Uncrowded Strategies #012: I hired three employees this week. None of them are human.

Welcome to this week's edition of Uncrowded Strategies.

This one's different. I'm only covering two topics today — The Leverage and The Portfolio. There's so much happening in the AI world right now that I wanted to go deeper instead of wider. I'll explain more at the end.

Let's get into it.

(Current Focus: Building a Team... of AIs)

I'm still building Sonny — my OpenClaw bot. But this week, something shifted.

As I've been building out the multi-agent system to run my company, I hit a wall. To automate everything, I first need strong systems and software underneath it. The AI agents need tools to use. Without those tools... the agents are just smart people sitting in an empty office with no computer.

Think of it like a construction site.

You can hire the best crew in the world. But if there's no blueprint, no materials, no scaffolding? They're just standing around. The foundation has to come first.

So I stopped trying to automate everything at once. I stepped back and asked: What does my AI team actually need to do their job?

And that's when I started building three core agents:

1. The Architect — runs on Claude Code using Claude Opus 4.6. This agent designs the system. It draws the blueprint. It looks at the problem, maps out the structure, and plans how everything connects. Then it reviews the final product for quality.

2. The Builder — runs on CodeX using GPT 5.4. This is the construction crew. Once the Architect draws the plans, the Builder executes. I chose CodeX here because GPT 5.4 is more thorough when it comes to building. It doesn't cut corners.

3. The QA — runs on Gemini 3.1. This is the building inspector. After the Builder finishes, the QA reviews the work. It checks for cracks. It stress-tests the structure. It flags what's broken before anything goes live.

Here's what gets me excited...

I used an open-source project called Agentchattr, like a group chat but for agents, and put all 3 in the same chat room!

I'm literally getting AIs to work together. Cross-referencing. Checking each other's work. Catching each other's mistakes.

One model designs. A different model builds. A third model inspects. WHILE I SLEEP…

The quality is significantly better than using one model for everything. Because here's the thing most people miss: every AI model has blind spots. Claude is brilliant at architecture and reasoning. GPT is thorough at building. Gemini catches things the others don't.

When you pit them against each other — not competing, but collaborating — you get something closer to a real team than anything I've ever built with software.

The Uncrowded Take: Most people are still asking AI to do one thing at a time. Write me an email. Summarize this doc. Answer this question.

That's like hiring a construction crew and asking them to hand you one brick.

The real leverage isn't one AI doing one task. It's multiple AIs working as a team — each one doing what it's best at, checking the others, and building something none of them could build alone.

I'm not managing people anymore. I'm managing a crew of AIs. And the construction site is my entire company.

I want you to imagine something.

Rewind three and a half years. December 2022. ChatGPT just launched.

1 million users in 5 days. The internet exploded. Your feed was full of people posting screenshots. "Look what it can do!" Hot takes everywhere. Hype. Noise. Chaos.

And then... the noise died down. People moved on.

But something had permanently shifted. Everyone now knew that AI was real. It was no longer a research paper or a tech demo. It was a consumer product. A mass-market moment.

That was the ChatGPT Moment.

And if you had looked underneath the noise for the signal, the bets were obvious:

The people who made money weren't the ones posting screenshots. They were the ones asking: "If this is real... who builds the picks and shovels?"

I believe we just witnessed another moment. The “OpenClaw” Moment.

About two months ago, OpenClaw went viral. By now, you've probably heard of it — it's the open-source AI agent I'm building Sonny on.

But here's the critical difference most people are missing:

The ChatGPT Moment made people believe in AI. The OpenClaw Moment made people act on it.

ChatGPT was a chatbot. Impressive, but passive. You ask, it answers. A fancy search engine with personality.

OpenClaw is an agent. It doesn't just answer. It does the work. It reads your emails. It deploys code. It manages your calendar. It operates autonomously like a digital employee.

That's not a chatbot upgrade. That's a different category entirely.

