A conversation with Stanford’s Prof. Stephen Boyd on AI education, open data, and where countries like Vietnam should be placing their bets.

When Professor Stephen Boyd of Stanford University, one of the world’s leading experts in optimization and mathematical engineering, sat down with the team at CEI, the conversation kept returning to the same uncomfortable question: is the AI field, for all its momentum, getting some fundamental things wrong?

His answer was yes. And his reasoning is worth sitting with.

The problem is the thinking

Boyd doesn’t dispute what large language models can do. He finds them genuinely stunning. But he’s troubled by what he’s seeing in the next generation of engineers: people trained almost entirely in the newest methods, with little to no grounding in the classical techniques: optimization, control systems, basic statistics that have kept infrastructure running for decades.
The deeper concern is cognitive. Before any method is chosen, Boyd insists on two questions that he finds most engineers today cannot answer cleanly: What is the problem? And how would you know if you’re doing a good job?

“Slow down. Tell me the problem. You can’t mention the method — please tell me the problem.”

Methods change every year, he notes. The questions don’t. And a field that skips the questions in its rush toward the methods ends up, in his words, going downhill from there.

Where Vietnam’s real opportunity lies

On Vietnam’s position in the global AI landscape, Boyd is direct: competing head-to-head with Meta or Google on foundation models is not the play. The compute requirements alone make it a losing race before it starts.
But that framing, he argues, misses where the actual opportunity is. The infrastructure for building powerful AI applications already exists through open-source models, open weights, publicly available tools that let a small team go from idea to working prototype faster than ever before. What’s missing is the domain knowledge, the local context, and the clarity of thought to apply that technology to problems that matter.

“Vietnam could easily come up with all sorts of awesome applications and if done openly, those solutions could translate to other developing countries.”

This is not a consolation prize. Domain-specific AI built on local expertise and real-world constraints is, in many cases, more durable and more impactful than another foundation model. The countries that will benefit most from the AI revolution are not necessarily the ones training the largest models. They’re the ones asking the clearest questions about what their communities actually need.

What this means for how we work at CEI

Boyd’s framing maps closely onto what CEI has been building toward. Our research spans environmental intelligence, materials science, and smart energy, domains where the questions are urgent, the local context is irreplaceable, and the opportunity to apply AI meaningfully is vast. We are not in the business of competing with Silicon Valley. We are in the business of asking what Vietnam’s environment, infrastructure, and communities actually need and building toward that.

The conversation with Boyd reinforced something we think about often: the most important thing a research center like CEI can do is maintain clarity of thought in a field that moves fast enough to obscure it. Not chasing the frontier for its own sake. Defining our own, from the ground up.

A platform for this conversation

This July, CEI is hosting HORIZONS 2026 at VinUniversity, bringing together researchers, technologists, and institutions from across the region to do exactly this: think carefully about where AI, environmental science, and emerging technologies intersect, and what responsible, locally grounded innovation looks like in practice.
Boyd’s questions are a good place to start. What is the problem Vietnam is actually trying to solve? And how will we know if we’re doing a good job?

HORIZONS 2026 takes place July 1–3 at VinUniversity, Hanoi. Co-organized with NUS, CUHK, and Fudan University. Learn more about the conference and how to participate: https://horizons2026.vinuni.edu.vn/.

Based on a conversation between Prof. Stephen Boyd (Stanford University) and the Center for Environmental Intelligence team, VinUniversity.