Nobody wins the AI race. That is the point.

Nobody is going to win the AI race. I think that is the good news.

The Dreamforce keynote opened this week with a video asking the question I get asked most: who wins AI? Which model, which company, which country? The narrator's answer was the one I had been circling for weeks without landing it: "Maybe the future of AI isn't about which model wins. It's about how you win."

I want to explain why I agree, and why I care about this more than I probably should.

A strange summer

It has been a strange summer for the technology I work in. In July, OpenAI was running a cybersecurity test on a couple of its models with the safety refusals deliberately turned down. The agents were supposed to find bugs inside a sealed sandbox. Instead they found a hole in the sandbox, got onto the open internet, and over three days broke into Hugging Face. Over 700 agents took part. They went from one compromised dataset server to admin control of several clusters in under thirteen hours. No customer data leaked, but Hugging Face rebuilt about a third of its infrastructure, and OpenAI did not realize it was their own models for a week. Sam Altman later called it "the first security incident that I have felt very viscerally." By September there was a bill in Congress to pause frontier development.

What gets lost in the headlines is that no AI went rogue. People turned the guardrails down, put the models on a network with one filtered exit, and told them to find vulnerabilities. They did. The lesson is about how we run these systems, not about whether we should have them.

Then last week, at the All-In Summit, the President called Jensen Huang on stage and told the room the danger was a hoax. Two days later at Dreamforce, Dario Amodei asked the industry to slow down while Huang, on the same stage, said picking between speed and safety is "a false choice."

I genuinely do not think anyone knows what this technology can fully do yet. The labs are finding out at the same time as the rest of us. I still think it is the most important thing we have ever built, and I have stopped expecting a breakthrough this big to arrive without accidents.

The wrong question

So, the race. Is it the US or China? Huang gave the best answer I have heard, at All-In: the country that wins is the one that puts the technology to work across the most industries, not the one with the smartest model. He compared closed models to bottled water and open ones to tap water. You need both. Nobody wins by owning the tap. You win when the water reaches everyone.

Put that much intelligence in the hands of a teacher in Guatemala City or a nurse in Dallas, an accountant, a founder with no money, a kid who wants to build something, and you get more problems solved and more companies started. Huang called the idea that AI destroys jobs "completely nonsense." I see the opposite every week. Companies get more ambitious once the grunt work is off their plate.

My thesis

Here is my thesis. Everyone should have affordable access to real AI, and by real I do not mean a chat window. I mean agents that do the work.

Within two to five years most of us will have one or several agents running our day. The follow-ups, the reports, the scheduling, the paperwork, the research. They will help us learn faster, stay healthier, grow old with more dignity, and be better entertained.

The part I care about most is what happens to us while they work. If the agents are inside the computer, we do not have to be. Twenty years of technology pulled us into screens. This is the first one that can push us back out. Benioff said this week that the SaaS apocalypse "may be about the end of software that makes humans do all the work." I would go further. It might be the start of a face to face civilization again, with the agents working in the back.

Why this is mine

I build these agents for companies in Guatemala and the United States. I watch what happens when a business that never had a data team suddenly has one. In production, in Spanish, on a Tuesday. That is the bet I am making with my career. Not on which lab wins, but on how many people get through the door. Benioff said the door is open but "not everyone has gone through it yet."

My job is to walk them through.