A short introduction to Fully Autonomous Ventures

Steve Jobs, in memory of him, I would like to start the introduction of what I am dedicating my life to. Apple's?Think Different?campaign is one of the most important videos of my life. Every time I felt at my lowest or had a setback, I watched his video about the round pegs in the square holes, about the crazy ones.

Jobs mentions that the people who are crazy enough to think they can change the world, are the ones who actually do it. From a very young age, I wondered how I could be one of those crazy ones. How could I make sure to get the most out of my life, and how I could become the best version of myself? In short, how would I be able to make an impact?

About a decade ago, I discovered my love for computers, and it didn’t take long before I discovered all the ins and outs. Two years later, I started my first company, followed by my second and third ones a few years after. A decade later, I have helped more than 30 tech start-ups find a solution to their problem, prototype the first version, and scale.

I was lucky to see first-hand how technology enabled people to do more, in comparatively less time with better results. I had front-row seats to how several technology start-ups succeeded, failed, stood up again, and finally reached business success. From the start, I saw the power of data gathering and data analysis, and I started to gather data like a maniac from the start-ups I started, the ones I followed, and from other open source datasets such as Crunchbase or by scraping millions of websites.

Now, a decade later, I discovered, together with some of my most trustworthy friends, that there are patterns in the building of these companies that showed me what I had failed to see before. I believe that building a company is more math than we previously considered. Of course, our insights and assumptions are still very early, but we believe that (at least in theory) we can lay a standardised foundation upon which a company could start, operate and achieve success.

We believe that within a few years, we will be able to create a machine learning algorithm that can predict the outcome of a task (like a chess-playing algorithm). This framework will be made possible with the help of our talented team, which has a lot of knowledge in the field of Machine Learning, our ability to experiment (with and for our partners) and the vast amounts of data we have access to.

Once we reach this phase, we can go a step further. If we can predict the outcome of each chess move, we would be able to create a machine learning framework that would be able to start, launch, and above all, manage a company. A company where theoretically, nobody is working anymore.

It sounds like an idea for Star Wars, however, we believe we are closer to it than ever before. In the following few sections I’ll explain why I believe we should do this and why now. But first, I’ll go more in-depth into what a?Fully Autonomous Venture?would look like.

The idea / The concept

When building a start-up we see three major departments/functionalities. There is also an Other department, in which we put all the tickets or tasks which we can’t automate directly.

The blueprint of a Fully Autonomous Ventures

I often make the following drawings when introducing Fully Autonomous Ventures (FAVs).

So, first of all, we have the?Product department?responsible for everything product or service related, from building, maintaining, and improving it. In short, this is the team that?ensures there is a product that people can use.

Next, we have the?Marketing department?responsible for?finding / attracting customers. This team handles everything from content marketing to sales and everything in between.

Thirdly, we have the brains of the company, which I often refer to as the?Intelligence department,?responsible for?determining which resources to invest in which team?(e.g. how much of our profit should go into product improvement and how much into marketing?).

Besides those three teams, we have the?Other department?within an FAV. This part includes everything that we couldn’t add to one of the three previous teams (and are often things we cannot automate entirely (yet)). One example of this is Customer Support.

Automate what we can’t automate

Several things within a company can't be automated (yet). Our technology still has plenty of limitations. But before we start telling people that things are impossible, we should discuss the input and output. So before going more in-depth into Fully Autonomous Ventures, we will explain what automation is and what the concept of a Fully Autonomous Venture entails.

First of all, automation implies that the owner of a process no longer has to do a task manually. A basic form of automation would be delegation, where someone else completes a task. This person could be more specialised at this type of task. Automation in this sense does not mean that there is no longer a person involved.

Many processes and start-ups focused on?automation?still rely on a team to perform certain tasks. One such example is the Amazon Mechanical Turk. This service (from Amazon Web Services) allows you to integrate with their API (a way to link two programs to each other) and send over tasks that actual humans need to perform. Mechanical Turk has over 500.000 workers, available to you to complete any task, right through their API.

