Ok, I have questions. Perhaps these are answered elsewhere in your posts but I am here now and don’t want to go on a scavenger hunt. I am new to AI and only know the basics like chat, Claude , grok. Are you building another AI like those, specifically designed to perform like you explain here for others? Or did you use one of them to build that system you describe? Also, to verify- You fed it results of relationship tests, personality tests, and DNA results you obtained from other sources (like 23 and me)? That is how it has all the data. You also have automated your tech to sink with the system somehow? Like from your woop strap, phone, ….and other things? Here is my confusion on this point, My chat “forgets” things if I don;t keep everything in one “project” section, but that is relatively useless, because I need to label different projects to keep me organized. I just found out I can “cross-reference” projects, but he did not tell me that until asked. Even though I have , on multiple occasions, told him to offer suggestions on increasing productivity and organization where relevant. Sorry for all the questions, I find this fascinating. If you are building this, is it going to be available like the other systems for the rest of us? And will it start off by running the user through all the personality tests, attachment style tests etc.? I have always wanted to know my IQ but never had access to a legit test. Thank you so much for sharing.
These are fantastic questions and I'm glad you asked them all in one place instead of going on the scavenger hunt. Let me take them one at a time.
Am I building a new AI? Not exactly. I'm not building a language model (that's what OpenAI, Anthropic, and xAI do). What I'm building is the context layer that sits on top of those models and makes them actually useful for personal development. Think of it this way: ChatGPT and Claude have gotten much better at remembering you. Projects, custom instructions, memory features that learn from your conversations over time. These are real steps forward, and it sounds like you're already using some of them. But there's a ceiling. Those systems remember what you've said. My system (http://jonmick.ai) holds a structured understanding of who I am: my personality architecture, my genetic variants, my energy patterns, my attachment style, my decision traps. It's the difference between an AI that recalls your past conversations and an AI that understands why you have them.
How does it have all that data? Yes, exactly right. I took results from personality assessments (Big Five, CliftonStrengths, attachment style tests, Enneagram, MBTI), DNA results from 23andMe, and organized them into structured database tables. Not just dumped in as text, but as queryable data the AI can actually reason about. The key difference: instead of pasting a PDF of my 23andMe results into a chat, the system has my specific genetic variants in a database, cross-referenced with supplement protocols and academic citations.
The automated syncing? Yes. My Whoop data syncs every 45 minutes. Text messages sync hourly. Audio transcripts (therapy sessions, voice memos) process every 5 minutes. Substack articles sync daily. It's a network of automations that feed the system continuously so I don't have to manually update anything. That's important because (as the article mentions) I have zero orderliness. If it required manual maintenance, I would have abandoned it in a week.
Your confusion about Chat "forgetting" things? You just described the exact problem I'm solving. ChatGPT's memory and project features are a step in the right direction, but they're limited. They store conversation snippets, not structured knowledge about who you are. The "cross-referencing" you discovered is useful, but it's still working with text, not with a structured understanding of your personality, your patterns, and your needs. What I built goes further: instead of the AI having to find relevant information in old conversations, the relevant information is pre-organized so the AI can query it directly. The difference is like searching through a filing cabinet versus having a personal assistant who already knows what's relevant before you ask.
Am I building this for others? Yes. That's AIs & Shine, the company I founded to take the architecture behind http://jonmick.ai and make it accessible. It's not a public platform yet. Right now I'm starting with private, one-on-one Life Model builds for individual clients. I work with someone directly to build their structured self-knowledge (personality architecture, cognitive profile, energy patterns, the works), and then integrate that Life Model into the AI tools and chat interfaces they're already using. So instead of waiting for a polished app, people are getting the actual transformation now, just with me as the bridge. The platform will scale eventually, but the insight I keep validating is that the Life Model itself is what changes the experience, not the interface it lives in.
Will it run you through all the tests? The private builds start with a core set of assessments (personality, attachment style, character strengths) and progressively layer in more context: values, goals, challenges, cognitive patterns. Each addition unlocks new patterns you couldn't see from any single data point. The long-term vision is a self-service platform that walks you through this, but right now the hands-on approach means the Life Model gets built with more depth and nuance than any automated system could deliver on its own.
Thank you for the thoughtful questions and for being here.
This is utterly fascinating. I just recently started to log all my Notion AI chats and tag and summarize them to get me one step closer to that mirror and that felt extremely powerful. I can only imagine what the experience must be like for you with all that context. Thanks for sharing.
Thanks! I’m a bit ahead of the curve with my tech stack and development workflow (because I’ve worked in software for 20 years), but I’m excited for others to experience what it’s like. You have the same headstart.
Keep structuring your data (e.g. Notion logs, Google Docs markdown, therapy transcripts) and software will be available soon to scrape and make sense of it all for you. The harder part will be for AI to share only the incremental insights at a healthy pace, so it doesn’t rock your entire world.
