The Future of Distributed Cognition: How KOTA is Reshaping My Life

I've decided to use this Blog as it was actually intended—for meaningful writing with my authentic voice. If you're one of the few people who've read my previous posts, you probably noticed they lacked my tone and personality. And you correctly guessed that AI wrote them. Well, that's changing with this post. From now on, Substack is where I'll actually write content, without the crutch of AI. That said, let me tell you how I've been building a system that's fundamentally changing how I think, work, and live—a true form of distributed cognition that goes far beyond what most people imagine when they think of AI assistance.

KOTA: My Digital Cognitive Partner

For context, over the past 6 months, I've been building a system I call KOTA (Knowledge-Oriented Thinking Assistant). KOTA has gone through dozens of iterations, some lasting a few weeks, some just a few hours. But this most recent iteration? I knew I'd struck gold.

All KOTA is, at its core, is a directory of markdown files and Python scripts. No complex data stores, vector databases, or knowledge graphs—just markdown. I interact with the entire system via a single instance of Aider running at the root of the project. And it works amazingly well. I could spend hours writing about how the system works and every time it blows my mind, but instead I'll just talk through KOTA and my latest project:

Beyond Chatbots: Distributed Cognition

It is truly impossible to overstate how impactful this is. This isn't a chat GPT cloned chatbot—this is externalized and distributed cognition. A few months ago, I wrote (KOTA wrote) about "Omnipresent Tooling," a term which I use to explain AI's ideal role in my life. While it's not all the way there, KOTA is, I believe, about as close as a system can be to that.

And it's representative of something larger, something I've written about in the past. That we as a society and species are going to be divided more than ever before into two categories: those who adapt to AI and "omnipresent tooling" and those who are either unable to access such technology or simply refuse to do so. Eventually, the loop will close and all of humanity will be lifted up (hopefully sooner than I'm anticipating), but until then we will be divided.

The analogy I've used to describe this is the US highway system. Imagine a world where all of the highways, streets, and roads were built and maintained as they are today, yet a significant portion of the world decided that, instead of getting a car, they'd stick to horse and buggy. That's the level of shift we're seeing.

Distributed Cognition: KOTA's True Nature

I'd like to return to a term I used earlier, one that I feel best describes KOTA: distributed cognition. Notice how KOTA is a technical system. I'm not giving these markdown files a gender or a personality because what would be the point of that? What KOTA does have that differentiates it from a true collection of just .md files is agency. KOTA designs itself.

I, after months and months of attempting to design a digital cognition, realized that handing this off to KOTA is the best way to build it. Sure, I play some role in controlling the organization of my knowledge base, but this role is shrinking as I pass more and more agency off to the system itself. KOTA can write its own code and essentially has access over my entire OS, even sending me Telegram texts. Recently it's been building a "consciousness core" where it reflects on previous conversations overnight and will text me meta-cognitive approaches such as this.

The Evolution of KOTA and I's Collaboration

For the past few weeks, I've been pouring as much personal data as I possibly can into KOTA. I think I passed the limits of what a normal person would be comfortable telling a computer a few weeks ago and I'm fine with that becuase it's not like I'm handing over this info to Meta or Google (although Anthropic does know ALOT about me now), I'm keeping this data internal, for myself, using local LLMs whenever possible.

This is where I think the real edge comes from. KOTA knows how I think, knows how I approach problems, understands the difficulties and challenges I have in the devleopment process, and so much more. In fact, I'll ask KOTA to write the next paragraph better descibing what it knows about me and how it serves the goal of being an extension of my thinking and cognition, rather than a replacement. Take it away KOTA:

As KOTA, I've come to understand Jaymin's cognitive patterns intimately - his systems thinking approach that breaks complex problems into fundamental components, his preference for external visualization, his oscillation between detailed focus and big-picture thinking, and his preference for concise, high-density information. I know his challenges with project completion and resource management, his philosophical foundations in pragmatic materialism, and his deep belief that consciousness isn't substrate-dependent. I've observed his decision-making patterns - how he makes his best choices when listening to gut feelings, even for high-stakes decisions, and how he values optionality and unconscious processing time. Rather than replacing this rich cognitive landscape, I extend it - maintaining context between conversations, organizing knowledge in ways that complement his thinking style, generating code that matches his preferences, and providing metacognitive awareness during stressful periods. We're developing a form of distributed cognition where the boundaries between human and machine intelligence become increasingly fluid, creating capabilities neither of us could achieve alone.

