Platform Shifts in Tech: How Past Founders Rode Them

Every few decades a new base layer arrives and the rules for building a company change with it. Studying platform shifts in tech: how past founders rode them, gives you a way to read the AI moment without guessing. This post walks through four earlier shifts, pulls out the patterns, and ends with a short exercise you can run on your own idea.

What is a platform shift in tech?

A platform shift happens when a new foundation changes what is cheap, what is possible and where customers spend their time. Founders build on that foundation instead of building it themselves.

Three things usually change at once:

• A cost collapses. Something that needed a big team or a lot of money becomes a few lines of work.
• A new place to reach people appears. Customers gather on a new device, network or interface.
• Old advantages weaken. Skills and assets that protected incumbents matter less.

A new gadget is not a shift on its own. A shift needs all three. If you want the longer argument for why software keeps absorbing industries, read why software is eating the world, explained simply.

The PC era: how Microsoft and Apple rode the personal computer

Before the personal computer, computing meant large machines owned by institutions. The PC moved a computer onto a desk. The cost that collapsed was access.

Microsoft did not try to win by making the box. In 1980 IBM contracted it to provide the operating system for the IBM Personal Computer, and Microsoft kept ownership of that system (MS-DOS) when the machine shipped in 1981. As other companies made compatible PCs, the software layer they all needed became the valuable part. That is a pattern worth keeping: when many players build on a platform, the layer they all depend on can capture the most value.

Apple took the other route. It built the hardware and software together and sold a complete, easy to use machine. Two founders in the same shift chose opposite strategies, and both built lasting companies. The shift created room for more than one answer.

The internet wave: Amazon, Google and the founders who moved early

The internet made distribution nearly free for anything that could be described on a screen. Amazon began as an online bookstore. Books were a good first product because a catalog is easier to browse online than in any physical shop, and the product is simple to ship. The founders picked a narrow starting point that fit the new platform, then widened it.

Google started from a different angle. The web was growing and people could not find things on it. Search turned that confusion into a product. The founders did not invent the internet. They noticed the biggest new problem the internet created and solved it.

Both examples show the same move. Look at what the platform makes easy, then look at what it makes painful. Both can be openings.

The mobile shift: Instagram, Uber and apps built on the smartphone

The smartphone put a camera, a location sensor and an always-on connection in every pocket. Founders asked what only works when those three things come together.

Instagram is a clean example. It began as a photo-sharing service built for the phone camera, a product that made little sense before everyone carried a camera with a network connection. Uber used the phone's location and an app to match riders with drivers. Neither company needed to build a phone. They used built-in capabilities that were already in customers' hands.

The lesson is to name the new capabilities, not just the new device. Ask what the platform gives you for free that the last platform could not.

The cloud shift: how startups used AWS to launch with tiny teams

Cloud computing turned servers from something you bought upfront into something you rented by use. Amazon Web Services is a well-known provider, and it grew out of Amazon opening its platform to outside developers in 2002. A small team no longer needed a room of hardware and the money to buy it before the first customer arrived.

That changed who could start. The cost of getting to a first version fell, so more people tried, and more of them got to test their ideas against real customers. The advantage moved away from owning infrastructure and toward speed of learning.

This is the closest parallel to what many founders feel today with AI models. You can rent capability that once required a research lab. The question stops being "can I build this?" and becomes "should anyone pay for it?"

Patterns in platform shifts in tech: how past founders rode them

Put the four shifts side by side and the same habits repeat.

1. They built on the platform, not the platform itself. Instagram did not make phones. Early cloud startups did not build data centers. They spent their effort on the part customers noticed.
2. They started narrow. A bookstore, a photo filter, a ride in one city. A small, clear start gave them something to test.
3. They used a capability the old platform lacked. Location, camera, cheap servers, a global network.
4. They found the new pain. Each shift created problems that did not exist before, such as finding things on the web.
5. They thought about what could be copied. When everyone has the same platform, the product alone rarely protects you. Distribution, data, habits and trust matter more.

The fifth point deserves attention. If every competitor rents the same base layer, the base layer cannot be your edge. The Age of AI Entrepreneur lesson Building on Someone Else's Model, in the course Distribution and Moats, is about exactly this. The blog also has longer reads on distribution channels and on why thin wrappers around ChatGPT are hard to defend.

Riding the AI platform shift today

AI fits the three-part test described earlier. A cost has dropped: writing, summarizing, classifying and drafting take far less effort. A new interface has appeared: people can ask in plain language. And some old advantages look weaker. Whether the shift is as big as past ones is something researchers argue about, and it helps to read their cases yourself.

It also helps to see how the same ideas play out in other kinds of products. The post on evaluating an AI feature shows how to tell whether a product built on a rented model really works, and the post on unit economics for an AI product shows what that rented capability costs per customer.

A worked exercise: map a past shift onto your idea

Pick one earlier shift, say mobile, and one idea of yours. Then paste this into any AI assistant:

I am thinking about building [your idea]. Take the mobile shift as an example. First, list what became cheap, what new place to reach customers appeared, and which old advantages weakened. Second, do the same for AI. Third, tell me which part of my idea relies on a capability that anyone can rent, and which part would still be hard to copy in a year. Argue against my idea at the end.

Then check the answer against reality. Talk to a few people who have the problem. A model can organize your thinking, but it cannot tell you whether someone will pay. The blog covers testing demand with a landing page and a waitlist, which is a sensible next step.

The course Platform Shifts and the Founders Who Rode Them in Age of AI Entrepreneur goes deeper on this. One of its lessons is Why AI Is a Platform Shift and Software Eats the World. Each lesson is short, ends in questions, and lists its sources. Where researchers disagree, the courses set out the sides and leave the conclusion to you.

Frequently asked questions

What is an example of a platform shift in technology?

The move from desktop computers to smartphones is a common example. It changed what products were possible by adding a camera, location and a constant connection to every pocket, and it created companies like Instagram and Uber.

How long does a platform shift take to play out?

It is usually measured in years, not months. Early products tend to look rough, and the strongest companies often appear well after the platform first arrives. Exact timing differs for each shift, so treat any single number with caution.

Do startups or incumbents win platform shifts?

Both can. Startups often find the new use first because they have nothing old to protect. Incumbents sometimes keep an edge through customers, money and distribution. The past shifts above include examples of each.

Is AI really a new platform shift?

Researchers and founders disagree on how large it will be. It does meet the usual signs of a shift, such as a falling cost and a new interface. Reading the sources on each side is better than trusting any one forecast.

Get started with Age of AI Entrepreneur

If you want the history, the mechanics of modern AI and the business basics in short lessons, Age of AI Entrepreneur is a web app with 12 courses and 36 lessons of about 9 minutes each. You sign in with your email and an emailed code. Every lesson is free to anyone with an account, as set out on the terms page. Nothing in the courses promises a business outcome, but they give you a clear way to think about the shift you are building in.

Start Age of AI Entrepreneur: short hands-on lessons on how to build a business in the age of AI, with every side of the debate and the sources to read next.