When a piece of technology feels inevitable — like electricity, or the internet, or the smartphone — it’s usually because we’ve underestimated how fast it has become infrastructure. That seems to be where we are with AI, and more specifically, with ChatGPT.
From our view here at Dellecod, we’ve been thinking a lot about what it means for people, for companies, and for systems of governance when something man-made starts feeling like air pressure. When a tool starts to behave more like an environment.
ChatGPT isn’t just “a product that’s growing fast.” It’s the fastest consumer product in history — and it’s not just outpacing other products, it’s beginning to reshape the way other software behaves. Calendar tools, team chat, code editors, search, notebooks, presentations — it's all being reimagined through the lens of having access to ambient intelligence, available 24/7, deeply integrated into the way we think and collaborate.
The podcast we listened to this week framed it bluntly: ChatGPT is on its way to becoming a monopoly. Regardless of how you feel about that word, what’s undeniably true is that centralized platforms that consume massive data and serve models to billions are defining the frontier. And the center of gravity is pulling hard toward OpenAI.
What worries people — and what we've been talking about internally — is the geopolitical asymmetry of letting one or two U.S.-based companies own the most strategically important digital infrastructure of this era. The number that keeps coming up is 150: that’s approximately how many people in the world can build foundational AI models today, according to NVIDIA’s Jensen Huang. It’s a chilling statistic, not just because it reflects how advanced (and fragile) our current systems are, but because it means control is already in the hands of a very small cohort.
And yet, under that elite surface, another story is unfolding — a more hopeful one. The subscription model of ChatGPT may be quietly reinventing internet monetization. For decades, we’ve been stuck in the ad-based economy. You got your Gmail and your YouTube for free, but your data was the product. This new chapter flips that. People now expect to pay for a smarter, non-intrusive assistant. We’re watching the birth of a premium layer of the internet. And users seem ready to adopt it.
Economically, the implications are sweeping. There’s real potential for 4–6% GDP growth, sustained without inflation, driven by AI-aided productivity. That kind of growth hasn't been seen frequently in recent U.S. history — maybe 4.6 years out of every decade between 1950 and 2010. But it's not unimaginable today, especially if we're able to rebalance value creation toward creation itself, instead of bloated administrative systems.
This is where the conversation veers into more uncomfortable territory: government scale. One perspective in the podcast was that around 50% of the U.S. federal government could be pared back without a drop in output, assuming the military stays untouched. Whether or not you agree with that number, the underlying point resonates — modern bureaucracies weren’t built with automation in mind. If AI can do the equivalent work of enormous departments, the incentives to keep headcount as a proxy for function will have to change.
In tech, we're already seeing those pressure lines. Big incumbents are challenged — but not gone. Microsoft had the early advantage due to its OpenAI partnership, and it’s still significant. But with Office apps increasingly redundant in the face of assisted writing tools, the moat is shrinking. Google, once synonymous with search, feels stuck in amber. Despite its immense datasets, the company has yet to deliver a truly compelling AI assistant product. The innovation engine feels oddly cautious.
Apple is a special case — a hardware company through and through. If AI becomes ambient, wearable, and spatial, Apple could retain its relevance through the sheer quality of its integration. Imagine in-ear devices that replace much of your screen time, or gesture-based AI interpreters layered into glasses. But culturally, Apple has struggled to lead on AI, and Siri’s own reputation hasn’t aged well.
Still, this isn't just a Big Tech arms race. The disruption is coming more laterally — from real estate to fintech to logistics. Look at Opendoor. It could’ve been a $50–100 billion company if it had taken the right path after its peak. Real estate remains the world’s largest asset class, but technological adoption has lagged, largely because it’s an operational business masquerading as one of data. Opendoor’s story suggests that even massive opportunity can be squandered if leadership doesn't match complexity with clarity. Still, the space remains wide open.
Fintech too feels ripe again. Affirm showed what happens when you pair underwriting innovation with strong distribution. Meanwhile, Wise, Trade Republic, and infrastructure plays like AtoB or Imprint are showing that even highly regulated, capital-intensive domains can yield to well-designed software. The opportunity here isn’t in building neobanks for the 100th time — it’s in embedding intelligence where finance and real-world transactions actually intersect.
At the center of all this, a quieter idea keeps emerging — the role of the outsider. The podcast brought up something that resonated deeply with us: that most fundamental innovations come not from domain experts, but from practitioners asking weird questions. PayPal, Square, Stripe — most early builders there weren’t seasoned financiers, they were engineers and generalists. They read obsessively. They poked at contradictions. They didn’t wait to become qualified.
Which leads us to a personal takeaway — one we talk about often at Dellecod. It’s not just about tools. It’s about becoming people who ask better questions. We don’t need every credential in the world to understand what’s happening. But we do need a framework — one governed by curiosity, clarity, and deliberate pace. Reading more helps. Writing helps, too.
This new AI-driven economy isn’t arriving in neat waves — it’s arriving like weather. You step outside and the air feels different. The sky is brighter or heavier. People are acting like something’s shifted, even if they can’t name it.
That’s our sense of it right now. Disruption everywhere, and yet emerging patterns. Old monopolies facing real pressure — but new ones forming, fast. A chance to build something better, still alive, but not guaranteed.
