AI It Forward

In 1990, at a Gordon Research Conference, a young graduate student presented neural networks to a room of crystallographers and was met with: "That's no big deal. We knew that decades ago." The room went silent, and Sung-Hou Kim stepped in to rescue the moment.

The story is about humility, not nostalgia: expertise can either humiliate the next generation or help them stand back up.

The sentence "We've done this for decades" is not wrong. The problem is what it does socially. It closes the door just when the room needs more people: alternative ideas, creative pressure, skeptical pressure, non-scientific perspectives, and people who ask the naive question that experts no longer ask.

The lesson: AI is most powerful when it helps us cross the lanes we once stayed inside. Sung-Hou's lab worked because it was not one lane — crystallography, biology, chemistry, computation, and people willing to collide ideas.

Berkeley stands as a public-institution ideal: world-class excellence plus a diversity of disciplines and people.

The arc in one line: tools were built with AI for decades. What changed is that now AI builds the tools, in English.

These are not accomplishments to admire; they are examples of what the interface has become. Describe a need clearly enough and AI can help build toward it.

The two Food Health apps are the same story shipped to the App Store: one idea, described in English, became a tracker and a diabetes-focused companion.

The bionic arm does not decide what matters; it extends capability. AI extends reach, speed, recall, and iteration, but the human still sets the direction and checks the consequences.

On data-center anxiety: expect surges, overbuilds, and pullbacks, like many venture-backed cycles. None of that changes the deeper point — content and human intention remain central.

AI is not only for people who code. Its most powerful current interface is language, and language belongs to everyone. The curve (from a 1991 Berkeley thesis) shows why memorization without fresh data fails; the brains panel reminds us that raw capacity is not the moat — human content is. Content does not just reign supreme; content makes kings.

The forward-looking beat. The baseball is prediction fused to dexterous execution — the human game; the robot hand and light bulb is the machine that carries the idea. Boston Dynamics, Amazon's warehouses, and Waymo are already moving, and the quiet frontier is lightness — mass, energy, biomimicry. The machine is rising; the question, as with every fire we have lit, is what we choose to do with it.

Fire illuminated, warmed, cooked, protected, and transformed human life — and it also burned people and destroyed cities. Serious tools require serious culture: not naive optimism, but responsibility plus access.

Berkeley is extraordinary, but the next brilliant person may not have Berkeley access. AI can help more people find serious lanes to swim in.

The callback: at the Gordon Conference, Sung-Hou modeled how an expert can rescue a beginner instead of crushing him — the AI moment in miniature. We can be the voice in the audience, or the person who helps the next person keep going.

The fire is already lit. What we do with it is still up to us.