My notes on some unusually candid remarks from Jensen Huang at a 2026 AI summit.
1. The old moat is gone: coding was just typing#
Said three years ago, this would have sounded unhinged. Said in 2026, it feels like Jensen Huang just said the quiet part out loud: “Programming? That’s just typing. And typing isn’t worth much anymore.”
For thirty years we were told: learn Python, learn Java, and you’d hold the key to the future. Huang is telling us that era is over. As AI code generation has improved exponentially, syntax stopped being a real barrier.
Does that mean it’s over for programmers? No — it’s more like a release.
2. The rise of the domain expert: not knowing how to code is actually an advantage now#
Huang’s logic is blunt but genuinely hopeful: once the technical barrier drops to zero, the business barrier becomes effectively infinite.
If AI can write a perfect function in a second, what’s actually valuable? It’s the person who knows what function needs writing. The doctor who understands the underlying biology. The manager who understands the tangled logic of a supply chain. The marketer who genuinely understands people and markets.
Tomorrow’s standout individual contributor isn’t the person heads-down writing code — it’s the domain expert who understands the business, understands the customer, and knows how to direct AI to do the work. Huang put it directly: a fresh computer science graduate with great coding chops is worth less than a seasoned salesperson who understands what actually bothers their customers — because AI can write the code, but it can’t read a customer’s mind.
3. A required course for founders: if you want to control innovation, go see a therapist#
When asked how a company should manage AI-driven innovation that feels “out of control” internally, Huang’s answer was memorable: “If you want to control innovation, you should go see a therapist.”
In the AI era, the old habit of running everything through KPIs and ROI calculations is poison for innovation. Huang’s management philosophy is closer to “let a hundred flowers bloom.” Don’t ask on day one how much money an AI project will make. Allow some chaos first, allow trial and error, let a swarm of small AI-driven projects emerge inside the company. The leader’s job isn’t to control that — it’s to act like a gardener, pruning after things grow wild, and keeping the ones that actually turn into something real.
Trying to plan an AI revolution with a spreadsheet is an epitaph for the old way of doing things.
4. Resetting the business model: from “tools” to “digital labor”#
This was the most disruptive business insight of the whole conversation.
Huang pointed out that for decades, tech companies sold tools — screwdrivers, hammers, software. Now, we’re entering the era of the “AI factory.” Even hardware giants like Cisco and NVIDIA are finding their customers aren’t really buying “faster networking” or “more compute” anymore — they’re buying digital labor.
A self-driving car isn’t a car — it’s a digital driver. A smart customer-service system isn’t software — it’s a digital support agent.
Once you realize you’re manufacturing labor rather than tools, your addressable market jumps from the roughly one-trillion-dollar IT industry to the roughly hundred-trillion-dollar global real economy. That’s part of why he’s willing to say things like: Disney would rather be Netflix, and Mercedes would rather be Tesla.
5. Your question is worth more than the answer#
On the subject of data sovereignty, Huang offered a genuinely philosophical warning: “My question is my most valuable IP. The answer is cheap.”
In the generative-AI era, getting an answer is easy. The hard part is asking the right question. How you prompt AI, and the reasoning you use to guide it, reflects your actual strategic thinking — and that’s the real secret. That’s why companies need their own “sovereign AI,” rather than routing every core conversation through a public cloud — because the way you ask questions is itself a competitive advantage that can’t be copied.
Closing thought: an uncomfortable truth, and the only real way through it#
As the evening wound down, the conversation ended on a lighter note, but Huang’s closing line is worth remembering. In an era where AI compute is scaling a million times faster than Moore’s Law, anxiety doesn’t help. He repeated his now-famous line, and in 2026 it lands with real weight: “I swear to God, go apply this technology. You’re not going to lose your job to AI. You’re going to lose it to someone who knows how to use AI.”
In this wild AI era, don’t be the “typist” left behind. Be the domain expert who knows how to direct AI. There’s no better time to start than right now.