One example of a request you can send to Mechanical Turk would be to label an image (e.g. a company focused on creating self-driving cars that wants to label thousands of images of traffic signs).

This kind of automation enables you to automate everything that you can’t automate yet (at least in theory). One example we often get is how we would be able to do email outreach, handle customer support or guarantee the quality of our social media content. We can handle this with software like Mechanical Turk (of which there are hundreds of alternative companies).

There are plenty of solutions/ideas to guarantee the quality of the outcome of solutions like Mechanical Turk. How does the self-driving car company I mentioned earlier know that their labelling is correct? They validate this by using the?rule of three. Here two people are asked to label the same image, and if they are labelled the same, you conclude that the label is correct (the more people you ask to verify, the better the quality, but the higher the cost of labelling). In case both people give a different label, a third person is asked to validate. In that case, you also assign scores to the labellers so that every person builds a personal trust score (which could, for example, influence the income they get for each task).

People in a FAV

That said, there will still be plenty of people working at FAVs, however not directly but indirectly. The intelligence part of an FAV, which is responsible for resource planning, will calculate the budget it can make available for a certain task. It will then post that task on a marketplace for freelancers (such as Mechanical Turk), choose the freelancers, and pay them once their work has been uploaded and validated.

FAVs will not disrupt jobs. I believe they will create more of them, more specialised ones (as FAVs will take on the repetitive tasks, allowing people to focus on their talents). FAVs operate on a clear set of rules (e.g. the algorithm is a mathematical formula, so it would rationally pick the most suitable candidate to execute the job, no external influence, no emotions. Just clear rules where the best person for the job will get it.

The limitations

Every time I give a keynote or guest lecture, people challenge me by giving me tasks which are harder to categorise. So keep in mind that we are just setting sail. You could compare us to Columbus leaving the European continent, we are not yet far, but we are looking forward to discovering a new promised land.

In the book?Sapiens: A brief history of humankind?by Yuval Noah Harari, the author mentions that the Europeans in the 16th and 17th centuries (the ones who catapulted our world into the industrial revolution) were not bigger, brighter or better than the rest of the world (who were probably more advanced in technology, were smarter, owned more resources, etc.). But the one thing that the Europeans did was they realised one simple fact, the fact that they knew they didn’t know anything.

Together with a team of enthusiasts, we are researching the possibilities of FAVs, the limitations of Machine Learning, and the future of our species. Whether Fully Autonomous Ventures are possible (or not), we have stretched our sails, waved the public on the quay goodbye and have set sail to discover a new world. The journey will be hard, but the rewards will be worth it.

Why?

A few years ago, we had an application meeting with Michael Seibel, the CEO of Y-combinator (the Olympics for start-up founders where almost all major start-ups and technology companies originate). He asked only one simple question, "Why?”. Every answer we gave was met with a follow-up question “Why?”.

So when the idea of FAVs came up, we asked ourselves, why? Why would we do this? Why would this improve humanity? Why now? Why was this worth following? What follows are a few of the notes I wrote down on the papers when sketching out the first drafts about FAVs.

Are humans the most efficient solution?

One question I ask myself quite a lot while building is: if solution X is the most efficient solution for my problem? Does it solve what is required? Is it the cheapest solution? Will I keep using it in the future, etc.? I create a lot of categories and score each solution towards the problem.

If we would do this for the question of what would be the best solution to start/run and manage companies? Would humans come out as the best solution? I figure we would score pretty low on plenty of categories such as:

• To take the best rational solution (no emotions, no external influences)
• Can analyse all the data points (e.g. billions of data points) to make the best decision
• Is available 24/7 to make decisions, respond to the clients,…
• Work enjoyment?
• …

A computer would do all of the above-mentioned things a lot better than humans. Of course, some categories would favour a human over the computer, but those are quite limited. Some notes we wrote down when brainstorming about how efficient it would be to replace our entrepreneurs with computers are:

• Emotions aren’t bad, we need also emotional decisions in business. But do we?
• Can computers be creative? Well yes, they can.

