Hacker News Reader

Comment by spaceheretostay | Hacker News Reader

All this fancy stuff, but where's my Sets at? Google took down the website like 15 years ago. I used that thing daily. Then they put it into Google Sheets. You could type "apple" then "pear" then "orange" and then select cells, drag, and you'd get a list of fruits! Some you never heard of before!

I looked around and Google has a patent on it, and seemingly no public implementation. This makes me super sad because I honestly considered Sets one of the most useful exploration and ideation tools ever created.

BigQuery in Sheets, cool I guess? I want Sets back.

View on HN

Try this:

http://wordgrabbag.com/

Here is Google's Patent:

http://patft.uspto.gov/netacgi/nph-Parser?OS=pn/7,350,187&RS...

that's the same functionality, but the results are not as good.

I put in: fat obese porcine

I get: wild-type hyperactive amnesic misshapen narcissistic slimy cancerous motile putrid gravid

What's a use case for this?

I wonder how enforceable that patent is.

Still not as cool as Google Squared, where 10 years ago, you could create rows of any nouns and cols of any adjectives, and it would automagically research and fill out all the data for you according to its best guess. Create rows of car models, cols of price, seating, horsepower, safety rating, stuff like that.

That’s all in Google Knowledge Graph. They dropped it because it makes finding information too easy. And hence, competitor search engines could get info out of it super cheaply.

Boo! I’d love to have nice hooks into Knowledge Graph.

This reason is so much more disappointing than what I had previously assumed.

Seems unlikely to be the real reason

Seems convincing. Likely the same reasoning behind removing Real-Time results.

I don't know how Google's implementation of sets worked, but a simple implementation you could do yourself in a Jupiter notebook is to load up some pretrained word vectors, normalize all vectors, look up the vectors for all words in your seed set, take the mean of them, and then look for the nearest neighbours.

Stick it on a website and then the rest of the world can use it too!

Have a look at https://projector.tensorflow.org/

Reply on news.ycombinator.com