August 2022

In my latest post, I shared that I was joining a new web3 venture.

Well… I decided to quit.

But in the process, I found out about one of my deepest fear.

Raising a monster

A few months in our web3 project, after talking to many actors in the industry and iterating over different ideas, I was getting triggered repetitively.

• The people we were talking to didn’t seem to really enjoy or be passionate about what they were doing. They were clearly “bullshitting” their way into financial success.
• Every people in web3 admits it, they joined to get rich, period.
• Every project I was looking into, had for only purpose to generate money. No actual value added. I was at first seduced by the DAOs and Desci projects, but I quickly discovered that they were just elegant facades hiding yet another shady money-making scheme.
• My co-founder started positioning us as the Criteo of Web3 (I did an internship there). We would use on-chain financial data to allow things that ecom could have never imagined (ie. getting your whole bank history as soon as you login on any ecom website…)

Needless to say, I started to freak out.

Doubt started cramping. Am I going to build the Facebook of web3? (read Facebook as in all the privacy concerns, not the business “success”).

In the startup world, the relationship the founders have with their company is often compared to a father and their baby.

That sounds cute. But not all babies are… You can definitely raise a monster.

I was projecting myself building a privacy concerning company, but a succesful one. One that just has too much momentum and financial stake (investors money, employees’ careers) to bury it down.

What would I do if it happens?

Digging up a deeper fear

For months, I kept thinking about it. And the more I looked at it, the clearer it was that this project wouldn’t end up well, at least in my book.

No matter what the outcome is, this will be a failure. Worst. The best outcome is actually that we fail as soon as possible, before we do anything irreversible.

So I decided to quit before it’s too late.

Figuring out what I want to do have been really tough. This question has been killing me for the past year or so. And I still have no answer.

But today, I recognize my number 1 blocker. My fear of raising a monster.

I have ideas. I’m positive that I can pull them off, more often than not.

What I fear, is that I indeed execute properly, but end up with a terrible business to run that I can’t escape because the stakes are just too high.

Beautiful idea or terrible mistake?

Lemme give you an example.

Lately, I’ve seen many succesful businesses using the productized service model.

You take a freelancing service - let’s say design - and you package it in a way that can be sold as a product (subscription model) - eg. DesignJoy.co.

Since February, I’ve been working with a startup as DevRel (marketing for developers). I got pretty good at it, or at least I know what it means. I say that because this job is fairly new and misunderstood. Plus, it’s quite rare to find people that are both good developers and good at creating content.

So I immediately thought of a productized service business around DevRel-as-a-service. Something like get 1 blog post per week and a weekly (actionable) performance report for 2k€/month.

I can find a few customers (early-stage startups) via cold email. Once I’ve landed a few clients, I start delegating the writing to freelancers found online and grow the business from there (social media and SEO to acquire new customers, processes to make myself dispensable, etc).

I bet I can grow this business and sell it for 1M€ within a year or two.

Caveat: it’s probably going to be a nightmare to run.

I’ve painted the best case scenario. The most likely scenario is this:

1. I quickly reach 3-4 clients 🎉
2. I can’t keep up with the pace (1 blog post per day, 4 reports per week) and hiring dev copywriters is tough. Plus, I don’t have enough time to focus on that part
3. I start having trouble with some customers (either quality is degrading or they are just 🍑 - it happens more than you might think)
4. I finally manage to hire some freelancers to delegate the writing, but they only stay for a few months so I constantly have to find new people and quality varies a ton so I end up rewriting 80% of the content

END: shit hit the fans on all sides. Clients don’t want to pay even though they published the content. Freelancers want their paycheck even though I didn’t use their 💩 content. But I can’t afford to get out of it. I don’t have time nor money.

Now that’s what I call a monster!

What’s the way out?

Honestly, I have no idea.

But I keep coming back to the idea I had at the beginning of the year, some type of R&D AI Lab.

I still think it’s great but I just failed at being more specific in describing it. Even for myself it was too blur.

The idea is highly inspired by Tiago Forte’s The Rise of the Fullstack Freelancer.

The concept is simple. With all the new tools, freelancers can afford to broaden their area of action. And despite the popular advice of focusing on a single thing, having a portfolio of activities can actually lead to better results.

Let’s take the example of a leadership coach. In the classic paradigm, they would focus on a single thing, running 1:1 coaching services.

This has many drawbacks:

• how are they going to acquire new customers
• their model is tightly linked to their time spent → not scalable
• the lack of variety can quickly become dreadful and lead to burnout

But in the fullstack freelancer world, they can solve all these problems by diversifying their activities:

• workshops and talks for branding and building authority
• books and courses for passive income streams
• podcasts and newsletters for another passive income (affiliates and ads) + acquiring new customers + branding + loyalty
• consulting/coaching/corporate training.

The great thing about this model is that each activity actually reinforces one another. Publishing a book leads to more workshops and talks. And podcasts bring more course students and book sales.

Back to my AI Lab idea

The advice I’ve seen most often when young people ask “successful” people what they should do is:

1. build an audience — it will serve you regardless of what you end up doing
2. build stuff — it will flex some good habits (shipping, action bias), teach you some new skills and let you explore different rabbit holes.

I guess the AI lab is not really an official structure, but rather guardrails to keep me in my zone of genius:

• AI research
• product development
• tech for good (health/wellness, climate, city/transportation)

It’s also a way to force me into thinking about better ways to do all of this because I hate how AI research is done (both public and private) and the developer experience is still very hostile.

How did I fail

I tried doing it over the last 4 months and completely failed.

What happened is that I was starting new projects every single week without finishing the previous ones. I would question every decision without having a frame of reference, because I didn’t really know what I was trying to accomplish anyway (some longterm vision I guess?).

The second reason, I believe, is that I just got too comfortable.

I often state that I want to do “tech for good”. Use data - and eventually AI - to solve real-world problems in health/wellness, climate or city/transportation.

So in practice, I should suffer my way through learning more about these topics, exchanging with experts and immerse myself over a long period of time.

Elon’s cousin reportedly spent a full year immersing himself in the solar energy industry - devouring niche books and attending every conferences - before building SolarCity.

But I haven’t. I just stick to simple, superficial ideas because they don’t require much thinking or context knowledge. They are comforting.

How can I do it differently

I’ve reconciled myself with the idea of having a regular 9-5 job. I’m actually looking forward to work with like-minded people (finding my tribe so to speak).

My goal is not to build a business. It’s to have more mature mental models of the world (esp. areas I’m interested in) so that I don’t raise monsters if I end up building a company.

Tips on building great companies by Tobi - CEO of Shopify:

Read a lot of books, build a great mental model of the world, build a prototype of something that would be neat to add to it, inspire other people to join the building it, and then just don’t stop.

So my approach is not to build a business, but a body of work around data/AI for good.

I want to rethink:

• how research is done in the context of R&D to solve data-intense problems
• how solutions are designed around AI (think resources, maintainability, robustness, fairness, privacy, etc)
• how data and AI can become a tool in tech-for-good (without being a hammer looking for a nail)

TK - what’s the plan practically speaking? do I need to define goals or “success”?