Google's new neural-net LLM architecture separates memory components to control exploding costs of capacity and compute

Titans architecture complements attention layers with neural memory modules that select bits of information worth saving in the long term.