version 1: https://scratch.mit.edu/projects/1242382682/ triple click the output list, copy, and then paste somewhere to see the full ouput. Press space to retrain (may take a while tho) This one is a second order markov chain instead of first order. basically, it takes in the context of the past 2 words instead of 1 words when deciding what to make the next word. training data from chatgpt. 280 blocks or something
