Home feed

The misfits. The rebels. The troublemakers. The round pegs in the square holes. The ones who see things differently. They’re not fond of rules. And they have no respect for the status quo. You can quote them, disagree with them, glorify or vilify them. About the only thing you can’t do is ignore them. Because they change things. They push the human race forward. And while some may see them as the crazy ones, we see genius. Because the people who are crazy enough to think they can change the world, are the ones who do.

~Steve Jobs

I remember writing about this in February.

At the time, I was researching UX3 and Large Blockchain Models, and honestly, I thought I was failing.

I was publishing research about something that almost nobody seemed interested in.

I would write these long papers about why user experience could not always be solved at the interface layer.

Why some systems needed to understand behaviour.

Why machines needed to interpret intent.

Why complexity should be absorbed by infrastructure instead of constantly being pushed back to the user.

I was looking at blockchain because that was the system I was studying.

But looking back at it now, I think I was asking a much bigger question.

How do we build systems that understand humans instead of forcing humans to understand systems?

That was the real beginning of my research journey.

I wasn’t really researching UX in the conventional sense.

I wasn’t going around asking people what button they liked or whether a screen was intuitive.

I was asking the system questions.

Why does the user have to understand this?

Why does this complexity reach the user at all?

Why can’t the system infer what the user is trying to do?

Why should a human have to become an expert just to safely use a piece of technology?

My UX3 research eventually framed this as a problem of where intelligence lives.

If the intelligence lives only in the interface, the user still has to carry the complexity.

If intelligence moves into the system, the system can begin to absorb that complexity before it reaches the user.

That was the idea.

And my Large Blockchain Models research pushed it further.

I was looking at models that could understand behavioural patterns and infer intent from onchain activity, rather than simply giving people more information to interpret.

At the time, it felt extremely niche.

Now I’m watching designers, product people, engineers and AI researchers have conversations about agents, intent, natural language interfaces and systems that can understand what people are trying to accomplish.

And I’m realizing that I wasn’t really researching blockchain UX.

I was researching the relationship between human intention and machine behaviour.

Blockchain was just where I happened to find the problem first.

That’s probably why leaving blockchain ended up being such an important decision for me.

I had spent so much time trying to fit my thinking into an industry that I forgot the industry wasn’t the subject.

The subject was always the thinking.

Now I can look at AI, architecture, finance, software, products, interfaces or whatever else catches my attention and ask the same question:

What happens when the system becomes intelligent enough to understand what the human actually means?

And that changes everything.

It changes what an interface needs to be.

It changes what a designer does.

It changes how products are built.

It changes what engineers have to worry about.

It changes what UX even means.

And maybe the funniest part is that I almost gave up on this research because nobody was reading it.

I got chicken feet and ran. 😂

But I think that’s when I actually became a researcher.

Not when people started reading.

When I kept asking questions nobody around me seemed particularly interested in yet.

I’m still very far from where I want to be.

But I’m finally starting to see the thread.

And that’s a pretty beautiful feeling.

If we are going to call for better marketing, then we better teach teach people better marketing or at least show them the way to it.

To put this into effect is why I'm fixing a flaw in highschool and college marketing teaching and particularly one of the topics — Transportation.

Marketing textbooks make a vague and incomplete view of the matter so I fixed it.

Updating what Transportation In Marketing really means.

From: Movement of goods and services from producer to consumer

To: Movement of feelings, ideas, perspectives, and solutions from one point to the other.

A much more suitable remark for the value of transportation in marketing.

The nuances and thought process all in the link below.

In 10-20 years most of anything finance and more will move on-chain; payments, records, even agents running in the background.

That’s the direction we’re already heading and privacy doesn’t disappear in that world.

And privacy will never be the default setting for everything.

Accountability still has to live somewhere, we already watched it play out.

• Monero got delisted from the big venues because you can’t easily report what you can’t see.

• Tornado Cash is still running but the founder is looking at real time.

• Silk Road is another story to talk about.

The blockchain tech will forever stay and privacy will come as an option and people who needs it will still use it.

But the main rails will keep a version they can actually live with.

Full default opacity will survive in the corners but everywhere else will become a feature you switch on when it matters.

That changes everything.

A website used to be: Information.

Then: Interaction.

Then: Transaction.

Now we're moving toward: Action.

And eventually: Autonomous action

THE NEW MONEY IS MOVING UP THE STACK

Think about this: AI makes execution cheaper.

Therefore execution becomes less scarce.

What becomes scarce?

Strategy Knowing what to build.

