Using ChatGPT

Aleksandr

Well-Known Member
Messages
1,285
A few people have been asking about how I was using chatgpt the other day, so I thought i'd start a thread. We can put a collection of good prompts and interesting answers here.
CC: @ruprmurdoch @esp90 @Yura

Always keep in mind it can get things wrong. It is just another tool to help you, don't use it as gospel.

When you do your ARL or TEI test it should come with supplement recommendations. Upload your test to chatGPT and ask it for supplement advice, you can also tell it symptoms you have and it will try to work out what causes what. Always cross check its advice with the ARL or TEI supplement recommendations.
The free version is more likely to hallucinate and give you wrong answers, always check it has its inputs correct. For example, i uploaded a screenshot of my HTMA results and it said i had low magnesium of 12. It got the 12 correct, but 12 is high out of the ideal zone, not low, so it got that part wrong. I then corrected it, and it said that 12 is normal in TEI and that was its mistake, but in reality 12 is high in TEI as well.

Anyway, i kept correcting it, added some symptoms, and it ended up giving me pretty solid supplements (much less $$ than ARL), and in less intense doses so I think less could go wrong, but potentially it wouldn't move the needle as fast either. Keep in mind ARL has like 30 years of testing etc behind it, chatGPT is an amalgamation of various sources - don't use it unless you do your own research on its answers. Don't forget this thing has all the info of RDAs and general health recs as well. It also said things like i don't need copper or manganese as these are a bit high on mine, but that ARL have given them too me. Its probably a downfall of following ARL and TEI - they have pre-formulated supps, but ideally we'd be getting our own completely individual formulations each time.

You have to dance that tango with AI to get the best results. Keep iterating and cement one decision point before jumping to the next. I.e. ask what it should give you (only) and keep prompting it if it isn't correct. When it is, stop there - thats a decision point. Record the results. Then tell it what ARL recommended, and see what it responds. Another decision point. Then you could ask it what TEI would recommend. Another decision point. Treat these 'decision points' like a save point in a game. Then you could give it your symptoms, see what it responds. Another decision point. Then, if you want, you could say something like "I want to avoid vitamin A". It will change it. So good to know what it thought before vs what it thinks now - its easy to lose track of things in it, this is why i'm saying to have these decision points.

I prompted
> "decoding htma (hair test mineral analysis) how to take your own supplements based on your hair mineral report"
response:
EDIT: wasted too much time trying to copy paste into here, keep getting php errors on the browser console - seems the forum is too buggy.
It spat out general info on HTMA, what slows or fasts are generally recommended, important ratios to focus on and how to correct each individual ratio (without the context of an entire htma), then asked if i wanted to upload a PDF or JPG of my HTMA results for tailored supplements, and i did.
 
Last edited:

ruprmurdoch

Well-Known Member
Messages
451
IT is hard to believe that arl or tei leaked out their algos
Second thing they had different ways to heal the same cases
 

Yura

Well-Known Member
Messages
1,409
" Don't forget this thing has all the info of RDAs and general health recs as well. It also said things like i don't need copper or manganese as these are a bit high on mine, but that ARL have given them too me. Its probably a downfall of following ARL and TEI - they have pre-formulated supps, but ideally we'd be getting our own completely individual formulations each time."

Exactly they have blends of stuff in one pill to make it easier for people. So they don't have to take 20 pills. Most people have issues take 1 pill every single day not to mention opening and preparing 20 bottles..
I understand that they were aiming to cover the most people possible and for that you have to make some serious compromises.. Which is good in bigger picture as it can help more average people, because if it's easier, mor simple most likely more people will stick with the program, but for individuals like us who are "advanced" it is not good strategy and we simply have to bend it to our needs..
THat's why I never liked this mantra "you must take their supps and do exactly what they recommend otherwise you will have shitty results' that's bullshit..
 

RebelWithACause

Well-Known Member
Messages
2,613
I agree. I found you can do a lot yourself. But of course it has higher margin of error and the whole nice thing about TEI/ARL is that you outsource it and take their recommendations without having to think about it.
 

Enkounter

Member
Messages
28
Maybe a bit of a tangent from the point of this post, but there’s ways AI could be implemented within it.. has there ever been or does anyone think there’s viability in a sort of master charting doc that everyone can collaboratively add all their markers, results, ratios etc into and record their progress and to subsequently compare to others and analyse with ease? Maybe there’s certain patterns this approach could reveal which is getting lost due to the fragmented ways we log things currently.

I’m not too savvy with Google sheets or that sort of thing but I can see that being the sort of program to achieve something like this. With all that data in one place surely it’d then give AI much more to work with, albeit with maybe a more generalised outcome, but could be factored back down to apply to the individual with more well-read personalised feedback
 

Yura

Well-Known Member
Messages
1,409
Maybe a bit of a tangent from the point of this post, but there’s ways AI could be implemented within it.. has there ever been or does anyone think there’s viability in a sort of master charting doc that everyone can collaboratively add all their markers, results, ratios etc into and record their progress and to subsequently compare to others and analyse with ease? Maybe there’s certain patterns this approach could reveal which is getting lost due to the fragmented ways we log things currently.

I’m not too savvy with Google sheets or that sort of thing but I can see that being the sort of program to achieve something like this. With all that data in one place surely it’d then give AI much more to work with, albeit with maybe a more generalised outcome, but could be factored back down to apply to the individual with more well-read personalised feedback
But this is what ARL/TEI should already be doing. Working with AI to improve their approaches.. But I am not sure how many driven people are working there who are really trying hard to make it better or just bunch of people who have it as god pay somewhat easy job..
 

ruprmurdoch

Well-Known Member
Messages
451
But this is what ARL/TEI should already be doing. Working with AI to improve their approaches.. But I am not sure how many driven people are working there who are really trying hard to make it better or just bunch of people who have it as god pay somewhat easy job..
just tax AI, it would limit spread of it. it is all about tax. or make incentives on places where human work.
 

RebelWithACause

Well-Known Member
Messages
2,613
But this is what ARL/TEI should already be doing. Working with AI to improve their approaches.. But I am not sure how many driven people are working there who are really trying hard to make it better or just bunch of people who have it as god pay somewhat easy job..
Wouldn't be surprised if they are riding on Eck's back lol. Would be cool to see them go to the next level. Then again they are already using a big computer that is fast. Maybe AI isn't even needed.