I also think this. But more in the sense where both end of the LLM are trained using words through repeating arithmetics, considering LLM itself is a repeating pattern of connections, the space in the middle if extracted the same way as the beginning and the end would become data that make sense to us.
Can the awareness simply come from injecting knowledge of itself during fine tuning and then during a chat a system prompt is injected to add a particular context that triggers its self-knowledge?
Example: Steam cloud save. Countless time I have to force quit Steam's background service because Cyberpunk 2077 and GTA V freeze while saving.
Steam doesn't show the launch game button, even though the game is killed from task manager.
There are no heartbeat like mechanism to keep in check the game process state.
And there are no alternative to launch the game that I bought from steam, so the instability becomes little of nuisance.
But that's a minor issue where unstable features are unimportant ones.
> It is all about how often stability is not a priority when it could and or should be.
And I'm on your team too. That's why my argument is that stability IS A feature too. Countering grant parent's argument that says stability is ignorable as long as the features are validated by the market.
> And in the no alternative scenario, it can be compelling to continue to ignore robustness to get growth and lock in.
As a paranoid user I agree this has been a nightmare. I constantly have to make sure I wont get locked in to services that I pay for, and often time I find myself using open source solutions rather than paid one because of the same reason.