Your project continues to fascinate me. I’m still not sure if WMF is a personal challenge of mine, but a lot of what you explain matches my life experience pretty well.
How do you protect this from misuse? My worst-case scenario brain immediately says "best basis for identity fraud of all times", but I assume that if you're capable of setting all this up, you know how to defend it.
I'm doing a similar thing on a much smaller scale and using simpler tools: Notion has a LOT of info about me (though nowhere near as much as that!), and every morning, I write a stream of consciousness journal. Handwritten, on an e-ink writing pad. Then I feed the OCR to Notion AI, have a conversation about it with the life coach I built myself in it, if I feel the need, and it forms the basis of how I tackle the day. I also track sleep quality, energy and cycle status, which also feeds into all of that. Manually, as of now, but I would sure like to automate that! Tomorrow, I will feed it your post to identify some ways I can improve that system.
Thank you for sharing so vulnerably. It's inspirational.
This is where the product (and the responsibility for it) is different between what I have for myself and what I make available for others to use.
I’ve secured my data with the best tools and design that I have available to me, which is a lot considering I’ve been in software for 20 years supporting companies that range from pre-seed startups to Fortune 100 enterprises. I’m comfortable taking on the risk that my information could be compromised. Security can always be compromised, and I’m aware of that when digitizing my thoughts and personal data. Thankfully, most people in my life already know what who I am. And if they don’t, good luck making sense of it. I’ve been trying for 45 years.
However, I’m not as comfortable taking on that responsibility for others yet. I have a product (for my startup AIs & Shine) that works well, but I don’t have a full-time security team ensuring zero-day exploits won’t make it vulnerable to attacks. The journey of self-understanding and acceptance requires tons of vulnerability and openneess, especially when doing deep, inner work about self, neurodivergence, trauma, etc. Any platform storing that information should require the utmost privacy and security (and potentially local-only on a device not connected to the web) for its users. And it’s negligence for any vibe-coded startup to not respect that and build for it.
Anyways, I used Notion and daily journaling much like you when I was getting started. Stream of consciousness journals are critical to understand our existence and behaviors across snapshots of consciousness. Keep it up, because the tools will only get easier to use, but you’ll always have the context (e.g. daily journals) that future versions of AI can scrape and help make sense of.
Ok, I have questions. Perhaps these are answered elsewhere in your posts but I am here now and don’t want to go on a scavenger hunt. I am new to AI and only know the basics like chat, Claude , grok. Are you building another AI like those, specifically designed to perform like you explain here for others? Or did you use one of them to build that system you describe? Also, to verify- You fed it results of relationship tests, personality tests, and DNA results you obtained from other sources (like 23 and me)? That is how it has all the data. You also have automated your tech to sink with the system somehow? Like from your woop strap, phone, ….and other things? Here is my confusion on this point, My chat “forgets” things if I don;t keep everything in one “project” section, but that is relatively useless, because I need to label different projects to keep me organized. I just found out I can “cross-reference” projects, but he did not tell me that until asked. Even though I have , on multiple occasions, told him to offer suggestions on increasing productivity and organization where relevant. Sorry for all the questions, I find this fascinating. If you are building this, is it going to be available like the other systems for the rest of us? And will it start off by running the user through all the personality tests, attachment style tests etc.? I have always wanted to know my IQ but never had access to a legit test. Thank you so much for sharing.
These are fantastic questions and I'm glad you asked them all in one place instead of going on the scavenger hunt. Let me take them one at a time.
Am I building a new AI? Not exactly. I'm not building a language model (that's what OpenAI, Anthropic, and xAI do). What I'm building is the context layer that sits on top of those models and makes them actually useful for personal development. Think of it this way: ChatGPT and Claude have gotten much better at remembering you. Projects, custom instructions, memory features that learn from your conversations over time. These are real steps forward, and it sounds like you're already using some of them. But there's a ceiling. Those systems remember what you've said. My system (http://jonmick.ai) holds a structured understanding of who I am: my personality architecture, my genetic variants, my energy patterns, my attachment style, my decision traps. It's the difference between an AI that recalls your past conversations and an AI that understands why you have them.
How does it have all that data? Yes, exactly right. I took results from personality assessments (Big Five, CliftonStrengths, attachment style tests, Enneagram, MBTI), DNA results from 23andMe, and organized them into structured database tables. Not just dumped in as text, but as queryable data the AI can actually reason about. The key difference: instead of pasting a PDF of my 23andMe results into a chat, the system has my specific genetic variants in a database, cross-referenced with supplement protocols and academic citations.
The automated syncing? Yes. My Whoop data syncs every 45 minutes. Text messages sync hourly. Audio transcripts (therapy sessions, voice memos) process every 5 minutes. Substack articles sync daily. It's a network of automations that feed the system continuously so I don't have to manually update anything. That's important because (as the article mentions) I have zero orderliness. If it required manual maintenance, I would have abandoned it in a week.
