Using AI For General Life Help

[pursuitofspirit]

Recently I have been experimenting using ChatGPT for general life advice, asking questions about conflicts at my work, how to best handle those situations, questions about not taking antipsychotic medication, etc.

I'm curious what the forum thinks: Is it a bad idea to lean on ChatGPT for advice such as this? 

For me it sort of feels like Jarvis in the Iron Man series - a sidekick that can objectively view situations and give you feedback without emotion or distortion - but I also feel like this may be limiting my own independent thought process at arriving at my own conclusions.

What do you all think? Would love to hear your thoughts.

[Natasha Tori Maru]

Keep in mind that it is answering based on the opinions, data, and language of the internet. 

In terms of medical advices, I would always refer to a trained professional. You never know if the language bot is giving you advice from some random forum posts from 10 years ago in an effort to maintain your engagement. And this leads to the next issue; ChatGPT is sycophantic, and modelled to keep your attention, awareness and engagement. You attention is your most priceless asset. 

Other than that, I have had many people close to me use it for therapeutic advice and claim excellent results. The one thing I will raise is that, without having thought process awareness (ie meta-thought) the bot will not be able to tell you where your cognitive bias is. Usually you need a trained therapist to be able to dig into your thinking patterns and challenge you to break them; ChatGPT won't challenge you in this way.

It has its pros and cons   

[Joshe]

It's a great idea. It's the #1 self-development tool in existence if used right. The main thing is to get clear on everything it's saying and evaluate it all for yourself. Also, ChatGPT is often a bit what you might call "woke" - trying to be careful with it's responses. I found Claude is usually better, but I combine them to seek out multiple perspectives. 

Many "intelligent" people are threatened by it or they don't understand or haven't discovered it's value, so they'll often drastically underestimate its value and tell you to not rely on it much, but that's foolish. 

I can spend 4 hours with an AI going back and forth on a topic. At the end of that 4 hours, I walk away with more knowledge and development in that 4 hours than what I'd get from a month of inquiring with books, blogs, forums, etc. 

AI has been huge for my development.

[Joshe]

14 minutes ago, Natasha Tori Maru said:

ChatGPT won't challenge you in this way.

It will if you ask it. All you gotta do is ask. And Claude certainly will. 

But yeah, if the user lacks meta awareness, it'll likely just amplify their blind spots. I catch it trying to appease me often and have to redirect it. 

Edited 3 hours ago by Joshe

[Cred]

 

[Natasha Tori Maru]

4 minutes ago, Joshe said:

It will if you ask it. All you gotta do is ask. And Claude certainly will. 

Yeah Claude does! The user said ChatGPT - so I assumed just that LLM was being engaged with. ChatGPT always tries to revert to sycophantic shit though - even with my programming.

6 minutes ago, Joshe said:

But yeah, if the user lacks meta awareness, it'll likely just amplify their blind spots. I catch it doing that often and have to redirect it. 

Meta-awareness is a skill very difficult to learn. Massive asset. I didn't become aware of this skill until my 30s! 

The blindspot amplification is REAL

[Joshe]

8 minutes ago, Cred said:

 

This is an example of an intelligent person telling you how bad it is. 

Those most dismissive of AI probably haven't actually sat down with it to work on serious problems for multiple 4 hour sessions. This guy is probably too busy with his business to do that. 

And he finds studies about how 8th grade students suffer cognitive decline from lazy AI use, then extrapolates that to adults. If that isn't lazy thinking, IDK what is. Had he consulted AI before he made this a pillar of his position, he could have learned something. 

If you're lazy and only want quick answers, then yeah, it's bad. If you care about getting things right and put in the energy, it's amazing.

I could use AI to teach myself anything in a mere fraction of the time it would take to learn in any other way. And this also applies to self-knowledge. Because of AI, I found out why I liked to throw 2 marbles in the bathtub and watch them interact when I was a kid. lol. This revealed things about my nature that have been with me since the start. I spent many hours with it untangling my childhood fascinations and walked away with a ton of self-knowledge. 

[Joshe]

Also, you can bypass it's bias with different tactics, for example: 

"So this guy is going through X and feels like Y. I'm pretty sure he's missing something. What do you think?" 

X and Y are your situation. 

This biases it in the opposite direction. 

It's a tool and you have to know how to use it.

Also: 

"Here's my situation. First, tell me what I might be missing or getting wrong. Then tell me where my instincts might actually be sound. Then tell me what questions I should be asking that I'm not."

Edited 2 hours ago by Joshe