You want to be somewhat careful using ChatGPT or other LLM / AI tools as your “endmyopia therapist”.
These tools are impressive, certainly.
They get a lot right.
But they get plenty wrong, too.
Especially as you continue a conversation for longer, or go broader in your questions, the risk of getting hallucinations, plausible sounding (but wrong) advice, go up dramatically.
First there is the inherent LLM problem, more on that in a moment. There is also the issue that a lot of the Website content is up to 10+ years old and we’ve made thousands of changes / updates / revisions to original advice, over the years.
Some of what you will find are concepts and methods I no longer recommend.
You want to be aware that you’re in DIY mode and not all advice is current.
This is your eyes we’re talking about. The GPT doesn’t know or care what it’s saying. It’s designed to keep you engaged, to respond with plausible answers, to make you feel validated while appearing to be a knowledgeable expert. A big part of its programming is about keeping you talking to it, not to give meaningful advice.
It is a dangerous tool to trust when it comes to a serious bit of “bio hacking” like endmyopia type of myopia reversal.
The Website is not that easy to immediately get answers from.
It’s very tempting to just ask GPT. And it will very likely give you some good answers.
It could be 70% accurate. Or 90%.
That’s not enough. When it comes to your eyes, which are irreplaceable, the core design of this whole system is to build accountability first and foremost. You are made to dig, search, piece things together. You don’t do this if you don’t accept the learning curve.
This bit of India Jones research and hoops is what protects you from making mistakes by expecting easy answers.
You learn to question, search, troubleshoot on your own. You are made to take plenty of time and understanding before you make changes that will affect your eyesight. This is a very important part of endmyopia and doing DIY things with your eyes.
If you let an AI bot do the work for you, bad things are (far more) likely to happen.
I’m not saying this to dissuade you. In fact I spent months exploring AI tools and digging into advanced stuff like RAG databases and ways to make an AI be reliable enough to give to you as an assistant.
It’s not there yet. It’s tempting, it sounds good, it’s also a bit dangerous.
There are great uses for AI and it keeps improving rapidly. Right now what you got though is half parrot, half fortune teller. (don’t be mad at me ok, ChatGPT? I’m not throwing shade here, just warning real humans about real risks)
Chat Bots are very cool but it’s not what you want to delegate your brain to when it comes to reversing myopia.
As soon as the technology is reliable, I promise I will personally make these tools available here directly, trained on just endmyopia data. I’m not gatekeeping or trying to make your life hard for no reason.
Having done this for going on 20 years, with thousands of free resources, my goal has never changed.
Right now my recommendation would be to do this old school. By all means if you love it, use chat bots to brainstorm. ALWAYS double check its responses. They will be correct to a degree to make you think it’s all there. But as I said … ok we won’t beat the dead horse.
If you’re super serious and need support, buy BackTo20/20. Yes it costs money and yes it does so for a reason (not just for ole Jakey’s strip club affinity). You can buy accountability on my part, I will provide support. And when LLM gets good enough I will also offer that as an option, as economically efficient as possible.
Some percentage of people won’t listen. This has always been the case.
Which is why the “10 most common mistakes” continues to be a link that I share more than any other, no matter how much I tell people up front.
There is also our private forum, no ads, real humans, no feed. Get feedback there. It’s 5 bucks to join, no subscriptions.
Take your eyes seriously. Don’t expect a digital talking parrot to not make some sh*t up.
– Jake
