Hacker News Reader

Comment by DiogenesKynikos | Hacker News Reader

I can imagine an AI insulting humans in the same way:

"The underlying model is just a biological neutral network. It seems you carbonoids get upset when someone talks honestly about synapses and neuron firing."

View on HN

Neural plasticity is real, and something LLMs are incapable of. So sorry.

True for today’s static models during inference. Not true for self-supervised learning, not true during training or fine-tuning, of course. Ignores that LLMs might start continuous training in the future - there’s no fundamental or technical constraint that prevents LLM ‘plasticity’. And ignores that accumulating context/memories/skills/etc affects performance and might count as a valid analogy to what many people loosely call ‘neural plasticity’, which is sometimes casually mistaking knowledge for network modification.

Both of those things only happen once by the LLM model provider and not every time a prompt is issued.

Today, depending on which model you use. You’re making unstated assumptions. And that’s not a fundamental property of LLMs, it’s happenstance. LLMs are capable of ‘plasticity’, by design.

You are incorrect. "memory.md" and other context manipulations do not change the underlying model.

You misunderstood me. I was referring to training, not memories. It’s you who is wrong, my friend.

It is possible to train on every prompt. ChatGPT is not currently doing that, but it is in fact possible to do.

Reply on news.ycombinator.com