot of evolution before ChatGPT

And people in academia were all over Word2Vec. Mikolov presented his work in our research group around 2014, and people were very excited. Granted, that was _after_ Word2Vec had been published, and this was a very pro-vectorspaces (although of a different type) crowd.

Reply on news.ycombinator.com