DeepMind claims its AI called RETRO matches the performance of neural networks 25 times its size, cutting the time and cost of training large language

[Techmeme]

Will Douglas Heaven / MIT Technology Review:
DeepMind claims its AI called RETRO matches the performance of neural networks 25 times its size, cutting the time and cost of training large language models — RETRO uses an external memory to look up passages of text on the fly, avoiding some of the costs of training a vast neural network