Awesome! It was a pleasure to read your very compact source code, BTW. edit: Can we rename this bot "The Feels Good Man"? ;)
nevermind. here it is: https://gist.github.com/abeppu/6958565
41–50 of 60 posts
Awesome! It was a pleasure to read your very compact source code, BTW. edit: Can we rename this bot "The Feels Good Man"? ;)
nevermind. here it is: https://gist.github.com/abeppu/6958565
Late this afternoon, I started @FeelBetterBot back up (I had kept it 'off' for most of the day, while lots of people were looking at it) ... and it lasted less than 3 hours before getting shut down again.
A while ago I made a Twitter bot ( https://twitter.com/HereHaveAKitty ) with a friend. We used the face.com API to detect sad faces in photos in tweets (as well as ASCII emoticons) and reply with a photo of a kitten. It lasted about as long as yours before being suspended. We tried again ( https://twitter.com/KittenDelivery ) but that too was shut down.
[1] http://www.kernelmag.com/features/report/6451/what-suicide-n...
Awesome project! Any chance you'd be interesting in publishing the code for the bot so we could make variations of happybots to spread the <3 and not-so-subtly protest the suspension of said happybots?
The code is here : https://gist.github.com/abeppu/6958565 (also linked from the article text.) I'm fine with other people making lots of happybots. I just hope no one would adapt the code to, say, harass or bully people who are already feeling shitty.
Also has to send some non-@replies, so it'd work better as a dual-purpose bot that does some other work too, I guess.
Earlier quoted context omitted.
How the hell does he get so many upvotes? He has hundreds of comments with nothing but run on sentences and they all have dozens of upvotes each.
Because the intersection of utter surreality and being really really nice to people is ... to me, both awesome and hilarious. I would suspect but have no idea how to test that the upvotes are coming from fans of it as performance art who're following the user, rather than particularly from people already interacting in the original thread.
I do want to say that it's one of the best reddit user comment histories I've ever gone through.
As someone who loves Twitter gimmick accounts, I would love to hear some strategies for keeping them in the clear. Cute project!
If you have only 20 different tweets, its easy for them to look like a typical bot to automated filters. Even if you are posting them slowly, it wouldn't be a stretch to assume one flag would be a ratio of similar tweets.
The response criteria could be another thing raises a flag. It wouldn't be difficult for a filter to recognize someone only replying to specific key phrases.
I just had an idea for Twitter though. Twitter should make it so that bots get banned after a time proportional to the number of followers they have. Essentially this says: if people like what the bot is producing then it should need less manual review. Of course, you'd need to make sure that those are real people who follow it, but I dont think that that should be too difficult.
If changing the name works (i.e. create a new account), then you have something else to write about: Twitter doesn't care if you spam people with tweets, as long as you are concealing the fact that its spam!