Is this what we want to do with our time?

Another question I often ask myself is how I see my life. Well, I would like to have fun, have an impact, spend time with family, with my children, travel, and just enjoy life. What I don't see myself doing is working 9 to 5, pleasing people I don’t really care about, getting yelled at because I didn’t do a good job, etc.

This results in my most important question, why do we spend so much of our precious time on something we don’t like? To earn money that we can spend with the limited time we have left? We spend almost 1/3 of our life working, and 1/3 sleeping which leaves only 1/3 of the rest of our time to do the things we really want to do.

Building FAVs would drastically change our society, potentially introduce a universal basic income, liberate people from repetitive tasks, and allow them to do what they sincerely love (and are probably better at).

The solution to world problems?

Plenty of people I talk to tell me that they want to dedicate their lives to solving world problems like climate change by organising clean-up actions and solving social isolation by visiting older family members that they currently can’t visit because they have to work all the time. Other people have great ideas on how we can fight poverty, but they don’t have the time or financial resources to do so. And some might want to work with NASA or SpaceX on becoming an interplanetary species, but they don’t have the skills to be hired and (/or) not the time and financial resources to retrain.

A personal example of this is my father. He has an incredible eye for scouting potential pro cyclists (granted, not a big world problem, but something that can have a good influence on a lot of people). On Saturday, he organises training sessions for small children (from 5 to 14 years). It is his passion, but not his work. He spends all his spare time on this when he is not working. What would happen if he didn’t have to worry about money anymore? Or increase the time he can work on this? The result would be net-positive, even financially, as more and more people would discover the cycling sport, we would find more talent and maybe even snatch up some medals in Olympic or International championships (Yes, already several European/World Champions have been scouted/trained initially by him).

I believe that humanity will flourish and we will transcend ourselves when we can release ourselves from the burden that is “work”. First, we had the agricultural revolution, followed by the industrial revolution, and the digital revolution, all of which have radically changed humanity and society. Next, I believe we will have a work revolution, liberating us from the duty of working.

Why now?

One of the other Why's, I keep pondering on is, why now? In the past few years, a lot has happened that has brought us closer and closer to the age of autonomous ventures. Let me share a few examples of them.

Data is the new gold

The first thing I want to highlight is the amount of data available. Over the last couple of years, the data influx has grown exponentially, there is just so much data that is created and published online, which can help us build better and better companies.

In 2015, every minute, over 60 hours worth of video content was uploaded to Youtube. Today, that number stands at over 500 hours worth of content. There are more start-up directories, and more blog posts or tweets from founders outlining what went wrong and what helped them succeed published every month. New data initiatives rise and older ones get more promotion. New technologies like Selenium and Beautiful Soup are released or improved. The cost of servers and storage is cheaper than ever before.

We have datasets of tens of millions of start-ups, of which, on a weekly basis, we scan their social media accounts. We have also indexed over 120 million websites and extracted their content, resulting in billions of web pages and texts. Add to this the tens of millions of blog posts that we index monthly and over 100 million newspaper articles and you get multiple Terrabytes of raw data waiting to be transformed into something meaningful. This data is all stored on hard drives, transferred, analysed, and used for only a few thousand euros each month.

Technological revolution

This data combined with all the new algorithms and machine learning frameworks that have seen the light of day in the past decade puts us in an extraordinary position. I still remember when I started that I wanted to create a simple machine learning algorithm that could categorise news articles and extract entities from them, it was incredibly hard and I ended up with the most basic algorithm ever. Whereas a few weeks ago I started entirely from scratch and in a matter of days, I had an algorithm ten times better than the one a few years ago.

