>To reach a more brain-like behavior LLMs need to integrate your inputs into their model dynamically, essentially retraining real-time based on the most salient input.

That's already possible with LLMs. The challenge is that 1. it would allow permanently jail-breaking models and 2. there'd be no way for them to efficiently transfer what they'd learned to a new model generation.

Oh do you have a source? I haven't seen it done in real-time.

Coincidentally the human brain is also jailbroken and nontransferable