Apiro Data

06/08/2026

The most digital generation on earth just made the most analogue film of the year a billion dollar hit.

Half of The Odyssey's audience is under 35. They queued for sold out 70mm screenings, then went and downloaded the audiobook of a three thousand year old poem.

I share what I think is really going on in the video, and it has very little to do with Homer.

What is the last thing you left the house for because watching it on your phone felt like a waste?

04/08/2026

Every morning, they are waiting for me.

For the past couple of weeks, my morning walk in Rhodes has taken me past Cat Café Rhodes.

Every day, I am greeted by beautiful faces looking up at me expectantly.

Anyone who knows cats knows that look. Equal parts curiosity and hope, with the quiet suspicion that I may be carrying breakfast.

I stop. Of course I stop.

Apparently, my morning walk now includes being emotionally manipulated by cats through a window.

What a great concept this place is.

Rhodes has thousands of stray cats, and the café gives rescued cats somewhere safe to receive care and wait for a home. Some are adopted and already moved to places like Germany, Luxembourg, and the Netherlands. Other cats can be sponsored by people who want to help, with no need to take a cat home.

Wanting to help and being able to adopt are two different things. Sponsorship gives people another way to make a real difference.

A safe place can change the whole direction of an animal's life.

If you love cats and you are in a position to help, take a look at Cat Café Rhodes and consider sponsoring one.

👉 https://catcaferhodes.com/

Have you ever sponsored a rescue cat and found yourself checking for updates as though they were already yours?

Cat Cafe Rhodes. Das erste Cat Cafe in der Dodekanes mit sozialem Hintergrund. Coffee*Rescue*Adoption. Unsere Kätzchen suchen ein Forever Home

31/07/2026

The Geopolitical Kill Switch: Why the World Panicked 🌍

In 2026, a US export ban on Anthropic sent shockwaves globally.

This 'geopolitical kill switch' forced nations to rethink their reliance on foreign tech. From France swapping Palantir for local tech to Canada leveraging its climate for supercomputing, the race for AI sovereignty is on. Is your infrastructure safe?

29/07/2026

Why Companies Pay a 20% Premium for Sovereign AI Security

Is the price of data security worth a 20% premium?

Companies are ditching centralized AI giants for 'Sovereign AI' to build liability firewalls, but it comes at a massive cost. Discover why the global AI landscape is fracturing into walled gardens and what it means for the future of communication.

27/07/2026

Airbnb was passed over by seven investors when it tried to raise $150,000.

Five said no.
Two did not even reply.

Twelve years later, the company opened on Nasdaq at more than $100 billion.

That number deserves a moment, because it tells you something important about how early-stage value actually gets read, or more accurately, misread.

The product that eventually became one of the most used online booking tools, wasn't obvious to the people whose job it is to spot obvious things. It wasn't a slow burn either. It was rejection, then more rejection, then a number that made those early passes look like the most expensive decisions those rooms ever made.

There's a version of this story that founders tell themselves as comfort. 𝘒𝘦𝘦𝘱 𝘨𝘰𝘪𝘯𝘨. 𝘚𝘰𝘮𝘦𝘰𝘯𝘦 𝘸𝘪𝘭𝘭 𝘴𝘦𝘦 𝘪𝘵 𝘦𝘷𝘦𝘯𝘵𝘶𝘢𝘭𝘭𝘺. And that's true, but it's also incomplete.

The more interesting question is what changed between rejection four and funding. Was the product fundamentally different? Was the market? Was it timing, distribution, a single conversation that opened the right door?

Usually it's all of those things at once, which is exactly why pattern matching on persistence wins misses the point. Persistence is necessary, but it is not enough.

What actually compounds is iteration under pressure, staying close enough to real usage that you keep adjusting until the product fits something the market genuinely needs, whether the market knows it yet or not.

The $100 billion number is the headline. The seven rejections are the actual case study.

24/07/2026

Nvidia's $95B Monopoly: Why New AI Chips are Impossible to Get

Nvidia has a $95 billion grip on the global AI hardware supply, leaving smaller players and even sovereign nations in the dust.

But even if you get the chips, can your data center handle the heat? The 140kW Blackwell bottleneck is real.

