For the last couple of weeks I haven’t been sleeping much.

And no, it’s not stress. Stress is always there anyway. Fundraising. Running the business. Keeping up with what is happening in the world. Founders live in this permanent background pressure. There is always something that needs attention. Something that breaks. Something that needs to be fixed yesterday.

But this time it is different.

This time something else is keeping me awake.

Something that many of you have probably seen all over social media in the last few months. Something that people discuss at events, in founder group chats, in random late-night voice notes.

AI agents.

And more specifically, OpenClaw.

See below to continue…

​Experience London’s first private Revivd tasting.
​First-ever frozen-to-sashimi experience.
​Immerse yourself in the art of post-freezing sashimi at an intimate, invitation-only gathering for chefs, culinary enthusiasts, industry insiders, foodtech investors, and discerning guests seeking something truly unique.

​Join Jihyun Lee, the founder of Revivd, to explore flavour, texture and a modern ritual in a setting designed for quiet indulgence and discovery.

​When: Monday 23 March 18:30 - 20:00
Where: Soho Works 180 Strand

​Spots are limited to keep the conversation intentionally intimate and high-value.

Request yours now.

Continue…

OpenClaw is an open-source AI agent framework created by Austrian software engineer Peter Steinberger. It didn’t come from a giant research lab. It started as a practical tool. A way for one person to interact with AI in a more useful way.

Instead of asking questions and getting text back, the idea was simple. Let AI actually do things.

The project existed quietly for months. It went through different versions and names. Then suddenly, at the beginning of 2026, everything changed.

The usage exploded.

Developers started building real workflows.

Founders started experimenting publicly.

Content creators started showing insane demos every day.

Within weeks, the creator became one of the most talked-about people in the AI ecosystem. He received strong interest from major players like OpenAI and Meta and eventually joined OpenAI to work on next-generation agent systems.

There were even rumours online that the total offer package to bring him in could be worth close to $1 billion. Nothing officially confirmed. But the fact that such numbers were even discussed tells you how strategic this space has suddenly become.

And once again, we see a familiar story.

European innovation. American scale 🫠 .

Full transparency. This space is new to me.

I had founders in our community talking about agents for months. I saw posts. I saw excitement. I saw early experiments. But I simply didn’t have time to properly look into it.

Running a real business does not give you space to explore every new technology that appears. You are constantly reacting. Delivering. Selling. Managing. Hiring. Solving problems.

Then OpenClaw started appearing everywhere.

People were posting about automating entire workflows. Replacing operational roles. Running AI setups 24/7 on small local machines.

At that point I realised I had no choice. I had to understand what this is and how real the potential actually is.

For the last couple of weeks I have been watching videos, reading tutorials, testing tools, breaking setups, starting again. Trying to separate hype from reality.

And I can tell you one thing.

It is both exciting and terrifying.

The uncomfortable realisation about our own business

When I started mapping our workflows at Rare Founders against what AI agents can already do today, the result was uncomfortable.

A massive part of our work is operational.

When you honestly break down how much time goes into these tasks, you start seeing a pattern.

80 to 90 percent of repetitive work is technically automatable.

Not perfectly.

Not instantly.

Not without effort.

But clearly automatable.

Recently I had to let go of a virtual assistant who was genuinely an amazing person. Hard-working. Positive. Loyal.

But I develop fast. I learn every single day. My expectations kept increasing. Eventually it became unfair for him to operate under that level of pressure.

Around the same time I started experimenting with agents. And I realised that almost everything he was doing could be handled by AI systems that do not get tired, do not forget instructions, and do not make emotional mistakes.

This was not a comfortable moment.

But it was a very honest one.

Social media today is full of stories about founders firing entire teams and replacing them with clusters of Mac Minis running automation 24/7.

A lot of this is hype.

A lot of this is pure bullshit.

The potential is real. The simplicity is not.

I am not technical. To even get basic workflows running I had to use a mix of tools. ChatGPT. Gemini. Perplexity. Forums. YouTube. Documentation. Trial and error.

I bought new machines to run local models. Some things worked. Many things failed. Eventually I went back to paid cloud models because reliability matters more than ideology when you are running real operations.

But even with all this friction, once you get one useful workflow running, you immediately see ten more opportunities.

It becomes addictive.

It becomes obvious that this is not just another productivity tool.

This is a new layer of how businesses will operate.

The real fear is not what agents can do today.

The real fear is how quickly they are improving.

Right now you still need curiosity and patience to make them useful. You need to be willing to look stupid. You need to accept that many things will not work on the first attempt.

But in one or two years this will be packaged into simple products. Clean interfaces. One-click setups. Plug-and-play automation.

When that happens, huge amounts of administrative and coordination work will disappear.

These are real jobs. Real livelihoods. Real stability for many people.

This transition will affect the economy. It will affect founders. It will affect society as a whole.

Sometimes I genuinely feel that as humanity we are moving faster than our ability to adapt.

But pretending this is not happening will not slow it down.

As founders we sit at the front line of change.

We experiment first.

We benefit first.

We also get disrupted first.

If you are reading this and thinking this is not relevant for your business, I strongly believe you are wrong. More likely you simply haven’t explored the real possibilities yet.

Your use cases will be different from mine. But they exist.

If you delay learning because of fear, privacy concerns, or noise on social media, you are limiting your own growth and competitiveness.

The train is already moving.

I am still early in this journey. Still testing. Still failing. Still learning every day.

But I am convinced that investing time into AI agents now will multiply everything we do at Rare Founders.

Operations will move faster.

Outreach will scale.

Content will be more consistent.

Data will become more actionable.

And once I reach a point where the progress is real and measurable, I will openly share how we are using agents across our database, community systems, events and investor workflows.

For now my message is simple.

Start experimenting.

Start learning.

Start getting comfortable with discomfort.

Because when this becomes easy, the gap between those who prepared and those who ignored it will be brutal.

And when everything reshuffles, you don’t want to find yourself at the bottom of the pile.

✅ If this resonated with you, share it with someone in your network who is building right now. These shifts are easier to navigate when we learn together.

(👉 Vote now — we’ll share the results in next week’s issue. All votes are anonymous.)

(Last week’s poll results)