3 ms·Wu Dao 2.0: 10 times the size of GPT-35 points by krick 5y agocarrolldunham 5y agoStop making these glorified Markov-chain text generators bigger, and add some actual persistent memory.