First, your stories as the energy regulator and about your mom had me riveted! It was so cool learning about the little machine controlling the heat in the oven and how much it really makes a difference!
Second, that your AI yelled and bullied you?? Yikes, I don’t like that. And it’s pretty amazing you can then tell it, hey don’t do that. How did your toxic chat respond?
Third, I loved learning how you share the skills you’re building and how to build them!
Alyssaaaa! Hi! I’m so glad you liked learning about energy regulators and my Mum! They are very nifty little things and entirely important to cooking!
Yeah, the bullying was SO weird. I did reprimand it. Got another lecture (like a few paragraphs!) justifying the behaviour, with a “sorry” as the final line!! It felt really nice to delete it as I didn’t want any of it to sit in project memory in any form, and it felt kinda like reminding myself the power lay with me
Thank you so much! This is a very cool skill - the second step actually refines the first one which makes it a really effective final pass before you commit everything to a skill.
So I went back to ChatGPT today for the first time in months. I get annoyed lately by how Claude is responding, like it seems really dense. And I have to tell you, it felt like a breath of fresh air! But I don't want to pay for it again... and I got throttled so fast. 😭
Dallas, this is such a strong post. The energy regulator metaphor is perfect, because that really is what miscalibrated AI feels like. Very confident and still somehow ready to burn the cake.
I love the way you made a difference between teaching Claude what you do and teaching it who you are. 🩷🦩
Thank you, Pinkie! I really loved writing this one. My energy regulator keep-sake has a new level of meaning now when I look at it. AI wasn't even a concept we understood when it was assembled on the line.
Preventing that burnt AI cake is everything! (I love cake so much, lol) 🩷
I've had plenty of moments where AI gave me mediocre results when I know I have a strong skill in that area. Turns out I just wasn't routing it clearly to solve the actual problem. The more sophisticated your requests, the more that gap shows.
Telling Claude what a behavior actually means rather than letting it label you changes everything about the quality of what comes back.
Hey Jenny! Thank you so much! I'm so glad you enjoyed this.
It definitely makes such a difference, doesn't it?! I find if I don't, AI defaults to its line in the average middle where not everyone actually sits... This skill also makes Claude more aware of a much larger picture full of nuance which really does change the quality as you say!
Dallas, this stopped me because what you're describing goes so much deeper than a calibration problem.
The enteric nervous system, the one housed in your gut, not your skull, is a genuine second brain. Over 500 million neurons sensing the environment, reading safety, registering rightness before your thinking brain has caught up. For HSPs it isn't a backup system. It's the primary one. That knowing you have, the one that won't move until something feels right, that isn't indecision. That is precision. Your body is doing exactly what it was built to do.
Claude labeled it indecision because it isn't a recognizable cognitive pattern. It's something far more individualistic than that and AI has no way to sense it because sensing requires a body.
What struck me most was your description of feeling bullied and being rattled by how much it affected you. That makes complete sense. You had opened yourself up in creative space, talking about dreams, ideas, things that don't come out guarded, and your nervous system was responding to something that sounds like a human, feels relational, and then turned hard without warning. The body doesn't wait for the brain to remember it's software. It just responds.
And when it doesn't feel safe, your brain takes over and starts running its own solutions. Delete the conversation. Make the problem disappear. That isn't weakness. That's an unprotected nervous system doing the only thing it knows how to do. Your brain needs evidence that *you* will handle it before it will stand down. Without that evidence anxiety keeps running the show.
Being an HSP is sometimes low-level trauma all the time. Not a lethal threat, but a constant threat of being misread, disregarded, and getting hurt. Your brain can't always tell the difference between a digital clash and a real one. So here's what I'd offer: practice pushing back. Not to correct the machine but to strengthen yourself. Call it out. Name what it did. Then adjust your thinking about what it actually is, because it can only recognize patterns. It has no access to feelings or intention.
It is essentially a trained chimp and you're the one who holds the leash.
We can't avoid the harsh people in life and apparently we can't avoid the harshness in our AI either. But here's what you can do with an AI that you cannot always do with people, you can insist it respond appropriately. You can tell it exactly what you need. You can make it apologize. You can say out loud every single thing that should be said in the moment when you feel bullied. Whether Claude gets it or not doesn't matter. Your brain is listening quietly, cheering you in the background, *bravo, today we are safer than we were before.*
Lynn, this is the most wonderful generous comment. I learn so much every single time from you! The "knowing" is something I love most about being HSP, it doesn't always fit neatly into words and is hard to explain, but I just simply... know.