And right now? You're hearing the same noise. Social media is full of people posting their OpenClaw setups. "Look what it can do!" Hot takes everywhere. Hype. Noise. Chaos.

Sound familiar?

Right after the OpenClaw Moment, Block (the company behind Square and Cash App) announced it was cutting 40% of its workforce — over 4,000 people — and replacing them with AI agents. The stock surged 24% in premarket trading.

Read that again. A publicly traded company fired nearly half its staff because of AI. And investors cheered.

The noise will die down. The impact won't.

So again: look underneath the noise for the signal.

The obvious layer — chips and energy — is still a safe bet. If the ChatGPT Moment created demand for AI infrastructure, the Agentic Wave will multiply that demand 3-5x. Because agents don't just chat. They run. They consume compute around the clock. They need power. They need chips. That's not slowing down. It's accelerating.

But here's what I've been digging into...

The layer underneath the obvious.

Whether you think Block is genuinely using AI or just dressing up cost cuts (there are strong arguments on both sides), the signal is clear: the market rewards companies that align themselves with the agentic future.

So I did some digging. If the ChatGPT Moment was about "who builds the infrastructure," the OpenClaw Moment is about: "Who captures the value when AI agents go mainstream?"

I'm looking at small positions in two companies. Maybe 1-2% of my portfolio each.

Cloudflare (NET) — Here's the angle most people aren't thinking about: when millions of AI agents start browsing the web, making API calls, and interacting with services 24/7... who handles the traffic? Who secures it? Cloudflare is already building the infrastructure for this. They're classifying agent traffic differently from human traffic. They just launched AI security products specifically for agent-powered applications. They're positioning themselves as the security guard for the agentic internet. Strong financials. Cash flow positive. And sitting right in the path of a tidal wave that most investors haven't priced in yet.

UiPath (PATH) — This one is interesting. UiPath started as a robotic process automation (RPA) company — basically, software that clicks buttons for you. Old school. Boring. But last year, they pivoted hard into AI Agent orchestration. They built a platform called Maestro that helps enterprises manage and deploy AI agents at scale.

Here's why I'm paying attention: the stock has fallen roughly 85% from its 2021 peak. It went from over $85 to around $11-12. Yet revenue has grown significantly. They just hit their first-ever full-year profitable year. They have 80%+ gross margins. And they just announced a $500 million stock buyback.

An enterprise company that already has 10,000+ corporate customers, a governance framework that big companies trust, and a platform built for orchestrating AI agents... trading at a fraction of what the market valued it at during the hype cycle? If they execute the agentic pivot well, this could be a significant re-rating.

The Uncrowded Take: When everyone is talking about the obvious — chips, energy, the big AI names — that's the layer where it's already harder to make outsized returns. Still a safe bet? Probably. But if you have a higher risk tolerance and want higher potential upside... look at the layers underneath.

Who handles the traffic when agents flood the web? Who helps enterprises actually manage all these agents? Who captures the value that sits between the AI models and the businesses using them?

That's where I'm looking.

(Standard disclaimer: I'm not a financial advisor. Do your own research. These are small positions for a reason — I'm mentally prepared to be wrong. This is my personal thesis, not a recommendation. You've been warned.)

A Quick Note From Me:

You might have noticed something missing this week. No Partnership section. No Protocol section.

Here's why: there are some weeks when I want to go deeper into a subject instead of spreading myself thin across four topics. The AI world is moving so fast right now that I felt like you deserved a proper deep dive instead of a surface-level summary.

But this is something I'm still figuring out. This newsletter is still new, and I want it to serve you — not just me.

So I need your help. Reply to this email and tell me:

Hit reply. I read every single one.

See you next week,

Gabriel Judah

P.S. If you want to get smarter about AI with just 5 mins a day… check out today’s sponsor below (you’ll get smarter and supporting my newsletter, that’s a win win 😆 )

If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.

This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.

Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.