# Jensen Huang Didn't Actually Admit AGI Has Arrived

I just finished listening to his latest podcast. It's mainly nostalgic reflections on glorious days of the past.

(Parentheses contain my personal editorial commentary)

## 1. Old Huang Almost Destroyed the Company

NVIDIA used to only make gaming GPUs. If it had continued this way, it might have just been an ordinary profitable company.

This leather-jacket-wearing man seemed destined for something extraordinary. He decided to bet everything on CUDA, forcing the tools for AI computing into ordinary gaming GPUs. In the first year, everybody was stunned—GPU costs surged 50%, profits nearly disappeared entirely, and market cap crashed from $8 billion to $1.5 billion.

(The worst part that old Huang didn't even mention: he came to China to schmooze with various entrepreneurs, like Lei Jun and others.)

Nobody knew what he was trying to do. He held on for 10 years before climbing out. Now the entire world's AI depends on NVIDIA.

He said if you want to win big, you have to dare to bet. To get everyone in the world using it, you have to go all-in.

## 2. Jensen Huang's Quirky Management Style

He has 60+ direct reports and doesn't hold one-on-one meetings. All issues are discussed collectively. Their business is too diverse: chips, CPUs, memory, networking, power supplies, cooling systems, server racks.

(These meetings test people's initiative—those who don't speak up voluntarily get called out.)

## 3. His Views on AI's Future

AI will become increasingly intelligent, but ultimately it all comes down to compute power. Intelligence is ultimately determined by compute power.

But token costs can drop 10x annually. Pre-training, fine-tuning, inference-time reasoning, and then comes agents. AI spawns swarms of smaller AIs to help do the work.

## 4. Jensen Huang's Vision Has Moved Beyond Chips

Old Huang wants to build planetary-scale infrastructure—a mega-project at Earth's scale: data centers, power grids, cooling systems.

## 5. The Future of Programming

Old Huang believes billions of people will be programming in the future, because programming will just become writing requirement specifications—telling AI what you want done and letting AI write the code.

He gave an example: AI reads X-rays more accurately than doctors. Yet the number of radiologists actually increased because they can now examine more patients.

(This also refutes the notion that education will become useless later—how could people use AI if they don't study?)

## 6. AGI and Human Nature

Old Huang said AGI has already been realized, but 100,000 AI agents working together still couldn't create an NVIDIA.

(This got misquoted by the media.)

Because intelligence will become a commodity like water and electricity. But what's truly amazing is human nature—character, trust, compassion, leadership—things AI can't replicate yet.

(AI still can't pull off revolutionary enterprises. No matter how smart, great companies still need great captains and helmsmen.)

## 7. Old Huang on Himself

- He said he's worked with TSMC for 30 years with hundreds of billions in business, yet never signed a single contract.
(Sounds a bit like boasting, to be honest.)

- When stressed, break it down into a checklist: What is this? What can I do about it? Then just do it.

- He hopes to die at his desk working.

- His mind is always thinking about the next 10 years ahead.

- He praised China as the country with the fastest innovation speed in the world.

## Summary

Really, when it comes to sitting down at the table, bringing your entire company with you, betting on destiny, betting on the future, betting on the next 10 years—most who did that died.

He bet right on the future.

So now the entire world is running behind him.