Building a Personal AI Infrastructure (PAI)

All right, I've got a million things to talk about and share in this episode, but ultimately I want to ask—and answer—the question of what are we actually doing with all these AI tools.

What I've been building ​

I'm just as excited as the next person about all the tooling. I've literally spent hundreds of hours on my MCP config and now I've probably spent a couple hundred hours on all of my agents and sub-agents and tool chains and everything.

And now with Claude Code, I have started unifying everything together!

So what I wanna do is show you what I've been building, how all the pieces fit together, and then give you a few examples of it in action.

A mission reminder ​

So before I get started, I want to just mention like what I'm actually doing and what my mission is and that sort of thing. So I have a company called Unsupervised Learning, which used to just be the name of the podcast, but now it's a company. And essentially my mission is to upgrade humans and organizations, but mostly humans.

I'm basically building a thing called Human 3.0 which is a human-focused future and that's something else I've already done a video on that which you can go check out. But I want to get that out of the way and let you know basically what all this tooling and all these products and stuff that you see me putting out in my channel like what is it all going towards? It's going towards that.

Personal AI Infrastructure ​

So the main topic and the umbrella of everything I'm gonna talk about today is what I call a Personal AI Infrastructure (PAI), which is PAI for an acronym. Everyone likes pie. It's also one syllable, which I think is an advantage.

And the larger context for this is the feature that I talked about in my really shitty very short book in 2016, which was called The Real Internet of Things. And basically has four components:

• Digital Assistant
• The API-fication of everything
• Augmented reality
• The ability for agents to make things happen once everything has an API

Those are the four components of the book and I think those are coming to pass now.

Progress toward the vision ​

But we don't have DAs yet. We have lots of different things:

• Digital companions
• ChatGPT memory and larger context windows
• Personality features in ChatGPT memory

So it's definitely moving in the direction of Digital Assistants. And then the second piece and sort of its progress so far is the API-fication of everything and of course that is being manifested as MCP. And you can see the quote here of what I said in the book in 2016 about everything getting an API and who knows what the protocol will be, but it will probably be HTTP-based and guess what it is.

So Meta and some other companies are working on the third piece which is augmented reality and then the fourth piece is just like orchestration.

The brain of Personal AI Infrastructure ​

What I'm most focused on is building the brain of this Personal AI Infrastructure which is the thing that actually understands you the best. It's basically your best friend. And here we actually want to go and overlay some images from the AI predictable future post.

What the brain needs to have ​

It's basically a full context dump of everything about you and your AI has access to all of that, and it has its own personality. And then what it does is actively and passively advocate on your behalf.

So we don't have that piece yet. Even with Claude Code it doesn't feel like a personality yet and I don't really have the Claude Code functionality and capabilities inside of ChatGPT. And unless you use one of the AI companion products that's not really unified with any of this other stuff. So I think we're still a little ways away from actual DA.

But what we're not far away from, or at least starting to happen, is all the capabilities that the DA will have starting to come together and that's what I'm really excited to show you today, starting with tools.

Unifying tools and giving them to your AI ​

So the thing I'm so excited about is I have now unified tons of different tools within my main Claude Code instance. And by the way, sometimes I use OpenCode as well. I kind of go back-and-forth but this week I've been mostly using Claude Code.

The unified toolset architecture ​

So this is the list of tools I have defined in my Claude Code and notice how I'm actually talking about these. I'm actually saying hey look here's all the tools that you have and here's how you can get to them.

And I am including in there more than just tools. It's resources available right? You have a staff right? You have agents.

Agents ​

Here's my current lineup of specialized agents:

.claude/agents/
├── engineer.md      # TypeScript/Bun development specialist
├── pentester.md     # Security assessment expert
├── designer.md      # UI/UX and visual design
├── marketer.md      # Product positioning & copy
├── gamedesigner.md  # RPG mechanics & narratives
└── qatester.md      # Quality assurance & testing

Fabric ​

You also have access to Fabric which you could check out a link in description. That's a project I built in the beginning of 2024. It's a whole bunch of prompts and stuff, but it gives you—or my Digital Assistant, whose name is Kai by the way—the ability to go and make custom images for anything using context.

• Problem solving for hundreds of problems
• Custom image generation
• Web scraping with jina.ai (fabric -u $URL)

Commands ​

And then I have all these commands that are also there. So .claude/agents are the agents, .claude/commands are the commands and then I have separately all my MCP servers.