We’re still asking questions. That seems to be the right place to begin.
From our view here at Dellecod, we’ve been thinking a lot about what it means for people, for companies, and for systems of governance when something man-made starts feeling like air pressure. When a tool starts to behave more like an environment.
ChatGPT isn’t just “a product that’s growing fast.” It’s the fastest consumer product in history — and it’s not just outpacing other products, it’s beginning to reshape the way other software behaves. Calendar tools, team chat, code editors, search, notebooks, presentations — it's all being reimagined through the lens of having access to ambient intelligence, available 24/7, deeply integrated into the way we think and collaborate.
The podcast we listened to this week framed it bluntly: ChatGPT is on its way to becoming a monopoly. Regardless of how you feel about that word, what’s undeniably true is that centralized platforms that consume massive data and serve models to billions are defining the frontier. And the center of gravity is pulling hard toward OpenAI.
What worries people — and what we've been talking about internally — is the geopolitical asymmetry of letting one or two U.S.-based companies own the most strategically important digital infrastructure of this era. The number that keeps coming up is 150: that’s approximately how many people in the world can build foundational AI models today, according to NVIDIA’s Jensen Huang. It’s a chilling statistic, not just because it reflects how advanced (and fragile) our current systems are, but because it means control is already in the hands of a very small cohort.
And yet, under that elite surface, another story is unfolding — a more hopeful one. The subscription model of ChatGPT may be quietly reinventing internet monetization. For decades, we’ve been stuck in the ad-based economy. You got your Gmail and your YouTube for free, but your data was the product. This new chapter flips that. People now expect to pay for a smarter, non-intrusive assistant. We’re watching the birth of a premium layer of the internet. And users seem ready to adopt it.
Economically, the implications are sweeping. There’s real potential for 4–6% GDP growth, sustained without inflation, driven by AI-aided productivity. That kind of growth hasn't been seen frequently in recent U.S. history — maybe 4.6 years out of every decade between 1950 and 2010. But it's not unimaginable today, especially if we're able to rebalance value creation toward creation itself, instead of bloated administrative systems.
This is where the conversation veers into more uncomfortable territory: government scale. One perspective in the podcast was that around 50% of the U.S. federal government could be pared back without a drop in output, assuming the military stays untouched. Whether or not you agree with that number, the underlying point resonates — modern bureaucracies weren’t built with automation in mind. If AI can do the equivalent work of enormous departments, the incentives to keep headcount as a proxy for function will have to change.
In tech, we're already seeing those pressure lines. Big incumbents are challenged — but not gone. Microsoft had the early advantage due to its OpenAI partnership, and it’s still significant. But with Office apps increasingly redundant in the face of assisted writing tools, the moat is shrinking. Google, once synonymous with search, feels stuck in amber. Despite its immense datasets, the company has yet to deliver a truly compelling AI assistant product. The innovation engine feels oddly cautious.
Apple is a special case — a hardware company through and through. If AI becomes ambient, wearable, and spatial, Apple could retain its relevance through the sheer quality of its integration. Imagine in-ear devices that replace much of your screen time, or gesture-based AI interpreters layered into glasses. But culturally, Apple has struggled to lead on AI, and Siri’s own reputation hasn’t aged well.
Still, this isn't just a Big Tech arms race. The disruption is coming more laterally — from real estate to fintech to logistics. Look at Opendoor. It could’ve been a $50–100 billion company if it had taken the right path after its peak. Real estate remains the world’s largest asset class, but technological adoption has lagged, largely because it’s an operational business masquerading as one of data. Opendoor’s story suggests that even massive opportunity can be squandered if leadership doesn't match complexity with clarity. Still, the space remains wide open.
Fintech too feels ripe again. Affirm showed what happens when you pair underwriting innovation with strong distribution. Meanwhile, Wise, Trade Republic, and infrastructure plays like AtoB or Imprint are showing that even highly regulated, capital-intensive domains can yield to well-designed software. The opportunity here isn’t in building neobanks for the 100th time — it’s in embedding intelligence where finance and real-world transactions actually intersect.
At the center of all this, a quieter idea keeps emerging — the role of the outsider. The podcast brought up something that resonated deeply with us: that most fundamental innovations come not from domain experts, but from practitioners asking weird questions. PayPal, Square, Stripe — most early builders there weren’t seasoned financiers, they were engineers and generalists. They read obsessively. They poked at contradictions. They didn’t wait to become qualified.
Which leads us to a personal takeaway — one we talk about often at Dellecod. It’s not just about tools. It’s about becoming people who ask better questions. We don’t need every credential in the world to understand what’s happening. But we do need a framework — one governed by curiosity, clarity, and deliberate pace. Reading more helps. Writing helps, too.
This new AI-driven economy isn’t arriving in neat waves — it’s arriving like weather. You step outside and the air feels different. The sky is brighter or heavier. People are acting like something’s shifted, even if they can’t name it.
That’s our sense of it right now. Disruption everywhere, and yet emerging patterns. Old monopolies facing real pressure — but new ones forming, fast. A chance to build something better, still alive, but not guaranteed.
We’re still asking questions. That seems to be the right place to begin.