Follow me for:

📰 News about AI, vibe coding and entrepreneurship.

💡 Tips on how to scale software companies. 

👀 A front-row seat to my building process.

Théo Canale

"Software is over. Everything will be handled by AI agents. Websites will be built for LLMs, not people."

This is absolute nonsense.

Don't get me wrong, I deeply believe in AI's transformative potential. 

But let's call this what it is: intellectual laziness masquerading as a "bold prediction".

Humans are visual creatures. We are not command lines.

We spent the last 20 years evolving UI/UX for a reason. We built beautiful dashboards, intuitive toggles, and drag-and-drop canvases because they make interaction easier, not harder.

Chat is a terrible interface for complex workflows.

Do you really want to manage a 50-person project by typing paragraphs into a text box? Do you want to edit a video by arguing with a chatbot?

No.

You want to click a button. You want to see the timeline. You want the friction removed, not shifted to prompt engineering.

Intelligence is valuable. But interface is control.

The immediate future isn't "No UI." It's "God-tier UI" powered by intelligence in the background.

If you are betting on a future where we all just type into a void... You are betting against human nature.

SaaS isn't dying. Lazy, friction-heavy software is.

The OpenAI x Jony Ive rumors sound like a total wreck. And looking at the Rabbit R1 and Humane Pin, they probably are.

The graveyard of "AI Hardware" is already overflowing.

𝐓𝐡𝐞 𝐨𝐛𝐯𝐢𝐨𝐮𝐬 𝐦𝐨𝐯𝐞 would be to stay in the software lane. 

Let Apple handle the glass. Let ChatGPT just be a super-app. Keep margins high and supply chain headaches low.

But Jony Ive is betting on the non-obvious move: Form factor dictates behavior.

If AI lives on your phone, it is trapped by the notification economy. It competes with Instagram, email, and dopamine loops.

If AI lives in a "dumb" pen (or badge, or orb): It bypasses the distraction economy entirely.

𝐓𝐡𝐞 𝐜𝐨𝐧𝐬𝐞𝐧𝐬𝐮𝐬: We don't need another device to charge.
𝐓𝐡𝐞 𝐜𝐨𝐧𝐭𝐫𝐚𝐫𝐢𝐚𝐧: We need a device that disconnects us from the feed but keeps us connected to intelligence.

It’s a massive gamble. 

It could be the next iPod. Or it could be a $1 billion paperweight that nobody asked for.

What do you think? Won't lie... I'm a big Jony fan, but nonetheless quite skeptical.

With the rise of agentic browsers, LinkedIn has a major decision to make.

Option A: Adapt to the future of the internet. 
Option B: Suspend the entire user base for "Breach of Contract."

Knowing LinkedIn, I’d start backing up data now.

Everyone in tech was glued to CES last week.

They watched Jensen Huang. They watched the consumer robotics demos. They watched the AI hype cycle spin faster.

But the most interesting story wasn't on the main stage.

It didn't happen in Vegas. It happened in Paris.

Dassault Aviation dropped €200M into a startup most have never heard of: Harmattan AI.

Just like that, France minted its first Defense Tech "unicorn" (€1.4B valuation).

On paper, this is the "Anduril Moment" Europe has been waiting for. The thesis is flawless and the details are quite promising:

1. Vertical Integration: Harmattan isn't a "legacy assembler" managing 50 subcontractors. They own the stack.

- Hardware: They are scaling to produce 10,000 drones per month.
- Product Line: They have the Sahara for ISR, the Sonora for training (already 3,000 units ordered by the UK), and the Gobi for kinetic interception.

2. Sovereign AI: Their compute runs locally. No cloud. No US servers. Pure "Strategic Autonomy".

3. Strategic Lock-in: This isn't just a VC bet. Dassault is integrating Harmattan's AI directly into the future Rafale F5 and the FCAS combat cloud.

It’s a perfect pitch and a great headline.

But when I look at the market, not the press release, this doesn't look like a renaissance.

It looks like a rounding error.

Here is the cold reality of European Defense that a €200M check cannot fix:

1. The "Zoo" Problem: The US operates 1 main battle tank. Europe operates 17. The US operates 6 fighter jet types. Europe operates 20. We don’t have a defense industry; we have a fragmentation museum. One unicorn doesn't solve a supply chain broken by 30 years of national ego.

2. The "Speed" Trap: Harmattan is fast. But the bureaucracy buying their drones is glacial. While Harmattan scales, France and Germany are still fighting over workshare for the FCAS program. You cannot fight an exponential war with linear politicians.

3. The "Budget" Illusion: We celebrate a €1.4B valuation? Anduril is worth billions more. The US defense budget dwarfs ours. We are bringing a knife to a nuclear gunfight and congratulating ourselves on the handle design.

My verdict?

Europe isn't marching. Europe is moonwalking.

The illusion is masterful: We extend one foot forward with "Strategic Autonomy" speeches and flashy funding rounds... But the friction of 27 fragmented bureaucracies drags the other foot backward.

To the audience, it looks like we are charging ahead. But in reality, we are sliding further away from relevance every day.

On critique le "vibe coding" et le "AI slop". Mais parlons de l'UX de nos services publics.

J'ai le transcript de la dernière réunion URSSAF:

"Chef, le parcours utilisateur est trop fluide." 
"Ajoute une étape." 
"Mais ils ont déjà tout rempli ?" 
"Demande-leur un code envoyé par courrier postal." 
"Génie!"

J'attends de voir le premier Agent AI capable de renouveler une carte grise sans faire un burnout.

Le vrai test c'est pas Turing, c'est le Cerfa.

Stop obsessing over Claude Cowork’s features and look at its birth certificate.

Built in under 14 days and entirely vibe-coded.

Still measuring your dev cycles in months? If so, life is about to get real tough.

Me: Rage-prompting after Gemini failed 5 times in a row.
Gemini: "The challenge lies in extracting the valid statements amidst the vitriol".

Turns out they built a Stoic philosopher.

I build autonomous AI systems that don’t just learn, they act.
I’m a systems architect and researcher with deep roots in machine learning, behavioral modeling, and real-world deployment. My work spans algorithmic trading, cross-chain infrastructure, and emotional AI, all tied together by a mission to build intelligent agents that can think, adapt, and execute in dynamic environments.
Currently looking for :
Funded PhD positions (agent systems, cognition, behavior modeling)

taplio

The Battle

Théo Canale

Nikhil .

Followers

Posts

Engagements

Likes

Performance ScoreThéo Canale's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Théo Canale

Nikhil .

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Théo Canale's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreThéo Canale's LinkedIn Performance