.claude/commands/
├── write-blog-post.md          # AI-powered blog writing
├── add-links.md                # Enrich posts with links
├── create-custom-image.md      # Generate contextual images
├── create-linkedin-post.md     # Social media content
├── create-d3-visualization.md  # Interactive charts
├── code-review.md              # Automated code review
├── analyze-paper.md            # Academic paper analysis
├── author-wisdom.md            # Extract author insights
├── youtube-to-blog.md          # Convert videos to posts
└── ... 20+ more specialized commands

Also any Fabric pattern because they're just commands and Fabric patterns are just Markdown files! 🤯

MCP Servers and Tools ​

What do I have in my MCP servers? Here's my actual .mcp.json config (with API keys redacted):

{
  "mcpServers": {
    "playwright": {
      "command": "bunx",
      "args": ["@playwright/mcp@latest"],
      "description": "Browser automation for troubleshooting web apps"
    },
    "content": {
      "type": "http",
      "url": "https://content-mcp.danielmiessler.workers.dev",
      "description": "My content archive going back to 1999"
    },
    "naabu": {
      "type": "http",
      "url": "https://naabu-mcp.danielmiessler.workers.dev",
      "description": "Port scanning and service detection"
    },
    "httpx": {
      "type": "http",
      "url": "https://httpx-mcp.danielmiessler.workers.dev",
      "description": "Web probes and tech stack analysis"
    },
    "brightdata": {
      "command": "bunx",
      "args": ["-y", "@brightdata/mcp"],
      "description": "Ultimate web scraping capabilities"
    }
  }
}

How Kai discovers its capabilities ​

And I basically say very clearly in this thing right at the beginning of my CLAUDE.md that look, this is everything you have access to and when you get ready to process any request from me, first go and look at all your tools you have available. And I actually don't have them defined directly in the CLAUDE.md. I tell Kai where they're actually located so it's going to go to those different locations and find them. This helps not load a giant context file into memory right from the very beginning.

Built-in tools ​

• fetch
• reading files
• etc.

What I'm building with this infrastructure ​

So basically, that's what the tool infrastructure looks like. Claude now understands all these MCPs which have a couple of dozen tools on them.

The power of unified tooling ​

With this setup I can:

• Fetch any quote or blog or content going all the way back to 1999 from my website
• Create any custom image using contextual understanding
• Run any of the 219 different Fabric patterns to analyze content
• Create threat reports, do security testing, anything
• Use all my different agents to perform various tasks
• Build something brand new using a combination of the agents

So basically, I have all this stuff that I want to be able to do and that I want to enable other people to be able to do. If I'm helping an artist try to transition out of the corporate world into becoming a self-starting artist (which is what I talked about in Human 3.0), if I want them to become independent and have their own studio, brand, and everything, I'm thinking about:

• What are their favorite artists?
• Where are they going physically in the world?
• Can they go meet them, talk to them, get coffee with them?
• What's the new art styles that are coming out?
• Are there some technique videos that they could watch to improve their painting technique?

What I'm about is helping people create the back-end AI infrastructure that will enable them to transition to this more human world where they're not dreading Monday, dreading being fired in office politics and all of that bullshit.

My personal use cases ​

Building modular services ​

What that looks like for me is I want to be able to:

• Gather all data sources
• Parse all those data sources
• Extract meaningful insights
• Create actionable intelligence

That's what I'm building. I've got services here that can do anything with incoming content. For example:

# Parse YouTube video and extract insights
fabric -u https://youtube.com/watch?v=... | fabric --pattern extract_wisdom

# Analyze a research paper
cat paper.pdf | fabric --pattern analyze_paper

# Create threat intelligence from news
curl https://news-site.com | fabric --pattern create_threat_report

Example: The Threshold product ​

I have a product that I've built called Threshold. And it actually looks at the top 3000 of my best content sources:

• My favorite YouTube people
• My favorite blogs
• RSS everything

And it actually runs through this prompt called Categorize, Label, and Rate. It sorts into different quality levels of content, which tells me "Do I need to go watch it immediately in slow form and take notes?" Or can I skip it? So it's a better version of internet for me.

And this is like a really crucial point. That app is actually made from components of these other services. I'm building these services in a modular way that can interlink with each other. By calling them in a particular order and putting a UI on that, and putting a Stripe page on that, guess what I have? I have a product.

Example: Intelligence gathering system ​

Another example of what I'm building right now. I have a whole bunch of people that are really smart in OSINT right? They read satellite photos and they can tell you what's in the back of a semi truck. Right? Super smart. And there's hundreds of these people.

Well I'm gonna:

• Parse all of what they're saying
• Turn that into a daily Intel report for myself
• Parse the daily ones and turn into a weekly one
• Turn that into a monthly one
• Look at all of them and find trends that these people are seeing without even knowing it

So I'm building myself an Intel product because I care about that. So I could see different cyber threat actors and what types of targets are they going after. I just do all sorts of stuff and it's all based on building and combining these small little pieces and unifying them together into a unified tool set here.