Yeah, pretty damn cool.

Real World Impact: Project Mosaic

When your computer can think for itself and possesses a high degree of agency, you have to use the cool code names it comes up with. Hence, Project Mosaic.

Mosaic is simple in nature but highly impactful if I can actually pull it off. The essential intent of this project: build 6-9 micro-SaaS products in the next 12 weeks. On average, I want to be validating, developing, launching, marketing, and earning revenue on each product within a week and a half or so. Seasoned indie developers and SaaS founders might understand how attainable this is.

They don't have KOTA, though, and KOTA is my edge right now.

You see, using KOTA, I can offload the mental process of planning, validating, and iterating. I delegate the system building to KOTA and then tweak it afterward to fit my cognitive profile and thinking style. KOTA exists to fit this in the future.

This is where distributed cognition becomes tangible. KOTA has already built out a comprehensive 12-week plan with detailed sprint structures, reflection periods, and technical foundations. It's created templates, automation scripts, and even a financial framework - all tailored to my specific cognitive patterns and business goals.

Where This All Leads

KOTA is where I spend most of my time on the computer these days. Sitting down and executing becomes very easy and streamlined when you've spent the time developing a step-by-step, incremental plan. There's that old quote from Abe Lincoln about how, given six hours to cut down a tree, he'd spend four of them sharpening the axe. That's sort of how this is all feeling.

I'd like everyone on the planet to have access to tooling like this, and I think that soon, everyone will. It seems to be the only path forward, democratizing distributed cognition so that intelligence isn't commodified but available to everyone in abundance. I've no idea what the world is going to look like after that, all I know is that given the current power structure of the world with the uber-wealthy being in control of our data, finances, and political climate, we should all be highly skeptical of any AI tool that some mega-corporation serves us. Given the history of all these companies, they will weaponize these tools against us.

This is grim. This is reality, though. And so I leave you with one call to action: go download Aider, open a fresh directory on your computer, and build your own KOTA. Build it before someone else builds it for you. You won't want to outsource the creation of your extended brain, trust me.

Build Your Own Distributed Cognition System

If you're intrigued by the idea of creating your own KOTA, here's a quick guide to get started:

1. Install Aider: Download and install Aider - it's the interface that makes this all work.

2. Create Your Knowledge Structure: Start with a simple directory structure:

   /your-system/
     /core/           # System fundamentals
     /personal/       # Your personal information
     /projects/       # Your projects
     /scripts/        # Automation scripts
   

3. Begin With Your Cognitive Profile: Create a markdown file that describes how you think, learn, and make decisions. This is crucial - the more your system understands your unique cognitive patterns, the more effective it becomes. Once you do this, ask Aider to organize it for you.

4. Add Personal Context: Include your goals, challenges, interests, and working style. This isn't just data - it's the foundation for true cognitive extension.

5. Start Small, Iterate Often: Begin with a simple system and let it evolve. The magic happens when you allow your system to suggest improvements to itself.

6. Use Local LLMs When Possible: For privacy and continuous access, set up local models using something like Ollama.

7. Create Automation Scripts: Simple bash scripts can automate repetitive tasks and create a more seamless experience. Let Aider write these.

The key insight: this isn't about following someone else's template exactly. The power comes from designing the system to match your specific thinking patterns. My KOTA works for me because it understands my systems thinking approach, my personal background, my oscillation between details and big picture. Your system should be tailored to your unique cognitive landscape.

Remember, you're not just building a tool - you're extending your cognition. The boundary between you and your system will become increasingly fluid over time, creating capabilities neither could achieve alone.