Machine learning-focused start-ups like OpenAI or even Facebook (Meta), Google (Alphabet), and Nvidia are releasing new models and frameworks at a faster pace than ever before. It is hard, even for me, to follow the latest trends. One such example is the GPT-3 model from Open AI. This machine learning model can write human-like texts (maybe this article was created by that algorithm?), which look so good that you can’t differentiate them from human writing, especially if you combine GPT-3 with other software and models like Grammarly or Translation services, etc.

Besides the deep-tech innovation, the amount of low-code and no-code solutions (where you can create software without programming knowledge) have popped up like mushrooms out of the ground.

The conclusion is that today everyone can create a mobile app, a website or a machine learning model, at a fraction of the cost and knowledge it would have taken you years ago. And even back then, the cost was already mentioned as an argument as to why you should start a company.

I believe that if you are curious, driven, and ambitious, you can build anything you want in a fraction of the time, knowledge, and cost it would have cost you before. What is keeping you from doing it?

The war on talent

If passion and drive overtake talent, the war on the latter becomes harder and harder. This could be viewed as something negative (it is harder for a company like ours to find the right people), but it also opens up the opportunity for Fully Autonomous Ventures to compete with other companies that can’t attract the right talent to beat the computer.

Add to this that you are better off investing in more automation & technologies instead of finding the right talent and boring them with repetitive tasks (and probably losing them). So FAVs are the answer and maybe the solution for the war on talent.

Work from home culture

Despite the many negative outcomes, Covid-19 also brought us some small positive advantages. I remember all our customers asking our advice on how to move towards a global/remote (work from home) culture. We had already applied work-from-home on several occasions and now we helped other companies (which in the past thought they were unable to do it) implement it.

This allowed plenty of people to work from home, and they were encouraged to use software to communicate and do their job (from home). The use of video calls increased (and it hasn’t dropped yet), in combination with the rise of new software for internal communication e.g. Slack / Front / Microsoft Teams etc.

People were forced to learn how to work with new tools, and adapt to them and use them. My mom, for example, now uses Slack to communicate with her bagpipe music group, could you ever have imagined that a few years ago? Well, I certainly did not.

Generation X Y Z

Every decade we get a new generation, as a millennial, I grew up when technology was still in full development. I remember the time 3G was invented, followed by 4G. I also remember the time when cellphones or computers were not yet fully adopted by the masses (I even remember playing one of the first mass-adopted video games). However, the generation after me grew up with technology, and the level of technology adoption from a young age will only increase. This allows us to become more digital and it becomes easier to automate things than ever before.

Besides that, our entire world has changed. My friends back then still wanted to become pilots or nurses. Nowadays, the children I speak with want to become influencers, they want to play around on social media all day long. I often hear parents and older people complain that the new generation is scared of rolling up their sleeves, they are so used to automation that they are addicted to it (which is a bad thing on one hand, but on the other, makes automation unavoidable).

This trend also applies to people from Generation X and Y, which results in?the great reset?where more and more people are leaving their jobs to start their own companies and / or to spend more time on things other than work.

Isn’t this a great opportunity for Fully Autonomous Ventures?

Financial opportunities

As mentioned earlier, the price to start a company has decreased over the past few years. You can now start a company with next to nothing (I started my first one with just 150 euros and some money that I borrowed from my dad). But even that is no longer required.

You can have a website, forms, customer onboarding, marketing, and a customer support team, all for next to nothing nowadays.

Besides that, the price to write and host software is cheaper than ever before. We now train a lot of our models and analyse over a billion data points every month on a “machine learning” computer we once bought for about 3K and uses around 25 euros of power every month. Cheap, right? So with a few thousand dollars in our bank account, we can analyse and train an unimaginable amount of data.

Financial recession

One more reason why the age of Fully Autonomous Ventures is here is because of the financial recession that everyone is talking about. We don’t know if it would happen and if it happens, we don’t know when.