And then given that I just hand those tools and all those services and all those endpoints to Kai right here at this infrastructure. I say, "Okay, here's my goal. Here's what I'm trying to do." And it spins it all up, spins it all up and goes and just executes on it using all those individual pieces.

Caveats of working with infrastructure like this ​

There are a number of things that will catch you when you're trying to build this for the first time.

1. One example is that even if you explained to your main agent exactly how you like them to do things, you have to be sure that the agent is actually explaining that to sub-agents when it gives out tasks.
2. Another example is that you want to be really good about writing descriptions for all your various tools because those descriptions are critical for how your agents and subagents are going to figure out which tool to use for what task. So spend a lot of time on that.

I've put tons of effort into the back-and-forth explaining different components of this plumbing.

What's so exciting is that it's like tightening up these repeatable modular tools. The better they get, the less they go off the rails, and the higher quality the output you get of the overall system. It's absolutely exhilarating.

What this becomes ​

Why we should care about AI progress ​

What I'm getting at is you see all these videos talking about "oh look what OpenCode can do". Look what Claude Code can do. Look what this new model can do. I encourage you to ask yourself—why do we care? What are we building that gets enhanced by all these pieces getting better and better?

My answer to that is this. This is what I'm building; this is what I've always wanted; this is what it will make me—10 versions of me, 100 versions of me, a thousand versions of me.

The future I'm building toward ​

Because 2 years from now, 5 years from now, 10 years from now Kai is going to be talking into my ear, putting stuff in my glasses when I walk around. I'm going to be able to look and say:

• "Ok that area is dangerous over there"
• "That area looks like food over there"
• "Turn left here" with a little pop-up icon

How did I get a pop-up icon on my interface? Because Kai queried an API for a company who provides location services and UI services. Kai is using a UI in my glasses which is actually an API for a company and the data is from a separate company. So basically all the companies that we know and love they all provide APIs which are designed not to be used by us but by our Digital Assistants. So Kai is basically building this world around me and curating this world around me and trying to make changes to it to optimize my life.

And he could do that better than anyone else because:

1. He could talk to all those thousands of APIs, which I can't do.
2. He knows everything about me.

Why I wake up at 3:30am to code ​

This is what I'm building. This is what I'm so excited about. This is why I love all this tooling. This is why I'm having difficulty sleeping because I'm so excited. I wake up at 3:30 in the morning and I go and code accidentally for six hours. Making a new service, tweaking the service, tweaking the system prompt, turning it into a sub-agent, handing it over to Kai. This is my life now, and I hope you see why I'm so excited about it, and I hope you're excited about it too.

I do encourage you to definitely hit subscribe if this is the type of thing that resonates with you. I don't say that enough. Please subscribe and follow, and I'll see in the next one!

Possible Examples of PAI Infrastructure for Different Professions ​

You might be thinking, "This is great for tech people, but what about everyone else?" Here's the thing—PAI isn't just for developers. Every profession has repetitive tasks, scaling challenges, and opportunities for AI augmentation. Let me show you what's possible.

For the Independent Artist ​

Imagine you're an artist trying to make it on your own. You're talented, but you're drowning in the business side. Here's what your PAI could look like:

Creative Enhancement Services:

• Style Evolution Tracker - Analyzes your portfolio over time to show how your style is developing
• Inspiration Curator - Scrapes Pinterest, Instagram, museums worldwide for references matching your aesthetic
• Technique Video Librarian - Finds and organizes tutorials for exactly the skills you want to develop
• Color Palette Generator - Extracts palettes from any image, mood, or music

Business Automation Tools:

• Gallery Pitch Writer - Generates professional proposals based on your work and the gallery's style
• Commission Manager - Tracks projects, deadlines, client communications
• Fair Price Calculator - Analyzes market data to suggest pricing for your work
• Social Media Scheduler - Maintains consistent online presence while you're in the studio

Growth & Community Services:

• Artist Network Mapper - Identifies potential mentors and peers in your area
• Opportunity Finder - Monitors calls for artists, residencies, grants that match your profile
• Collector Relationship Manager - Tracks interactions, preferences, purchase history
• Show Documentation Bot - Automatically photographs and catalogs your exhibitions

For Teachers & Tutors ​

You became a teacher to change lives, not to drown in paperwork. Your PAI handles the administrative overhead while amplifying your impact:

Student Success Services:

• Learning Style Analyzer - Uses interaction data to identify how each student learns best
• Progress Tracker - Real-time dashboards showing exactly where each student stands
• Struggle Point Detector - Alerts you when a student is stuck before they fall behind
• Parent Communication Bot - Generates personalized weekly updates for every family

