Guess the word
Each guess must be a word (of any length) or short phrase. The similarity value is how similar Word2vec thinks your word is to the secret word. The highest possible similarity is 100 (indicating that the words are identical and you have won).
For today's secret word, the nearest word has a similarity of 82.08, the tenth-nearest has a similarity of 58.13 and the one thousandth nearest word has a similarity of 22.47.
The "Getting close" indicator tells you how close you are --if your word is one of the 1,000 nearest normal words to the target word, the rank will be given (1000 is the target word itself). If your word is not one of the nearest 1000, you're "cold". (By "normal" words", I mean non-capitalized words that appears in a very large English word list; there are lots of capitalized, misspelled, or obscure words that might be close but that won't get a ranking).
You will need more than six guesses. You will probably need dozens of guesses. There's a new word every day, where a day starts at midnight UTC.
Secret words may be any part of speech, but will always be single words. It's tempting to think only of nouns, since that is how normal semantic word-guessing games work. Don't get caught in the trap! Since our Word2vec data set contains some proper nouns, guesses are case-sensitive. But I removed all but lower-case words from the secret word set, and if your word matches the secret word but for case, you win anyway. So if you want to know if the word is more like nice or Nice, you can ask about both.
Prior art. I discovered French Toast about 20 years ago, and immediately knew that I had to tell Dominus about it. He replied that in fact he and Ranjit had invented it ten or so years before that, and called it "Plenty Questions".
But I wasn't thinking of French Toast when I invented this. Instead, I was thinking about Wordle and Worldle and about the dimensionality of data sets. I considered Filmle, a movie guessing game: You say "The Matrix", it says "Keanu Reeves has a bacon number of 3 where bacon is the lead actor in the target film." (and so on for Carrie-Ann Moss, Laurence Fishburne, and Hugo Weaving for actors at corresponding billings). But I don't care about movies. Too bad, because IMDB actually has exactly the data that you would need to implement it in a really easy form. Then I remembered word2vec, and here we are.