Hi fam!

This week we zoomed out to look at the bigger picture of the AI race — from the battle for the best models to the companies controlling chips, distribution, and the future of AI-powered software.

Then we looked at something much more practical: a simple checklist for writing better image prompts, helping you get much clearer results from AI image generators. And finally, we shared a simple trick that can make meetings far more productive — using AI to generate smarter questions before the meeting even starts.

From industry shifts to everyday productivity.

Let’s dive in. 🚀

So I saw this chart floating around LinkedIn and it kind of blew my mind.

Imagine the AI world as three layers.

At the bottom: chips.
In the middle: AI models.
At the top: distribution.

And the crazy part?

The winners in each layer are almost already decided.

The competition here is about who has the smartest AI model at any time. OpenAI was ahead for most of the time, Claude and Grok had their hit at the throne, but Gemini is just crushing it through 2025.

But the real moat might actually be who owns the GPUs and the users. Google has an established advantage here, and they are starting to monetise on this. But we still wonder whether it will be enough to break OpenAI?

I am getting more prone to betting on Google :-)

Here’s a wild one.

Oracle is reportedly restructuring and potentially cutting tens of thousands of jobs.

And the reason is… AI.

Think about it. Did we already talk about this and predicted the layoffs?

Oracle is one of the biggest enterprise software companies in the world.
But now they’re pouring billions into AI infrastructure and data centers.

So what happens?

Some of the old work disappears.

Internal reports suggest AI coding tools are now writing large chunks of software that engineers used to write manually.

So the company is basically saying:

This is probably the first wave of AI replacing white-collar tech jobs at scale.

Not the last.

Alright, this one is fascinating. And close to my heart as I use all three of them.

Three companies are racing to become the place where software gets built in the AI era.

And each one is taking a totally different path.

Lovable might be the fastest-growing SaaS product right now.

The pitch is simple:

You type:

And it just… builds the app.

People love it because non-technical founders can suddenly build software.

But serious developers sometimes complain they don’t have enough control.

So it’s amazing for ideas and prototypes.

Cursor is basically ChatGPT built directly into a coding environment.

Developers love it because it can:

If Lovable is the idea machine, Cursor is the power tool.

Replit is trying something different.

They want to be the entire operating system for developers.

You can:

All in one place.

Think Google Docs for coding, but with AI agents building stuff.

But looking around me, it seems most builders actually use all three.

My new workflow looks like this:

1️⃣ Idea → Lovable
2️⃣ Prototype → Replit
3️⃣ Production code → Cursor

That’s a pretty wild shift from how it was done before. 🤔

Tech moves fast, but you're still playing catch-up?

That's exactly why 200K+ engineers working at Google, Meta, and Apple read The Code twice a week.

Here's what you get:

All delivered twice a week in just 2 short emails.

Join 200K+ engineers

Most meetings don’t go wrong during the meeting.

They go wrong before it even starts.

You join the call.
You know the topic.
But you’re not entirely sure what you should ask.

So the conversation flows… but not always in the most useful direction.

That’s where a very simple AI trick can help.

Instead of improvising questions during the meeting, you can ask AI to generate them before the meeting even begins.

The Simple Trick

Paste a short description of the meeting topic and ask AI something like this:

Or:

Within seconds you get a list of questions you might not have thought of.

Some will be obvious.

But often a few will immediately improve the conversation.

Example:

Imagine you’re preparing for a meeting about a new marketing campaign.

You could prompt AI with something like:

You might get suggestions like:

Suddenly the meeting becomes much clearer.

Why This Works

Good meetings usually depend on good questions.

But coming up with them on the spot is difficult.

AI can generate a wide range of possible questions quickly, helping you enter the meeting prepared and focused.

Instead of reacting to the conversation, you can help guide it.

You don’t need AI to run your meetings.

But using it to prepare smarter questions can completely change how productive those meetings become.

Sometimes the smartest thing you can bring to a meeting… is a better question.

Ever generated an AI image that was almost right?

The style works.
The idea is close.
But something feels slightly off.

The truth is simple: AI image generators depend heavily on the prompt.

A vague prompt produces vague results.

But a well-structured prompt can dramatically improve the outcome.

👉 In our latest article we share a simple checklist for writing better AI image prompts, with practical examples that can help you get much stronger results.

Because sometimes the difference between an average image and a great one is just a better prompt.

Interviews | AI tools | News in AI | How to… | AI tips & tricks | AI Side hustles

Our master plan is to publish an NFT image in each newsletter and hand it out to our subscribers. The earlier you subscribe, the smaller the series are. It's a future collectible, a piece of digital art that captures the essence of this moment in time.

Thank you for being a valued subscriber. Together, let's embrace the magic of AI and creativity!

Was this email forwarded to you? Sign up here 👇