I hadn't heard about Jev but when I read about what it does, I understood we are in an AI bubble.
I wrote a function pretty much exactly like Jev about a year ago... I would provide my function an array with a bunch of labels as first argument and a string of text/content as second argument and it would give me a relevance score for each label in the array against the provided text/content.
It's really easy to implement. Took me like a day or so by hand.
You just compute the embedding vector of each label in the array then compare each one against the embedding vector of the text/content and it then return an object which maps each label to its similarity score.
I used it to allow users to tag employee data from LinkedIn.
I had essentially forgotten about that code until now...
jongjong · · focus · HN ↗
I wrote a function pretty much exactly like Jev about a year ago... I would provide my function an array with a bunch of labels as first argument and a string of text/content as second argument and it would give me a relevance score for each label in the array against the provided text/content.
It's really easy to implement. Took me like a day or so by hand. You just compute the embedding vector of each label in the array then compare each one against the embedding vector of the text/content and it then return an object which maps each label to its similarity score.
I used it to allow users to tag employee data from LinkedIn. I had essentially forgotten about that code until now...