Lisbon is loud. Between the mechanical clatter of robots in the Meo Arena and the roar of a sold-out crowd, it is easy to lose the signal in the noise.
This year, the noise came from everywhere. Paddy Cosgrave, the CEO of Web Summit, opened the week by noting a shift in geography. The startups are no longer just coming from Berlin and London. They are coming from Poland, Qatar, China and Brazil. The stage belonged to Chinese robotics and Brazilian fintech as much as it did to Silicon Valley.
Strip away the spectacle and one truth remains. We are witnessing the collapse of the “technical priest.”
For the last twenty years, if you wanted to build, you needed to speak the language of the machine. You needed to know the syntax. That era is over. The tools are no longer the bottleneck. Our imagination is.
The barrier to entry hits zero
On stage, Anton Osika, the CEO of Lovable and a former particle physicist, shared a staggering number with TechCrunch’s Connie Loizos. His company reached $100 million in revenue in just eight months. That is not normal hyper-growth. That is a market correction. His thesis is that the “99%” who could not write code are now entering the room as builders. “The best people to focus on are the people who don’t know how to write code today,” he said.
What does this mean for the industry?
It means the unit of production is no longer the line of code. It is the product judgment. The moat protecting technical teams has evaporated.
This logic traveled from the server room to the supply chain. Later in the week, I sat down with Oisin Hanrahan, the founder of Keychain, off stage and discussed the future of manufacturing. Oisin, who previously sold Handy to Angie’s List, described a similar decoupling in the physical world. “Private label is going to win,” he told me. We are moving from a world of “Big Brand” dominance to a world where influencers and niche creators can spin up a manufacturing line as easily as they spin up a website. “This is the era where consumer preferences have changed,” he noted.
When a fashion influencer can architect a supply chain (Keychain) and a software stack (Lovable) in a week, the definition of a “tech company” dissolves.
The sovereign builder has arrived!
From discovery to decision
If the machine does the labor, what is left for the human?
The answer lies in the death of “discovery.” During the “Rewriting the workweek” panel I moderated, I explored this with David Shim of Read AI and Gautier Cloix of H Company.
The consensus was clear: we spend too much of our lives hunting for information. We dig through dashboards, Slack channels, and emails just to find the starting line.
That friction is disappearing. As David explained, the new class of agents does not just retrieve data; it contextualizes it. It tells you what happened in South Africa while you were asleep and presents six clear options. You do not search. You swipe right. You decide.
We are moving from an economy of search to an economy of choice.
This sentiment echoed loudly in “Smarter plays: How AI is changing the game,” a session featuring tennis legend Maria Sharapova and IBM’s Sarah Meron. Sharapova was initially skeptical of data in sports, fearing it would kill instinct. She found the opposite. “In the tiebreaker of the third set, you are no longer thinking about certain patterns,” she said. “Because you’ve already learned them, you know that they exist.”
The data removes the mental clutter. It allows the athlete—or the executive—to skip the rote pattern recognition and go straight to the tiebreaker decision.
Even on opening night, Khaby Lame, the most followed person on TikTok, reinforced this from a creative angle. You cannot automate the human element that connects. “I think it’s not the same thing [as] to put your heart to make people laugh,” he said.
The machine provides the leverage, but the human provides the taste.
The dog in the data center
Speed is dangerous!
If everyone can build and everyone can move fast, the risk of drift and failure multiplies. You cannot have a sovereign builder without sovereign rails.
In a technical deep dive, I interviewed Stephen Bye, the CEO of Ookla—the company that likely measures your internet speed. He shared an old industry saying that feels painfully relevant for the agentic age. “To run a factory or a network, you only need two things: a man and a dog. A dog to stop the man from touching the network and the man’s there to feed the dog.”
In Stephen’s view, the AI is the dog. It is the ultimate guardrail.
We often talk about AI as a generator of content. We rarely talk about it as a guardian of reliability. But as Stephen noted, the AI sees the outage before the customer calls. It prevents the human from making a clumsy manual error. Back in the Lovable session, Anton Osika used a similar analogy: “In these days you would never launch a space rocket [with just] a human. You have machines running this because they are more precise.”
This is the paradox of 2025. The front-end of the economy is becoming more democratic, creative, and chaotic. The back-end must become more rigid, automated, and precise.
Sovereignty is the new luxury
If you rely on rented rails, you are not a builder. You are a tenant.
This was the quiet undercurrent of the policy tracks. In my conversation with Theresa Swinehart from ICANN, she reminded us that the internet is a decentralized addressing system. Owning your Top-Level Domain (TLD) is owning your digital land. “The address on the outside of an envelope… actually connects people,” she said.
The concept of sovereignty is rewriting the financial logic of the creator economy as well. In the an insightful session, Mark Nelsen, Global Head of Product at Visa, argued that we need to move beyond simple payments to actual asset ownership. He described a near future where creators do not just seek loans, but self-fund through Web3 technologies—issuing tokens that guarantee a percentage of future earnings to their fans. It is a radical re-imagining of the IPO, shrunk down to the individual level.
Nelsen pointed out that while creators generate billions in value, they often struggle to get a credit card. They are businesses without a banking stack. They need financial sovereignty to match their creative reach.
The long game versus the fast game
If the US model is defined by the sprint, the Chinese model is defined by the marathon.
During a panel dissecting the global AI race, Einar Tangen, a Senior Fellow at the Taihe Institute, and Dr. Jostein Hauge, a political economist, outlined a stark divergence in philosophy. The US approach, driven by Silicon Valley venture capital, follows the “travels fastest” doctrine. It incentivizes rapid innovation, short-term valuation spikes, and a rush to market. It creates a volatile environment where policy shifts with the political winds.
China is playing a different game.
Their approach is guided by the “travels furthest” doctrine. It is not driven by quarterly returns but by five-year plans. As Mr. Tangen noted, the focus is not on speculative chatbots but on practical integration into manufacturing and logistics.
What does this look like on the ground.
It looks like restraint. While the West races to build hyperscale data centers, China has actually curbed expansion due to fears of overcapacity. They are prioritizing purposeful deployment over speculative infrastructure. It also looks like total sovereignty. Driven by a deep distrust of US-made components, China is aggressively building a domestic supply chain for semiconductors.
As the panel highlighted, China controls 98% of the global production of rare earth magnets. That is a choke point. While the US struggles with the labor shortages delaying its CHIPS Act factories, China is leveraging its vast engineering talent to transition from imitation to innovation. They are not just building the AI; they are building the grid, the robots, and the raw materials that the AI relies on.
The lesson here is uncomfortable but necessary. Innovation without a supply chain is just a prototype. Sovereignty is the only true scale.
The takeaway for leaders
We spent the last decade learning to speak to machines. We will spend the next decade teaching them to listen to us.
If you are leading a team through this transition, here is your playbook.
1. Audit your “Discovery” tax
Look at your calendar. How many hours does your team spend finding the status of a project versus moving the project? Use agents to automate the context gathering so your team can focus on the decision.
2. Hire for instinct, not syntax
If you are vetting talent solely on their ability to write boilerplate code, you are hiring for 2020. Look for the “Sharapova” trait—the ability to internalize data and make the intuitive tiebreaker call.
3. Install the dog
Identify the “kill chain” in your business—the systems that cannot fail. Put the AI in charge of watching them. Guardrails are not about slowing down; they are about moving fast without breaking things.
4. Own the rails
Whether it is your supply chain data, your cloud infrastructure, or your domain, do not rent your core value. In a world of infinite content, the only scarce asset is sovereignty.
The tools are here. The busy work is dying. The only question left is what you will build with the time you get back.
