Useless Duck Company
3 HN recommenders · recommended in 4 months, 2018–2023
17:41 · 622K views · May 15, 2018
A maker builds a talking banana that reads YouTube and Twitch comments aloud on stream, then fights viewers who abuse it with profanity. He ends up building a pronunciation-based filter to catch bad words.
HN shares it as a funny, technical story about phonetic profanity filtering, relevant to Soundex and content moderation threads.
5 comments from Hacker News
You might enjoy this video. It’s part humor and part technical discussion on how his talking “Banana” bot was brigades by users on Twitch in an effort to make him say racist words and phrases. After he got a strike on Twitch, he patched the software and then it got circumvented... terrific little video about the battle between him and his viewers and ultimately ends with his approach to sterilizing input so that the banana couldn’t speak anything harmful. I thought it was great. * this video
You say that there's no "non-hacky" way, but... this video He made a filter that did analysis of the pronunciation of the messages that were being sent. His breakdown of it starts at around 13:30
If you are interested in phonetic algorithms, sorting, recognizing and filtering, you will enjoy Talking Banana Twitch ban story by Useless Duck Company : this video
Reminds me of the guy that streamed a talking banana on Twitch, where viewers could make it say things. People submitted variations of the n-word and got him banned, and after trying to filter out all character combinations he could think of he wrote a phonetic filter. That apparently worked much better than trying to think of every permutation of characters that sounds like bad words. this video
Related, here's a pretty impressive attempt to detect bad words, to allow a talking banana on a Twitch live stream, without being banned: this video