🎈 The app boom
An estimate that circulates among people who follow the software market says that about two-thirds of the apps that exist today were created in the last three or four years. The exact number matters less than the direction: the amount of software has grown much faster than the number of people willing to use it.
💸 Inflation, in the economic sense
In economics, inflation is when there’s too much money in circulation and each dollar is worth less. Something similar is happening with apps: there’s too much software competing for the same attention, and each individual launch is worth less to the people who see it.
This doesn’t mean software has lost its value. It means generic software has lost its value. What still matters is software that solves a real pain point in a way that’s hard to copy and reaches the right people.
✓ What got easier
- ✓Go from idea to working prototype in hours.
- ✓Test multiple versions without hiring anyone.
- ✓Build on your own what used to take a team.
✗ What got harder
- ✗Get noticed among thousands of similar launches.
- ✗Convince someone to pay for one more subscription.
- ✗Keep an advantage when anyone can copy you in days.
💡 Practical takeaway
Before opening your editor, ask the question inflation forces you to ask: if ten more apps just like mine appear tomorrow, why would anyone still choose mine? If you don’t have an answer, the problem isn’t technical.
⚰️ Is SaaS dead?
Many people keep saying SaaS is dead. For the most part, they’re right: simple tools that only organized data or generated text are being replaced by a two-line request to an AI assistant. Why pay R$ 50 a month for something the chat does for free?
🧭 The three places where there’s still room
- •The right niche: underserved groups with money and little time to learn AI.
- •The right use case: a narrow, painful, frequent task instead of a generic platform.
- •Speed: get there first and learn faster than others can copy you.
⚠️ Common trap
Read "SaaS is dead" and give up, or read "there's still room" and build just anything. Both takes are lazy. What's changed is the bar: fewer ideas make it through, but the ones that do face less serious competition.
🔀 The product cycle got scrambled
The traditional product roadmap has seven steps: come up with an idea, validate it, prototype, market it, develop it, launch it, and improve it. This was a logical sequence when development was expensive: you only spent money on engineers after you were sure it was worth it.
With AI, building a prototype takes an afternoon. That changes the logic: sometimes it makes more sense to prototype to validate than to validate and then prototype. A working prototype answers questions an opinion survey never can, such as "does the person actually click?".
How the sequence usually looks today
Idea grounded in a pain point
Ideation is still at the center. Nothing replaces understanding the problem.
Rapid prototype
An ugly but usable version, built in hours with a coding assistant.
Validate with the prototype in hand
You show real people something real and measure what they do, not what they say.
Distribution from day one
Marketing is no longer the final step. Without an audience, there's no way to validate anything.
Develop, launch, and improve in short cycles
The final steps become an ongoing cycle instead of a straight line.
💡 Simple rule
The right sequence depends on the product and niche. Ask which step is cheapest to do right now and which one answers the most dangerous question. Start with the one that does both.
🎓 When experience is already validation
If you're a consultant, freelancer, or have worked in a field for years, you've probably noticed a recurring gap: every client needs the same little thing, and nobody offers it simply. When that's the case, you don't need as much validation. Your experience has already been the market research.
🧮 How the customer does the math
The customer compares your price with the alternative. If the alternative is hiring a developer for a few weeks, or paying for a premium AI subscription that costs hundreds of dollars a month and still requires them to learn how to use it, a product that costs R$ 49 and solves exactly that problem seems cheap.
The price isn't compared with zero. It's compared with the cost and effort of solving the problem another way.
✓ Signs your experience counts as validation
- ✓You've done this for five or more different clients.
- ✓Clients asked you to do it again or referred others.
- ✓You charge for it today, even if you do it manually.
✗ Signs you still need to validate
- ✗You've only solved it for yourself.
- ✗Every client asked for a different version.
- ✗Nobody would pay unless you were there to help.
🧑🚀 One person, three roles
At companies, products are usually built by people in separate roles. The product manager decides what to build and for whom. The engineer builds it. QA tests it and finds bugs. Each role pushes back on the others, and that friction prevents many mistakes.
🕳️ The role that gets skipped most
People who know how to code tend to jump straight into engineering because it's the fun part. AI agents make this worse: building becomes so easy that nobody stops to ask whether they should.
That's why most of the rest of this course is about the product role: the questions a good product manager would ask before writing a line of code.
💡 Wear one hat at a time
Set aside separate times for each role. In one session, you only decide (product). In another, you only build. In another, you only try to break what you've built. Mixing all three at once means the builder always wins.
⚡ Speed loves money
There's an old business saying: speed loves money. It was always true, but it's truer now. If any product can be cloned in weeks, the first mover's advantage is measured in weeks too.
✓ Speed that helps
- ✓Launch the minimum version and learn from real use.
- ✓Make decisions in days that used to take months.
- ✓Quickly drop what isn't working.
✗ Speed that gets in the way
- ✗Skip the question "is this necessary?".
- ✗Launch something broken that ruins the first impression.
- ✗Change your mind every week without finishing anything.
⚠️ Speed isn't rushing
Being fast means shortening the time between a decision and the learning it generates. Rushing means skipping the thinking. The questions in the next tracks take an afternoon; they're what make speed safe.
💡 Measure in days
For each idea, set a short deadline (seven to fourteen days) to get a first version into someone's hands. If it can't fit into that window, the idea is probably too big.
🧪 Quick module quiz
Three questions. Click an option to see the answer.
1. Why does "app inflation" make it harder to launch a product today?
2. When can validation be lighter?
3. Which role tends to disappear when one person does everything with the help of agents?
📋 Module summary
Next module:
1.2 - The Average Problem