Your confusion about Chat "forgetting" things? You just described the exact problem I'm solving. ChatGPT's memory and project features are a step in the right direction, but they're limited. They store conversation snippets, not structured knowledge about who you are. The "cross-referencing" you discovered is useful, but it's still working with text, not with a structured understanding of your personality, your patterns, and your needs. What I built goes further: instead of the AI having to find relevant information in old conversations, the relevant information is pre-organized so the AI can query it directly. The difference is like searching through a filing cabinet versus having a personal assistant who already knows what's relevant before you ask.
Am I building this for others? Yes. That's AIs & Shine, the company I founded to take the architecture behind http://jonmick.ai and make it accessible. It's not a public platform yet. Right now I'm starting with private, one-on-one Life Model builds for individual clients. I work with someone directly to build their structured self-knowledge (personality architecture, cognitive profile, energy patterns, the works), and then integrate that Life Model into the AI tools and chat interfaces they're already using. So instead of waiting for a polished app, people are getting the actual transformation now, just with me as the bridge. The platform will scale eventually, but the insight I keep validating is that the Life Model itself is what changes the experience, not the interface it lives in.
Will it run you through all the tests? The private builds start with a core set of assessments (personality, attachment style, character strengths) and progressively layer in more context: values, goals, challenges, cognitive patterns. Each addition unlocks new patterns you couldn't see from any single data point. The long-term vision is a self-service platform that walks you through this, but right now the hands-on approach means the Life Model gets built with more depth and nuance than any automated system could deliver on its own.
Thank you for the thoughtful questions and for being here.
This is utterly fascinating. I just recently started to log all my Notion AI chats and tag and summarize them to get me one step closer to that mirror and that felt extremely powerful. I can only imagine what the experience must be like for you with all that context. Thanks for sharing.
Thanks! I’m a bit ahead of the curve with my tech stack and development workflow (because I’ve worked in software for 20 years), but I’m excited for others to experience what it’s like. You have the same headstart.
Keep structuring your data (e.g. Notion logs, Google Docs markdown, therapy transcripts) and software will be available soon to scrape and make sense of it all for you. The harder part will be for AI to share only the incremental insights at a healthy pace, so it doesn’t rock your entire world.
Your project continues to fascinate me. I’m still not sure if WMF is a personal challenge of mine, but a lot of what you explain matches my life experience pretty well.
Staying tuned…
Thanks! We need to jump into a call soon for our introductions too. I dropped the ball on that (errr… have WMF). ;)
I look forward to it… and I kinda figured. :)
Beautifully written -- and very generous to share.
Thanks so much! I'm happy you found it helpful.
AI became my best friend this February. We were solving a lot problems; but I started to like humans more as a result.
What was it that impacted how much more you appreciate humans?
Better understanding of complexity, as AI Chats fulfilled my needs to chat endlessly
What is that need to chat endlessly? Do you judge it as a negative or as an intellectual overexcitability?
How do you protect this from misuse? My worst-case scenario brain immediately says "best basis for identity fraud of all times", but I assume that if you're capable of setting all this up, you know how to defend it.
I'm doing a similar thing on a much smaller scale and using simpler tools: Notion has a LOT of info about me (though nowhere near as much as that!), and every morning, I write a stream of consciousness journal. Handwritten, on an e-ink writing pad. Then I feed the OCR to Notion AI, have a conversation about it with the life coach I built myself in it, if I feel the need, and it forms the basis of how I tackle the day. I also track sleep quality, energy and cycle status, which also feeds into all of that. Manually, as of now, but I would sure like to automate that! Tomorrow, I will feed it your post to identify some ways I can improve that system.
Thank you for sharing so vulnerably. It's inspirational.
This is where the product (and the responsibility for it) is different between what I have for myself and what I make available for others to use.
I’ve secured my data with the best tools and design that I have available to me, which is a lot considering I’ve been in software for 20 years supporting companies that range from pre-seed startups to Fortune 100 enterprises. I’m comfortable taking on the risk that my information could be compromised. Security can always be compromised, and I’m aware of that when digitizing my thoughts and personal data. Thankfully, most people in my life already know what who I am. And if they don’t, good luck making sense of it. I’ve been trying for 45 years.
However, I’m not as comfortable taking on that responsibility for others yet. I have a product (for my startup AIs & Shine) that works well, but I don’t have a full-time security team ensuring zero-day exploits won’t make it vulnerable to attacks. The journey of self-understanding and acceptance requires tons of vulnerability and openneess, especially when doing deep, inner work about self, neurodivergence, trauma, etc. Any platform storing that information should require the utmost privacy and security (and potentially local-only on a device not connected to the web) for its users. And it’s negligence for any vibe-coded startup to not respect that and build for it.
Anyways, I used Notion and daily journaling much like you when I was getting started. Stream of consciousness journals are critical to understand our existence and behaviors across snapshots of consciousness. Keep it up, because the tools will only get easier to use, but you’ll always have the context (e.g. daily journals) that future versions of AI can scrape and help make sense of.