Part 1: From Durable Code to Disposable Intelligence
The Shift You’re Not Seeing
Everyone’s talking about how easy it’s become to build software. AI tools write code, generate UIs, and even prototype apps with zero dev experience needed. That’s true – but it hides a more serious issue.
Software didn’t just get easier. It got harder.
Not at the level of typing code, but at the level of making smart bets – choosing what’s worth building, what will last, and where to stake your ground.
If you’re not deeply plugged into AI or tech, that may sound strange. Let’s break it down.
When Code Was Gold
Not long ago, software followed a simple formula:
You built a product once.
You sold it to lots of people.
You made money each time – without doing more work.
That’s leverage. That’s why tech was such a powerful business model.
Remember CD-ROMs? You’d buy Microsoft Office or a computer game on disc. One team wrote the code. Millions of people paid for it. You could stamp and ship those discs for pennies. Microsoft got very, very rich doing that.
That core idea – build once, sell forever – defined decades of software business. And it’s fading.
Code Is Cheap. Strategy Is Expensive.
AI flipped the economics of code.
You can now build an MVP in a weekend.
You can create tools, apps, automations – faster than ever.
The friction is gone.
But so is the long-term value.
Why? Because:
AI evolves constantly. What works today might be useless in six months.
Everyone has access to the same tools. Your edge disappears fast.
Much of what’s being built is shallow. Disposable. A proof-of-concept, not a product.
You’re not building durable software anymore – you’re spinning up temporary intelligence.
The Chariot Analogy (And Why It Matters)
Let me give you a metaphor. Stay with me.
3,000 years ago, in the Near East, the most powerful military tech was the chariot. If your army had chariots, you had the advantage – on flat land.
But chariots don’t work in mountains.
So the tribes in the hills? They couldn’t be conquered. Their terrain was too rough. Too inconvenient. The chariots just didn’t go there.
Today’s AI giants – OpenAI, Google, Anthropic – are the chariots.
They dominate wide, flat markets. General tools. Mass-scale products.
But they can’t reach every terrain.
Your job? Go to the mountains.
Find the messy, weird, high-friction spaces they’ll never bother with:
Complex regulation
Niche industries
Regional workflows
Human nuance
Legacy systems
These are the places where you can build something that lasts.
What Building Looks Like Now
If you’re trying to create meaningful tech today, your playbook has to change.
Here’s where the real work is:
Pick your terrain wisely. Don’t compete on general features. Go narrow and deep.
Own your data. Private, unique, hard-to-get data is your moat.
Use distribution as leverage. If you already have customer trust or industry access, use it.
Expect change. Build knowing your stack, your model, and your assumptions may all shift in months.
Avoid building features disguised as startups. If OpenAI can bundle your product into ChatGPT tomorrow, you don’t have a company – you have a menu option.
Why This Matters to You
You don’t have to be an AI expert or software engineer to care about this.
If you’re in business: it affects what tech you should buy or build.
If you’re investing: it changes what makes a startup valuable.
If you’re leading a team: it redefines how to think about long-term planning.
Software isn’t just lines of code anymore. It’s a question of terrain, timing, and trust.
In Part 2, we’ll get into specifics:
What niches are worth targeting?
What kinds of AI products are defensible?
What should you do now if you’re building, investing, or leading in tech?
Let us know if you want it sent straight to your inbox.