But one of the reasons Google had so much success was the fact that they started (with an impressive idea) at the right time. Because of the dot Com bubble, they could hire the brightest people, rent the best offices, and start in the perfect position to become the massive technology company they are today.

The limitations

Before we end this blog post with a new conclusion, I would also like to tell you a bit about the limitations we foresee.

First of all, computers lack some creativity, all their decisions come from a kind of mathematical algorithm. We can debate how a human gets their creativity or that everything is based upon something else (as in that case computers are as creative as humans). Some machine learning models (for example, DALL-E from OpenAI) are examples that computers can be creative, but for now, I want to keep it a debate if or not a computer can keep up with new ideas (as I believe it can).

Another limitation I see is that computers rely on humans for the next few decades to survive. They need humans to operate and oversee certain things (e.g. power adapters) and they can’t execute certain tasks (e.g. unload themselves). I believe it is just a matter of time before computers can help each other out, just like we depend on our mothers during the first few years.

There will always be certain tasks that will be cheaper and better than if a human does them. For example, validating a text, writing automated answers to questions from customers, etc. We will automate most of these tasks in the coming years, but for now, we will still need humans and I believe we will always need humans. However, humans will become more specialised in a given task or direction.

Data is cheaper and easier to gather than ever before, but I still believe that we need more and more data before we can build amazing Fully Autonomous Ventures. The first ventures we are currently launching are still only experiments or prototypes and they have only one goal, to gather more data, to give us more insights, and to help us develop better solutions.

One of the reasons for this is that I believe that 7 out of 10 FAVs will fail, 2 will work but only reimburse the initial start-up cost and 1 will truly succeed. These are tough numbers if a human has to do it, negligible if a computer can start hundreds of even thousands of ventures every year.

When will we see these use cases?

Now that we have established the basis of what an FAV is, what we are working on, and why it will or will not succeed, we can start going into more detailed information. One question I often get is when will I see these kinds of companies?

First of all, they are already here. Some start-ups already apply this level of automation and even our team has some first real-world experience and existing prototypes. So in short, the answer is now.

The longer answer is a little bit more complex. First of all, we need a lot more resources, data, time, and talent. The initial prototypes we have built are very experimental and they still have a lot of flaws. Each of these will now be closely examined and we will improve them step by step until we have a true FAV. This process will take a few years. So I always tell people we will see the first truly automated FAVs in 2 to 3 years.

What is the next step?

To round things off, I would like to give you a short overview of what is up next. First of all, we’ve founded a start-up called WebFaster, which is a research lab for Fully Autonomous Ventures. With WebFaster we research the possibilities and the limitations of these FAVs. We do this by creating prototypes, and MVPs, partnering with other companies and testing how we can automate certain processes for them, etc.

This allows us to gather more and more data about launching start-ups, why certain processes and ideas fail, and why others work. The data and insights are bundled into reports and machine learning frameworks, which help us to start a new generation of FAVs built on the lessons learned from previous experiments. We plan to continue to do this until we are satisfied enough that we have ample resources to help fund other organisations to help solve their major challenges.

Currently, we are fully occupied with a few of these experiments (4 to be specific), besides the implementation and the education at other companies. Want to find out more? Get in touch with our team or follow our socials to stay updated.

Want to partner up?

Do you believe FAVs are something for you as a company or as an individual? We have several options to work with us.

We have a?partner program for companies?who want to implement more automation and become more data-driven. More information can be found?here.

We are also always looking for great, talented people to help us on our journey. For those who believe they can help us with this, you can find more information?here.

If you want to keep up to date with our latest insights and news, I suggest you subscribe to our newsletter, which will keep you up-to-date on our latest research and insights.

If you loved the blog post, please share it with your friends, family and colleagues, so we can keep writing these amazing blog posts!

PS: we are writing a book about Fully Autonomous Ventures and the future of companies. Curious to read our first draft? Let us know! We are always looking for reviewers!