23/07/2026

An email just retired part of some peoples career.

Slight exaggeration.

But that was my first thought yesterday afternoon, seeing Microsoft in my inbox with a subject line where the word retirement sat a little too comfortably next to SMS authentication.

Well, this breakup has been coming for years.

Microsoft announced it this week in the politest way possible. Passkeys become the default sign in from September, and from February 2027 they stop sending SMS and voice codes altogether. The six digit text is being shown the door.

Most of the people I work with come from the messaging industry, and I promise you not one of them fell off their chair reading that email.

The industry has watched enterprises drift away from text codes for a long time, and it knows exactly why. Fraud got out of hand, with fake verification traffic pumped through networks to inflate bills until trust wore thin. Prices in some destinations climbed so high that a login code cost more than the customer's first purchase. And regulators kept adding rules, one country at a time, until sending a simple text felt like filing a customs declaration.

There is a part Microsoft's email skips over.

Passkeys assume a smartphone, a credential manager and a user comfortable with both. There are plenty of markets where that assumption falls apart fast, and where local regulation will have its own opinion. Moving every user in every country off SMS in seven months is a bold promise.

Nobody I know calls this a death.

It's a forced transformation, and the timeline now belongs to someone else.

If you work in messaging or identity, can Microsoft really take passkeys global by February 2027, or does that deadline quietly slip?

22/07/2026

Two developers build an app. PhD-level research, solid backend, auth, payments, onboarding, the full stack.

Then they ship it and… nothing much happens.

This story is so common it's almost a rite of passage, and I say that without any condescension, I've watched it play out dozens of times, and the people it happens to are often genuinely talented builders.

I have also been there myself and I can tell you that the problem is that building well feels like progress. It is progress, and it feel so good! But it's progress on only one half of the equation.

Distribution isn't a feature you add later. It's a parallel workstream that should have started the moment the idea was validated, or honestly, before that. The question 𝘸𝘩𝘰 𝘯𝘦𝘦𝘥𝘴 𝘵𝘩𝘪𝘴 𝘢𝘯𝘥 𝘩𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘳𝘦𝘢𝘤𝘩 𝘵𝘩𝘦𝘮 is just as hard as 𝘩𝘰𝘸 𝘥𝘰 𝘸𝘦 𝘣𝘶𝘪𝘭𝘥 𝘵𝘩𝘪𝘴? In my experience it's harder and it takes just as long to figure out.

Recently I have been asked a lot about AI and marketing and this is my honest take...

AI can help you move faster once you know what you're saying and who you're saying it to. But it can't tell you that. If you don't yet understand your customer's actual problem, their language, their objections, AI will just help you produce more of the wrong message, a lot faster.

Find someone who actually knows this space. Not to hand it off entirely, but to shorten the learning curve. The founders who figure out distribution early don't do it by accident. They treat it with the same rigor they give the product.

The build is never the finish line, no matter how much many builders would love it to be.

22/07/2026

Global AI models are missing the mark in Southeast Asia. 🌏

From linguistic nuances to massive liability risks, here’s why 'Sovereign AI' like Singapore’s Sea-Lion is the future. Plus, the 91% blind spot that could cost your company millions.

17/07/2026

Through the years, I have found that the real chain of command in most organisations isn't on any org chart.

It's the person who knows which spreadsheet to never touch. The one who quietly fixes the thing that would have escalated to the CEO. The one who everyone routes around the official process to reach, because the official process hasn't worked for the pst 5 years.

I see this on almost every engagement. The formal structure says one thing. The actual flow of decisions, information, and unspoken authority says something completely different.

And when organizations try to embed AI into their operations without understanding that real chain, the informal one, they hit a wall fast. You can automate a process nobody actually follows. You can build a dashboard nobody trusts. You can redesign a workflow that bypasses the three people the work actually runs through.

The technology is rarely the hard part. Understanding who really makes things move, where decisions actually slow down, and which workarounds have quietly become the system people follow, that's the hard part.

Before any operating model change, AI rollout, or transformation program, the most valuable thing I do is just watch. Who do people actually go to? Where do things stall? What does the silence in a meeting mean?

The org chart is a hypothesis. The real chain of command is what you find when you get close enough to see it.

That's where the work starts.