This is a test of generative Large Language Models on Scratch, using faster Tokenizers than ever before to keep speeds high, as the Scratch Engine itself is too slow to handle normal AI Calculation. In addition, Scratch is not getting GPU-accelerated, so we can't expect standard LLMs to run the same Speed here. After lots of Optimisation, i present to you: April Fools lol
