If you give all the love in the world to a schizophrenic, will he stop hallucinating?

We're looking at cause here not 'cure' or treatment.

Regardless schizophrenic individuals probably would do better with being treated with kindness, respect, compassion, and love than with derision and rejection. At least observational that has been my experience.

If you put an LLM on better hardware, will it stop hallucinating?

Anthropomorphization of LLMs is bad, but the reverse is worse. This is not psychology.

Although... has anyone tried quantizing schizophrenia?

depends on if you consider message passing (neuro-transmitters) in a graph NN as a hardware or software.

I'm not sure the distinction is meaningful at a high level. That which can be implemented in hardware can be emulated in software.

It may be significant with respect to a specific implementation. Do you mean to imply that the current crop of LLMs hallucinate only due to some flaw in the hardware they run on?

There is no software at all, it is all hardware. These are mere abstractions because the reality offers some guarantees which allows one to transfer one system onto another as if the underlying system didn't matter. Those guarantees allows one to speak of commonalities between one stuff and another.

ref : The Universe is Hostile to Computers

https://www.youtube.com/watch?v=AaZ_RSt0KP8

In that vein, there is no hardware either. It too is just an abstraction. However, that is not particularly meaningful. We draw lines around those abstractions and are able to compartmentalize them when discussing them.

I think we agree. If hardware was perfect, we wouldn't need ECC in RAMs. The reason to do this is we are poking the walls of compartments here.

Once someone's lost touch with reality, maybe not. In the earliest stages? Maybe, yeah:

> McFarlane believes that psychosis can be prevented with a range of surprisingly low-tech interventions, almost all of which are designed to reduce stress in the family of the young person who is starting to show symptoms.

> McFarlane cites research done at UCLA suggesting that certain kinds of family dynamics — families that don't communicate well, or are overly critical — can make things worse for a young person at risk of schizophrenia.

https://www.npr.org/sections/health-shots/2014/10/20/3566400...

I can believe it. Not smoking is a great way to prevent certain types of lung cancer. However, if you have lung cancer, quitting smoking won't cure it.

Hmmm:

"Maybe it isn't entirely a hardware issue."

Maybe read what I said first before going off.

"And where does someone “learn” depression or schizophrenia anyway, especially if their parents are not acting the same way?"

Man, I fear your family...