Editors' pick

Gates on AI

A three-piece reader, meant to be read in order.

These are three pieces we pulled from Bill Gates's longer arc of writing on AI. He himself placed them on the same page for readers. We turned them into a small reader: they are one conversation, not three.

The first piece names what the machine actually is, and which industries it will invert. The second walks through the specific risks and explains why panic is not the answer. The third, from August 2026, is Gates saying, more plainly than before, that this transition is a historical one and none of us can sit it out.

Piece 1 · November 2023

AI-powered agents are the future of computing

What is actually changing under our feet

Gates has been thinking about 'personal agents' since 1995. This is where he brings thirty years of that thinking down to earth: a real agent turns services that used to be affordable only to the wealthy, a private doctor, a private lawyer, a private tutor, a personal assistant, into something ordinary people can use every day. He walks through four areas one at a time.

"The most exciting impact of AI agents is the way they will democratize services that today are too expensive for most people."

Read the original at Gates Notes →

Piece 2 · July 2023

The risks of AI are real but manageable

What is at risk, and why panic is not the answer

Gates uses the history of every past technology shock, cars, calculators, printing, phishing, as the reference frame. He lays out each risk, then gives one very practical conclusion: we have learned to handle each one before. This time we just have to learn faster. The essay ends with a named English teacher who did not ban ChatGPT, did the opposite, and won.

"The best reason to believe that we can manage the risks is that we have done it before."

Read the original at Gates Notes →

Our three companion essays:

Piece 3 · August 2026

The turbulent AI era is here

What we must do now

This is Gates three years on. He admits that his 2023 confidence, 'we will manage these risks,' now needs revision because the world did not manage them. He lays out the three things that must happen: build new governance institutions, set aside a 'Human Reserved' domain of work that only people should do, and rebalance the way we tax labor versus capital.

"We need both: deep concern about the AI harms we need to minimize, and grounded optimism about the positives if we maximize them for everyone."

Read the original at Gates Notes →

Our three companion essays:

Why we published these together

The most valuable thing about these three pieces is that Gates refuses to let the reader pick a side. He asks you to keep both eyes open at once: to see what AI will run over, and to see what it can lift up. That 'hold both at once' posture is exactly what Luminis is trying to do.

We put the three sources together and paired them with three companion essays, one for parents, one for teachers, one for students, so that you finish this page not with a single feeling but with a small handful of specific things you can carry into next week's dinner, next week's classroom, next week's conversation with a child.