There is often conversation around the concept of flywheels and how powerful they can be. In fact, they are the models used for some of the largest growth in companies in history. We see this throughout the business world with companies such as Amazon and Disney being key players in this idea.

In the age of data, this is taking on another form. Most are now aware of the concept of network effects, something that is standard within the digital world. From here, companies such as Google, Amazon (again) and Meta have used to develop huge growth mechanisms.

Here is an example of this works for AI agents:

For Web 3.0, this is crucial for its success. It is a system that Web 2.0 perfected. However, the former is lagging a great deal.

The Flywheel Effect: Model For Massive Growth

Platforms that focus upon flywheels and get that going can see massive gains over time.

One company to learn from is Disney. The House of Mouse dates back to the 1950s. Walt Disney and crew understood this concept back then. Over the years, we can see the evolution of the company.

1957:

1967:

Today:

Source

Notice how we see massive advancement in complexity.

Disney is an interesting story since Bob Iger is really making some bad decisions. The companies theme park attendance is way down. Disney will lose over a billion dollars this year on its movies. The streaming platform remains a black hole.

Yet in spite of this, the company keeps chugging ahead. Why is that? The answer is the flywheel. Once set in motion, these do not break overnight. It takes a long time to built and the decline is also slow.

In other words, there is insulation. That does not mean, however, Disney is bulletproof. Over time, things can erode where the flywheel goes in the opposite direction. The network effects present in the digital world can reverse.

Flywheels In The Age Of Automation

Once the flywheels are in motion, the goal is to get them spinning faster. Properly constructed, we see how it all builds on top of itself.

Look at the diagram at the beginning of the article. The AI agent flywheel is reinforcing. This is why data is so important. It is the onset of the multiplication factor. It is something that those who simply look at things in a linear fashion overlook.

Google is another company that seems to make some boneheaded decisions. Going back a couple years, Zuckerberg did the same with the bet on the Metaverse. In both cases, the impact is minimal.

By now you should be able to guess why.

Both companies have enormous network effects that start with data. These are data generators from which all else can spring. Google has been downloading the Internet since the earliest days. It probably has more data than any other entity in the world.

As we head into the age of automation, this only accelerates. We know that companies are already incorporating AI agents into their systems. A prime example is chatbots.

When a model becomes popular, say ChatGPT, we see the reinforcing nature of the system take over. With an estimated 800 million users, people are doing prompts throughout the day. This leads to more data being generated. Both the user input, via conversations, and the synthetic data are then fed back into the training of ensuing models.

Of course, the rush is now to build other applications tied to this. When this occurs, we have the situation repeating itself. Now people are generating images, videos, or whatever the application produces, another dataset that is fed back in.

We can see how this easily moves to the world of robots. Since they contain a variety of camera and censors, it is logical to conclude the data will only skyrocket. After all, our eyes do not record what we see, at least not to an artificial database. With robots, the "eyes" are cameras, recording everything that passes before it.

Web 3.0 has to latch onto this concept. It is crucial for everyone to realize the impact this makes. It is why the stranglehold that Web 2.0 has is so powerful.

To break it, flywheels have to be set up elsewhere. This is the model that leads to massive growth.

Posted Using INLEO