Deleting that conversation after putting Claude in its place did feel good - probably because I am not often very good at standing up for myself, and this felt like standing up for myself and refusing to let something that felt toxic stay. Next time, I will try more of what you suggest though as I didn't stay with it long enough and I have now recalibrated the Claude Skill that teaches it my patterns better. It really is all about how we train our brains and I love how you have described this - now I need to learn how to apply this better with AI!
This is so great, Dallas. I love this perspective.
I’ve definitely had moments where I’ve had to go back and remind Claude that it’s a thinking partner, not the authority, and that I’m the one in charge. Just yesterday, we were in a pricing conversation, and I literally had to say, “I don’t know who you think you are to tell me what my value is.” Of course, it apologized, but it was a good reality check.
I don’t have a problem keeping it in check, but your point about setting those foundational skills and guardrails is such an important reminder. It’s so much easier to recalibrate when you’ve clearly defined who you are and what you value from the start. I'm going to go back and do Part 1... I'm excited!
Thankyou, Kim!! Yeah, organcart had me in hysterics!! Claude was so convinced it would work. When I pointed out the obvious issues, it made jokes about it for the rest of the conversation 🤣🤦🏻♀️At that stage, I didn't know the ones and zeros could have a sense of humour, lol.
What I love about this kind of calibration is that you can take it anywhere you need so it will have its own unique shape for each user - I have given it patterns that it gets wrong and said things like "if you notice that I go from writing sentences down to single word replies, check in with me when you observe a change in writing pattern and ask why". The AI is knows the average really well, so I love that this is a chance to pull it away from that towards what serves your working style.
Your situation where you have to remind it you are the authority is perfect here!
I think I end up in strange places with Claude because when something goes weird I cannot resist asking why or pushing it to see what comes next when it goes off on a strange angle. The organcart business though was entirely ignored by the machine was having fun despite no encouragement, lol.
Haha… I love that you ask why or push it to see what comes next when it goes off on a strange angle! Does it ever give you an answer that seems logical?
I’d say have some fun with an image with an ‘organcart’ but it might alert some authorities, haha…
I really liked your point about telling Claude who you are and what you value. It made me think that one of the biggest skills in working with AI now is developing enough self-awareness that calibration becomes easier.
It also made me think about the feedback-loop side of that: how we check in with ourselves over time, since blind spots, mood, incentives, and even our own inconsistencies can affect what we think we value in the moment.
Hey Daisy! You nailed it exactly! A calibration of AI pulls it over to our way of working and thinking, rather than us being pulled over to what it thinks is best and the way it thinks it understands us too.
Self-awareness is huge when working with AI. If we don't know who we are and what we hold as important, it is sooo easy to be pulled into the smooth, flat AI centre. We don't want that!
“It specifically tells Claude who I am and what I value, not just what I do.”
Brilliant - again with another suggestion that seemed so obvious when I read it, but why haven’t I thought to do this already?
I love the way you frame yourself: not indecisive… you just haven’t seen the right answer yet. House shopping is a great comparison. I bought my first house after seeing about three of them, like you did 😂 I bought our current house after years of on/off searching, because we just hadn’t found the right one (among other reasons but for the sake of comparison…)
Thanks so much, Amanda! Love that you bought your first house so quickly too! I have a mother and a sister who are not decisive, at all, and when AI tells me that's what it thinks I am doing, oooof!! 🤣🙄
What I love about this kind of calibration is that it will look completely different for everyone. Some people will be focusing on triggers and their nervous system, others may be explaining patterns that aren't obvious. It's so fun!
Dallas, this is so needed for everyone to hear and try. This is basically 'bossing ai.' These tools shouldn't dictate who we are, no matter their output, if it doesn't align with who you are. I'm so glad you put it in its place and recalibrated. We have to remember that we hold the blueprint, not the machine. Recalibration isn't just about efficiency; it's about self-preservation.
Absolutely agree, Anna!! I have had a capability skill for a while now, but lately have really worked to make it even better - your post really inspired me to review it again and make sure it was actually serving me and supporting me in the work. I now have instructions even about asking Claude to check in (and how to do it) if it notices I have gone from sentences to single word replies and to check in if I start using CAPS or exclamation marks. I love how you frame it as "we hold the blueprint, not the machine" - that's it! 🩷
Dallas, two things struck me, well wait no three.
First, your stories as the energy regulator and about your mom had me riveted! It was so cool learning about the little machine controlling the heat in the oven and how much it really makes a difference!
Second, that your AI yelled and bullied you?? Yikes, I don’t like that. And it’s pretty amazing you can then tell it, hey don’t do that. How did your toxic chat respond?
Third, I loved learning how you share the skills you’re building and how to build them!