Content Creation Tools:

• Standards-Aligned Lesson Builder - Input topic, get differentiated lesson plans
• Smart Quiz Generator - Creates assessments from any content with varying difficulty
• Interactive Activity Designer - Turns concepts into engaging, hands-on exercises
• Visual Aid Factory - Generates diagrams, charts, and infographics on demand

Administrative Automation:

• Grading Assistant - Applies rubrics consistently, highlights areas needing attention
• IEP/504 Compliance Checker - Ensures all accommodations are properly implemented
• Meeting Note Summarizer - Extracts action items from parent conferences
• Schedule Optimizer - Maximizes actual teaching time, minimizes transitions

With this infrastructure, you can give every student the personalized attention they deserve while actually having evenings and weekends to yourself.

For Church & Community Leaders ​

Leading a congregation means being everywhere at once. Your PAI helps you be present for everyone while protecting your own spiritual health:

Pastoral Care Services:

• Member Care Tracker - Maintains confidential records of prayer requests, hospital visits, life events
• Crisis Response Coordinator - Instantly mobilizes support teams when emergencies arise
• Counseling Session Manager - Secure notes, follow-up reminders, resource recommendations
• Check-in Scheduler - Ensures no one falls through the cracks during difficult times

Community Building Tools:

• Small Group Matchmaker - Connects people based on life stage, interests, spiritual goals
• Volunteer Gift Matcher - Aligns people's talents with ministry needs
• Event Planning Assistant - From weekly potlucks to annual conferences
• New Member Journey Designer - Creates personalized integration paths

Communication Amplifiers:

• Sermon Series Architect - Plans thematic content months in advance
• Omnichannel Broadcaster - One message to email, text, app, and social media
• Prayer Chain Manager - Securely distributes requests while maintaining privacy
• Story Collector - Gathers and organizes testimonies for perfect timing

For Starting Entrepreneurs ​

Starting a business is like drinking from a firehose while juggling flaming torches. Your PAI becomes your co-founder who never sleeps:

Business Foundation Services:

• Legal Document Generator - Creates customized contracts, NDAs, terms of service
• Entity Structure Advisor - Walks through LLC vs S-Corp vs C-Corp decisions
• Compliance Calendar - Tracks every license, permit, and filing deadline
• Financial Autopilot - Bookkeeping, invoicing, expense tracking, tax prep

Market Intelligence Tools:

• Competitor Surveillance - Monitors pricing changes, feature launches, pivots
• Customer Interview AI - Transcribes and extracts key insights from discovery calls
• Market Sizing Engine - Calculates TAM, SAM, SOM with real data
• Opportunity Radar - Spots emerging trends in your industry

Revenue Generation Systems:

• Outreach Personalizer - Crafts individual cold emails that don't feel cold
• Social Proof Harvester - Collects and organizes testimonials across all channels
• Content Machine - Generates SEO-optimized blog posts, social media, newsletters
• Pricing Optimizer - A/B tests pricing models based on conversion data

Scaling Infrastructure:

• Process Documentor - Turns your actions into repeatable SOPs
• Talent Scout - Writes job posts, screens applicants, suggests interview questions
• Customer Success Engine - Automates onboarding, support tickets, health scores
• Fundraising Assistant - Builds pitch decks, financial models, investor CRM

It works for any profession ​

Whether you're an artist, teacher, pastor, or entrepreneur, the pattern is the same:

1. Identify your unique value - what only you can do
2. Map out everything else that drains your time
3. Build AI services to handle the draining tasks
4. Use freed time to amplify your unique value

This isn't about replacing humans. It's about replacing the robotic parts of human work so humans can be more human.

A fun confession about this post ​

Okay, and now for a bit of a confession—you may have guessed already, but just to show how powerful this entire thing is, this system that this post talks about actually wrote this post.

All the code, all the asides, all the images were generated using it. The whole post was written by this system. I typed probably 10% and as far as the flow of the text, I just dictated the whole thing using Wispr Flow. All the paragraph breaks and formatting choices were done by it, following my rules laid out in the custom commands.

So the system really works!

Summary ​

1. Everyone's excited about AI tools (me included), but I think it's critical to think about what we're actually building with them
2. My answer is a Personal AI Infrastructure—a unified system of agents, tools, and services that actually knows and serves you
3. This isn't just for tech people—anyone can build modular services that map to their actual life goals and amplify their work
4. By combining agents, MCP servers, custom commands, and Fabric patterns, you create something way more powerful than individual tools
5. The future is your DA actively advocating for you in the world, constantly, making your life better moment by moment, using all these services that you have set up

I really hope this gets you as excited as I am to build your own infrastructure. We've never been this empowered with technology to pursue our human goals.

Go build!