Train a Markov chain on your tweets so you don't have to think of tweets anymore

Gemfile

# frozen_string_literal: true
source 'https://rubygems.org'
gem 'marky_markov', '~> 0.3.5'
gem 'twitter', '~> 6.1.0'

speak.rb

require 'bundler'
Bundler.require

client = Twitter::REST::Client.new do |config|
  config.consumer_key = '...'
  config.consumer_secret = '...'
  config.access_token = '...'
  config.access_token_secret = '...'
end

tweets = client.user_timeline(client.user.id, count: 200, trim_user: true, exclude_replies: true, contributor_details: false, include_rts: false)

markov = MarkyMarkov::Dictionary.new('twitter_dictionary')
tweets.each {|t| markov.parse_string t.full_text }
markov.save_dictionary!

puts markov.generate_n_words 200