zlacker

>>cyanba+(OP)
If I remember correctly, the original version of wordle used a word list that was run past the creator's wife, who had learned English later in life. The result was a really accessible game - none of the words felt like ones you wouldn't know. It probably makes sense to reuse words than risk losing that accessibility.

(I kept a copy of original wordle, and it seems to have 2,315 words that are possible answers.)

>>trotha+lg
Yes, that's correct! Took her about a year off and on, he had made a little app for her to go through and categorize everything.

As an aside, for about $200, you can ask a true/false question of every word in the English language with a frontier LLM, and get mostly good answers. I make word games in my free time and was sort of shocked when I realized how cheap intelligence has been getting.

>>knuckl+xx
$200? Does this use reasoning? Does it involve forgetting to use KV caching?

This should cost well under $1. Process the prompt. Then, for each word, input that word and then the end of prompt token, get your one token of output (maybe two if your favorite model wants to start with a start-of-reply token), and that’s it.