Since I'm dealing with how good claude code is, I gave a research task to claude to help my in-laws (who are psychologists) understand the coming storm at a more visceral level.

AI produced this: https://nexivibe.com/future.mental.health.web/index.html

I can't argue with the urgency. I then had claude produce a prompt based on this that I could feed Grok as a project to do therapy, and the result was shockingly good.

Obviously, Dr. Oz is going to be a political lightning rod for some people, but if you factor that out and try using AI for therapy... this is coming. This is happening right now.

▲

hvb2 4 hours ago | parent | next [-]

Sure, the results CAN be good. The real question is if your health is one of those cases you want to roll the dice with. I don't, but that's up to everyone.

Tesla fsd was the same thing. You needed to have people to try it. Some paid for it with their lives. Will it make progress, sure. Do I want to be the one that dies to make that progress? No

▲

zmmmmm 3 hours ago | parent | next [-]

The real healthcare system is rolling a dice too. A different dice, but an expensive one with a lot of delay and inconvenience - it's not like there's a choice that's problem free.

▲

YZF 3 hours ago | parent | prev | next [-]

The unfortunate reality is that when you have a doctor you're also rolling the dice. I've had been misdiagnosed a few times in my life. These misdiagnosis were very disruptive but luckily not life threatening.

▲

bastawhiz 3 hours ago | parent [-]

And you trust an LLM trained on Reddit and Twitter not to misdiagnose you?

▲

YZF 3 hours ago | parent [-]

That's not what I said. There are probably plenty of situations (more standard) where doctors do a good job. The doctors also have the benefit of performing a physical exam, ordering various tests etc. so they have more data.

My point still stands though. They definitely mess things up. At the very least LLMs can be an additional tool that doesn't have the subjective bias, the cost issue, or lack of time/desire to investigate deeper or listen to the patient.

EDIT: LLMs would be better if they could access case data. Maybe that's something to consider (anonymized).

	▲	mathgladiator 2 hours ago \| parent [-]
		I just upload everything to Lord Grok and the answers I get are amazing. I then run it claude with Ralph and the results define explanation. I am humbled...

▲

mathgladiator 4 hours ago | parent | prev [-]

I have rolled the dice with it, and it's been great. I use it to tune fitness and every mental dimension. Time will tell.

I will tell you this, using AI is better than using HN and random people who don't know me. I had the best technical conversation of my entire life with AI. I feel like I can handle it, and maybe I will dive into AI psychosis at some point. who knows!

I can say that it is better than the silence I've had for years from no community giving two shits about what I wanted to build. I can admit the number of mistakes I've made, but at a certain point - enough is enough.

▲

jplusequalt 3 hours ago | parent [-]

>I can say that it is better than the silence I've had for years from no community giving two shits about what I wanted to build

It sounds like you were lacking community, then went deeper into isolation.

	▲	mathgladiator 2 hours ago \| parent [-]
		You are not wrong. I blew up a great career that burned me out to go on a journey that was absolutely crushing to emerge with a machine that allows me to have the deepest technical conversations of my entire career. Professionally isolated for sure. But, im married and have friends so its not that bad and dont worry. I just came to many observations about myself and AI has helped process so many emotions that I would never share with another human.

▲

hrimfaxi 4 hours ago | parent | prev | next [-]

It's unfortunate that there are not citations for some of the numbers. I found this piece compelling:

> Marcus. He is twenty-eight and did two tours in Iraq. He has nightmares four to five nights a week. He drinks to fall asleep. His girlfriend left because she couldn't live with someone who woke up swinging. He tried therapy three times. The first therapist was a civilian who asked how combat made him feel. The second used Cognitive Processing Therapy but sessions were every two weeks, and when she said “trauma narrative,” his chest closed. The third was private practice, $160 a session. Marcus started to describe what happened in the house in Mosul — the one with the family in the back room — and the therapist's face changed. A microexpression, less than a second. Marcus caught it. He stopped talking. He never went back. The 73% problem, made human. Marcus has things inside him that are killing him slowly, and he has never found a room safe enough to say them. Not because the therapists were bad. Because the therapists were human, and the things Marcus needs to say are the things that change how a human looks at you.

	▲	alexpotato 4 hours ago \| parent [-]
		This reminds me of a story about how veterans with PTSD from recent wars were reading the Odyssey b/c there is a point where Ajax describes a "flashback". More details here: https://www.vice.com/en/article/how-ancient-greek-tragedies-... P.S. Almost every time I hear discussions about veterans and PTSD, the veterans say something like "having a therapist be someone who is ALSO a veteran and has seen combat is worth a million points" (I'm sure there are therapists who are excellent and are not veterans, just pointing out what the veterans value).

▲

bahmboo 4 hours ago | parent | prev | next [-]

Just 9 parts plus foreword and appendix. It’s cool that it could generate this but it’s going to miss 99% of the audience. I intend to be constructive, not dismissive of your efforts.

	▲	mathgladiator 4 hours ago \| parent [-]
		Missing the audience is the story of my life. Thanks for being constructive.

▲

voxl 4 hours ago | parent | prev [-]

I think there is a coming storm as you say: the disinformation crisis on overload. Even what you present makes my eyes roll as a more competent person in the room. It's not an impressive website, it's a carbon copy of plenty of sites that exist.

It is full of text that is hard to decipher as being meaningful in any way, and that has unfortunately been true of LLMs for a long time, they are great at bullshitting.

Yet, to determine if the content makes any sense takes a substantial effort, an effort I am highly confident you did not do.

Still, I believe you're right, this will be enough. This will be "not everything you read on the Internet is true" x1000.

	▲	mathgladiator 4 hours ago \| parent [-]
		I am having it reviewed by a handful of psychologists, and they are having to grapple with it too. I'm confident that I reviewed it well enough to be concerned, and I'm even more confident that it is the future.