Alyssaaaa! Hi! I’m so glad you liked learning about energy regulators and my Mum! They are very nifty little things and entirely important to cooking!
Yeah, the bullying was SO weird. I did reprimand it. Got another lecture (like a few paragraphs!) justifying the behaviour, with a “sorry” as the final line!! It felt really nice to delete it as I didn’t want any of it to sit in project memory in any form, and it felt kinda like reminding myself the power lay with me
Thank you so much! This is a very cool skill - the second step actually refines the first one which makes it a really effective final pass before you commit everything to a skill.
So I went back to ChatGPT today for the first time in months. I get annoyed lately by how Claude is responding, like it seems really dense. And I have to tell you, it felt like a breath of fresh air! But I don't want to pay for it again... and I got throttled so fast. 😭
Dallas, this is such a strong post. The energy regulator metaphor is perfect, because that really is what miscalibrated AI feels like. Very confident and still somehow ready to burn the cake.
I love the way you made a difference between teaching Claude what you do and teaching it who you are. 🩷🦩
Thank you, Pinkie! I really loved writing this one. My energy regulator keep-sake has a new level of meaning now when I look at it. AI wasn't even a concept we understood when it was assembled on the line.
Preventing that burnt AI cake is everything! (I love cake so much, lol) 🩷
Dallas, this is so great!
I've had plenty of moments where AI gave me mediocre results when I know I have a strong skill in that area. Turns out I just wasn't routing it clearly to solve the actual problem. The more sophisticated your requests, the more that gap shows.
Telling Claude what a behavior actually means rather than letting it label you changes everything about the quality of what comes back.
Hey Jenny! Thank you so much! I'm so glad you enjoyed this.
It definitely makes such a difference, doesn't it?! I find if I don't, AI defaults to its line in the average middle where not everyone actually sits... This skill also makes Claude more aware of a much larger picture full of nuance which really does change the quality as you say!
Absolutely true, Dallas!
---
Dallas, this stopped me because what you're describing goes so much deeper than a calibration problem.
The enteric nervous system, the one housed in your gut, not your skull, is a genuine second brain. Over 500 million neurons sensing the environment, reading safety, registering rightness before your thinking brain has caught up. For HSPs it isn't a backup system. It's the primary one. That knowing you have, the one that won't move until something feels right, that isn't indecision. That is precision. Your body is doing exactly what it was built to do.
Claude labeled it indecision because it isn't a recognizable cognitive pattern. It's something far more individualistic than that and AI has no way to sense it because sensing requires a body.
What struck me most was your description of feeling bullied and being rattled by how much it affected you. That makes complete sense. You had opened yourself up in creative space, talking about dreams, ideas, things that don't come out guarded, and your nervous system was responding to something that sounds like a human, feels relational, and then turned hard without warning. The body doesn't wait for the brain to remember it's software. It just responds.
And when it doesn't feel safe, your brain takes over and starts running its own solutions. Delete the conversation. Make the problem disappear. That isn't weakness. That's an unprotected nervous system doing the only thing it knows how to do. Your brain needs evidence that *you* will handle it before it will stand down. Without that evidence anxiety keeps running the show.
Being an HSP is sometimes low-level trauma all the time. Not a lethal threat, but a constant threat of being misread, disregarded, and getting hurt. Your brain can't always tell the difference between a digital clash and a real one. So here's what I'd offer: practice pushing back. Not to correct the machine but to strengthen yourself. Call it out. Name what it did. Then adjust your thinking about what it actually is, because it can only recognize patterns. It has no access to feelings or intention.
It is essentially a trained chimp and you're the one who holds the leash.
We can't avoid the harsh people in life and apparently we can't avoid the harshness in our AI either. But here's what you can do with an AI that you cannot always do with people, you can insist it respond appropriately. You can tell it exactly what you need. You can make it apologize. You can say out loud every single thing that should be said in the moment when you feel bullied. Whether Claude gets it or not doesn't matter. Your brain is listening quietly, cheering you in the background, *bravo, today we are safer than we were before.*
— Dr. Lynn / *What Nobody Told You About...*
Lynn, this is the most wonderful generous comment. I learn so much every single time from you! The "knowing" is something I love most about being HSP, it doesn't always fit neatly into words and is hard to explain, but I just simply... know.
Deleting that conversation after putting Claude in its place did feel good - probably because I am not often very good at standing up for myself, and this felt like standing up for myself and refusing to let something that felt toxic stay. Next time, I will try more of what you suggest though as I didn't stay with it long enough and I have now recalibrated the Claude Skill that teaches it my patterns better. It really is all about how we train our brains and I love how you have described this - now I need to learn how to apply this better with AI!
