Salt, water, and flour

There are millions of bakeries in the world. Some are good, some are bad.

And every one of them uses the same essential ingredients.

Flour, water, butter, sugar, salt, yeast… Some combination of these staples makes everything you love, everything you hate, and everything you think is just ok.

In the world of code, the ingredients are just as simple.

You have user interfaces (how users interact with data), APIs (how data platforms talk to each other), and you have databases (how data is stored).

Every app you’ve ever hated, liked, or loved is essentially a combination of these ingredients.

From Spotify to Notion to any SaaS product you can name, they all use the same essential ingredients.

And AI lets us build these platforms and better understand data contained within databases.

That’s it.

The problems that AI can solve are exactly the same as the problems software can solve in general: Anything that can be solved with a UI, APIs, and databases.

These ingredients were off-limits to the general population until recently. Closely guarded secrets of a brilliant, gate-keeping, coding elite.

But now we all have access to these tools.

It’s no more than the mastery of a few simple ingredients. It’s no less than the ability to bake anything we can dream of.

The hardest tech stacks have become the easiest

A decade ago, the race for the easiest website builders was raging.

From Squarespace to Wix to WordPress (and now Webflow), the trend was clear: Make it easier for people who don’t know how to code to make a website.

While these tools helped millions of people get out there, they were never the best. How could they be?

The trade-off for their ease was bloated code bases, tons of plug-ins, and a hornets’ nest of issues thinly veiled by a deceptively simple user interface.

But the real pros?

They were using React, Next.js, and custom CMSs like Contentful, etc.

The real pros weren’t afraid to use an API or two and get their hands dirty.

My agency began as a WordPress development firm, in large part because 25 years ago, the first languages I learned were HTML, PHP, and JavaScript—the core frameworks for WordPress.

But I always admired the “real” apps that were lightning fast using more advanced tech stacks.

This envy is what led me to transition my company to these tech stacks a couple of years ago.

But what I didn’t know was that AI would make coding these “difficult” languages easier (and faster) for us than the alternatives.

I never dreamed that coding in the best tech stack would also be the easiest.

But today, thanks to AI, it’s actually much easier for me to prototype and execute a complicated idea in Next.js—with APIs everywhere—than it is to use a drag-and-drop builder like Webflow.

When the best solution is also the easiest? You know you are in an exciting time.

Upskill or die

A recent article says that Accenture is planning on “exiting” staff who can’t be upskilled on AI.

This means, quite simply, folks who are too stubborn to embrace technological change.

“The only constant is change” - Heraclitus

I would argue that the fear of this tech goes up proportionately with one’s lack of knowledge in how to actually use these tools.

If AI is a concept that might get you fired, you’ll hate it.

If AI is a tool that lets you build literally any software application you can imagine, you’ll be excited.

Upskilling isn’t scary. It’s not mean. It’s not complex.

It’s just about approaching these tools with curiosity instead of fear.

It’s about attempting to go beyond just typing into Claude and using them to actually build software.

This is where the magic happens.

This is where the next phase of your growth occurs.

The movie ‘Memento’ is the best example of how coding with AI works today

In Christopher Nolan’s movie, the lead character has short-term memory that only lasts for a few minutes to a few hours. Afterwards, he remembers nothing.

In vibe coding (coding using AI), this is called a “context window”.

When the context window is full, the AI restarts, and remembers nothing about what you were just doing. Zero memory.

But just as the main character in Memento can help himself remember certain things by taking pictures or tattooing things on his body, so you can give the next memory-free AI agent some kind of context with updated documents and by compacting conversation...

Does it work? Kinda. How well did tattoos work for the guy in Memento?

It’s the same thing.

Sometimes AI kills the wrong guy.

The two-tiered pricing model every service provider must use

(and every client must accept)

Option 1: Provider uses AI = lower price.

Option 2: Provider doesn’t use AI = higher price.

If you want to hire a service provider and drive down the price, drive up your dependence on automation. You can’t have a Michelin-star steak for McDonald’s pricing.