🏗️🤖 Robopreneurs: A new type of startups in the age of AI
A bit of history on how startups have evolved.
In 1993, despite the internet promising to change the world, only 3% of people used it. It didn’t have a graphical interface, making it less user-friendly.
And one who complains when Netflix freezes…
Until Mosaic and then Netscape came out, the first browsers with graphical interfaces aimed at consumers.
For the older ones, the only photo you can listen to, haha.
This catapulted internet adoption, with a 1000% increase in browsing traffic just a month after its launch.
With the growth of the internet, entrepreneurs began to see its potential. However, building a website was difficult; you had to handle everything: software (writing code, which was more complex back then) and hardware (building the servers where the code runs).
Fortunately, alongside the entrepreneurs, investors eager to burn cash also appeared.
For this reason, the companies that began to emerge during the dot-com boom looked something like this:
Executive teams filled with MBAs
Initial investment of $10 million or more
A team of 20+ engineers from the start
Large and heavy teams that generally took one or two years to launch a product.
After the bubble burst and the MBAs fled the valley, there was a long period of “drought” for startups. The entry cost remained high, and investors' willingness to take risks was low.
But two milestones catalyzed a new surge in startups.
The first was the arrival of open-source programming frameworks like Ruby on Rails, which made programming easier. For non-technical people, a framework is a series of prefabricated code blocks that allow you to “assemble” web pages with less pain. With these new tools, you could have fewer developers generating applications faster.
On the other hand, “cloud” servers became a reality. Instead of needing engineers to build and maintain physical computers in your company to run the code, you could pay AWS or another provider to do it for you, at a fraction of the cost.
With these advances, “lean startups” were born. These teams were small (3 or 4 people) and mostly technical, capable of launching products faster and iterating with the help of their customers.
But “lean startups,” to grow, necessarily had to “bulk up”: hire more engineers, sales teams, marketing specialists, data analysts, UX designers, and a long etcetera.
All this was possible because investors returned to the scene, bringing venture capital financing to levels similar to those of the dot-com bubble.
If you have been following this newsletter, you already know where I’m going.
Just as open-source frameworks and the advent of cloud computing opened the door to a new type of startups, the same will happen with artificial intelligence.
This new technology will allow startups to be born with even smaller teams. But most importantly, it will allow them to remain small for longer, achieving enormous scales with lightweight teams.
Stories like Instagram (13 employees when it was bought for $1B) and WhatsApp (35 programmers supporting 450 million users when it was bought for $16B) will become more common. My prediction: we will see more unicorns with teams of fewer than 10 people.
## How?
AI will enable the automation of many processes that are key to scaling a startup, allowing it to do without several roles.
1\. Development
Most programmers are already using AI for their work or plan to do so.
conducted with around 90,000 programmers.
If you pair this statistic with the fact that this technology allows you to code 55% faster, according to some studies, then it’s clear we will see a considerable increase in productivity in terms of application creation.
Write one line of code, and AI autocompletes the remaining 16.
In simple terms: you will need fewer programmers and will be able to achieve more.
2\. Data Analysis
Any respectable startup has a monitoring and evaluation area (or its equivalent). The goal of this area is simple: collect data, clean it, organize it, and generate valuable information to guide product, marketing, and/or sales decisions.
Today, ChatGPT can be your data analyst. With the Code Interpreter plugin (which is still not open to everyone but has already shown huge potential), you can give it a database and ask questions in natural language.
There are also other tools like Akkio, with which you can train your own neural network to make predictions without needing to be a data scientist.
If you have 14 minutes in your day and are interested in data analytics, it’s worth watching
made by its CEO.
For more complex analyses, we will still need specialized people, but for 80% of tasks, perhaps the founding team itself, with the right tool, can handle it.
3\. Testing
To develop good applications, you must write good code. An important part of writing quality code is having automated tests that ensure everything is in order every time you make a change.
But testing takes time (and is boring).
Here too, AI can be a support that keeps teams “light,” with tools that can suggest and build tests when you need them.
4\. Designing Interfaces
Gone is the stereotype of applications made only by programmers: functional but horrible in design.
Products like Galileo or Diagram allow us to generate complete interfaces using only text.
Need a website to launch your product? The new AI from Framer creates it in seconds. Literally.
5\. Marketing
“Social media. 10x faster with AI” is the promise of Ocoya, which aims to facilitate the creation, automation, scheduling, and analytics of social media campaigns.
What previously required an external agency or dedicated internal teams can now largely be automated with AI.
6\. Sales
Sales robots on autopilot.
With tools like Apollo, you can search for anyone's contact:
With all the filters you can imagine.
Then generate an outreach sequence:
And let your sales robots generate your first meeting.
Disclaimer
This does not mean that startups will not need data analytics, design, marketing, or sales teams. When you reach a certain level of maturity, it is the right decision to bring in specialized people.
My thesis is this: startups that leverage AI will be able to start small (1 to 3 people), generate products quickly (days, not weeks or months), and remain small for longer as they scale.
We will see solid companies, with millions of users, created by a handful of employees and an army of robots.
