[CG] Meta’s Multi-token Model, A New Beginning for AI?

Just after I wrote that I was going to limit the number of notes that I would forward, I was gifted with the following note. I believe that it is an important innovation that can make LLMs significantly better. But I still believe that some hybrid method for evaluating results is stll a requirement. For example, I don't believe that it can detect and eliminate all the bad stuff on social media. John -------------------------------------------------------------------------------- Meta’s Multi-token Model, A New Beginning for AI? A New Type of Faster… and Smarter LLMs? https://medium.com/@ignacio.de.gregorio.noblejas/metas-multi-token-model-a-… Meta offers this with new research: a model that predicts multiple tokens at once in every prediction, not just one, and unlike previous proposals, with no training overhead. This not only speeds up the model's text generation but could also make it smarter, meaning we might be about to enter a new training paradigm for frontier AI.