More Builders, Broader AI Education: Seizing the AI Revolution

When Microsoft President Brad Smith was asked, in five words or less, what government must do to help America “win the AI race,” he replied: “more electricians, broader AI education.” His point was clear: we need the infrastructure to power AI and—crucially—a citizenry prepared to build and shape it.

On May 9, 2025, this conversation played out live before the Senate Commerce Committee, where OpenAI CEO Sam Altman testified alongside Dr. Lisa Su (CEO of AMD), Michael Intrator (CEO of CoreWeave), and Brad Smith himself. You can watch the full hearing here: https://www.youtube.com/watch?v=pxGE41V04fs

From “Using AI” to “Building AI”

Back in the early days of the internet, millions learned to navigate with browsers and to “Google it.” Today, a similar scene is unfolding: countless people can prompt ChatGPT or generate images with Stable Diffusion. But those surface-level interactions are quickly becoming table stakes. The true innovators are those who peer under the hood—people who grasp how transformer architectures and attention mechanisms drive model behavior, who fine-tune these models on domain-specific datasets to unlock new capabilities, and who weave LLMs into production workflows using techniques like Retrieval-Augmented Generation. They’re the ones designing resilient inference endpoints that auto-scale across cloud or edge environments to meet real-world demand. In other words, the winners in this new era won’t merely be consumers of AI tools; they’ll be the builders of AI systems.

What “AI Education” Really Means

When we talk about AI education, it can’t stop at demonstrating a few impressive demos. Real AI education immerses learners in the full lifecycle of an intelligent system. It begins by grounding them in the mathematical foundations of machine learning—linear algebra, probability, and the theory behind neural networks. From there, students dive into the practical toolchains: writing PyTorch or TensorFlow code, experimenting with Hugging Face pipelines, and orchestrating training jobs in Dockerized environments. Alongside these technical skills, they learn to wrangle data responsibly—cleaning and labeling it, detecting bias, testing for robustness, and ensuring explainability. The journey continues as practitioners design and deploy production-grade services, optimizing for latency, cost, and reliability. Finally, they learn to shape AI into products that solve real problems, measuring impact, designing user experiences around model outputs, and navigating regulatory safeguards. This comprehensive approach ensures that graduates don’t just know how to click “generate”; they know how to architect, build, and scale AI in the world.

Why Now Is the Moment

The convergence of open-source models, accessible cloud resources, and immersive training programs has created an unparalleled window of opportunity. Leading organizations and nonprofits are offering free or subsidized GPU credits and community tutorials. Startups and established companies alike are sponsoring bootcamps that teach MLOps in weeks rather than years. Meanwhile, the sheer pace of innovation means that early adopters who master these deep skills will find themselves riding a rocket of demand. But as Brad Smith cautioned, if we treat AI like a mere “phone-a-friend” tool, we risk ceding influence to those who build the engines of this revolution.

A New Call to Action

For students and career changers, the path forward is clear: seek out hands-on AI courses that go beyond prompting to cover generative AI toolkits, machine-learning pipelines, fine-tuning large language models, and advanced patterns like Retrieval-Augmented Generation. Dive into the open-source ecosystem, run training jobs yourself, and deploy real applications.

For companies, the challenge is equally urgent: invest deeply in in-house AI training programs that move teams from one-off experiments to a culture of continuous innovation. Embrace “fail fast” practices in high-frequency projects, encourage cross-functional learning and internal knowledge-sharing, and empower every employee to contribute ideas for AI-driven improvements.

In five words or less, it remains: “more electricians, broader AI education.” Now is the time not just to use the future, but to build it.