Distribution Knowing how to reach people.

Trust Getting people to believe you.

Data Having useful proprietary information.

Taste Knowing what's actually good.

Relationships Having access to customers and networks.

Systems Connecting everything together.

Capital Being able to scale what works.

This is why I wouldn't advise you to compete purely on:

I can build websites.

Think about this:

AI makes execution cheaper.

Therefore execution becomes less scarce.

What becomes scarce?

Strategy

Knowing what to build.

Distribution

Knowing how to reach people.

Trust

Getting people to believe you.

Data

Having useful proprietary information.

Taste

Knowing what's actually good.

Relationships

Having access to customers and networks.

Systems

Connecting everything together.

Capital

Being able to scale what works.

This is why I wouldn't advise you to compete purely on:

"I can build websites."

This is another uncomfortable truth.

AI can now produce:

images videos websites ads logos articles code voice presentations product descriptions

So content itself is becoming abundant.

Which means the scarce thing becomes:

attention + trust + distribution + originality.

That's why simply "creating content" isn't enough.

You need a reason people should care.

We're entering:

AI vs AI.

Attackers use AI.

Defenders use AI.

Agents attack systems.

Agents defend systems.

AI generates code.

AI scans code.

AI generates vulnerabilities.

AI searches for vulnerabilities.

That creates a massive new market for:

AI security + agent security + identity + permissions + monitoring.

There are already reports of AI-assisted attacks becoming more sophisticated, including campaigns using large numbers of agents.

Now let's separate real concerns from internet conspiracy.

There are legitimate questions around:

AI concentration

A relatively small number of companies control enormous amounts of:

compute models cloud infrastructure data distribution

That's worth watching.

Surveillance

AI makes analysing enormous amounts of information dramatically easier.

That's a real concern.

Employment

AI may not simply eliminate jobs.

It may eliminate tasks, restructure jobs and reduce the number of people needed for certain workflows.

Information control

AI increasingly sits between people and information.

That means whoever controls:

models + search + recommendation + distribution

can potentially influence what information people encounter.

Autonomous systems

The biggest long-term concern may not be "AI becomes evil."

It may be:

Humans give increasingly powerful systems authority they don't fully understand.

That's considerably more realistic than most movie-style conspiracies.

This is probably one of the biggest business opportunities.

Imagine a small company with:

2 human employees 8 AI agents automated CRM automated email automated customer support automated research automated reporting automated lead qualification automated content production automated bookkeeping

That company could potentially operate like a much larger company.

This is why AI automation is such a strong area for freelancers.

You're not selling AI.

You're selling:

less manual work + faster execution + lower operating cost + more capacity.

Google just launched the Gemini app for Windows.

You can access it directly from the desktop using a keyboard shortcut.

This sounds small.

It isn't.

The direction is: AI shouldn't live in a browser tab.

It should live where you work.

Your computer becomes increasingly: AI + operating system + applications + agents.

Eventually you may not open 15 apps.

You tell your computer what you want done.

We're also seeing serious investment into alternatives and specialized AI chips.

Today, EUCLYD raised more than €200 million to develop energy-efficient AI processors and data-centre systems.

The bigger story isn't just Nvidia.

It's:

Who can make AI computation cheaper, faster and more energy-efficient?

That's a gigantic market.

It may be:

electricity.

It may be:

electricity.

AI requires:

GPUs data centres cooling networking storage electricity land semiconductor manufacturing

AI infrastructure spending has become enormous.

And investors are now starting to question whether the spending pace can continue indefinitely. Reuters reported growing concerns about a potential slowdown in AI capital expenditure while major technology companies are projected to spend hundreds of billions on AI infrastructure in 2026.

That's why I wouldn't only watch OpenAI, Google and Nvidia.

Watch:

power + chips + cooling + networking + data centres.

That's where another generation of businesses could be created.

One of the more interesting recent developments is AI being used beyond content and software.

For example, recent research has demonstrated AI systems controlling fusion plasma on extremely short timescales, reacting to instability before it develops.

We're seeing AI applied to:

drug discovery materials science biology physics chemistry robotics energy climate modelling

This is where AI becomes much bigger than the write my email era.

This is another trend people aren't talking about enough.

Google DeepMind is working on Gemini Robotics systems designed to give robots whole-body intelligence, dexterity and teamwork capabilities.

So AI isn't staying inside computers.

It's moving into:

robots → factories → warehouses → vehicles → laboratories → physical environments.

The long-term transition is:

AI software → AI + machines → autonomous physical systems.