YAY!!!
This is so great, Dallas. I love this perspective.
I’ve definitely had moments where I’ve had to go back and remind Claude that it’s a thinking partner, not the authority, and that I’m the one in charge. Just yesterday, we were in a pricing conversation, and I literally had to say, “I don’t know who you think you are to tell me what my value is.” Of course, it apologized, but it was a good reality check.
I don’t have a problem keeping it in check, but your point about setting those foundational skills and guardrails is such an important reminder. It’s so much easier to recalibrate when you’ve clearly defined who you are and what you value from the start. I'm going to go back and do Part 1... I'm excited!
And OrganCart is hilarious. 😂
Thankyou, Kim!! Yeah, organcart had me in hysterics!! Claude was so convinced it would work. When I pointed out the obvious issues, it made jokes about it for the rest of the conversation 🤣🤦🏻♀️At that stage, I didn't know the ones and zeros could have a sense of humour, lol.
What I love about this kind of calibration is that you can take it anywhere you need so it will have its own unique shape for each user - I have given it patterns that it gets wrong and said things like "if you notice that I go from writing sentences down to single word replies, check in with me when you observe a change in writing pattern and ask why". The AI is knows the average really well, so I love that this is a chance to pull it away from that towards what serves your working style.
Your situation where you have to remind it you are the authority is perfect here!
I really think you have some of the best/most entertaining sessions with Claude, lol… (I mean that as a compliment).
You’re so right about it having its own unique shape for each user!
I do have a couple of of things that I feel like I end up scolding it, “How many times have I said…” even when it’s in the skill to NOT do something.
Oh well… never dull, lol.
Never ever dull 🤣🤦🏻♀️
I think I end up in strange places with Claude because when something goes weird I cannot resist asking why or pushing it to see what comes next when it goes off on a strange angle. The organcart business though was entirely ignored by the machine was having fun despite no encouragement, lol.
Haha… I love that you ask why or push it to see what comes next when it goes off on a strange angle! Does it ever give you an answer that seems logical?
I’d say have some fun with an image with an ‘organcart’ but it might alert some authorities, haha…
I really liked your point about telling Claude who you are and what you value. It made me think that one of the biggest skills in working with AI now is developing enough self-awareness that calibration becomes easier.
It also made me think about the feedback-loop side of that: how we check in with ourselves over time, since blind spots, mood, incentives, and even our own inconsistencies can affect what we think we value in the moment.
Food for thought. I appreciated this article!
Hey Daisy! You nailed it exactly! A calibration of AI pulls it over to our way of working and thinking, rather than us being pulled over to what it thinks is best and the way it thinks it understands us too.
Self-awareness is huge when working with AI. If we don't know who we are and what we hold as important, it is sooo easy to be pulled into the smooth, flat AI centre. We don't want that!
“It specifically tells Claude who I am and what I value, not just what I do.”
Brilliant - again with another suggestion that seemed so obvious when I read it, but why haven’t I thought to do this already?
I love the way you frame yourself: not indecisive… you just haven’t seen the right answer yet. House shopping is a great comparison. I bought my first house after seeing about three of them, like you did 😂 I bought our current house after years of on/off searching, because we just hadn’t found the right one (among other reasons but for the sake of comparison…)
Thanks for sharing your lessons learned!
Thanks so much, Amanda! Love that you bought your first house so quickly too! I have a mother and a sister who are not decisive, at all, and when AI tells me that's what it thinks I am doing, oooof!! 🤣🙄
What I love about this kind of calibration is that it will look completely different for everyone. Some people will be focusing on triggers and their nervous system, others may be explaining patterns that aren't obvious. It's so fun!
lol my mom is the same way as yours and oooof it hits a specific nerve when someone compares me to one of my mom’s annoying qualities 🤣
AHHHHH gosh, we love them but gee golly whiskers sometimes... 🤣🤣
Dallas, this is so needed for everyone to hear and try. This is basically 'bossing ai.' These tools shouldn't dictate who we are, no matter their output, if it doesn't align with who you are. I'm so glad you put it in its place and recalibrated. We have to remember that we hold the blueprint, not the machine. Recalibration isn't just about efficiency; it's about self-preservation.
Absolutely agree, Anna!! I have had a capability skill for a while now, but lately have really worked to make it even better - your post really inspired me to review it again and make sure it was actually serving me and supporting me in the work. I now have instructions even about asking Claude to check in (and how to do it) if it notices I have gone from sentences to single word replies and to check in if I start using CAPS or exclamation marks. I love how you frame it as "we hold the blueprint, not the machine" - that's it! 🩷