The AI That Solved a 50-Year-Old Problem and Won a Nobel Prize.

Part 1 of the AI Receipts series.

Lately, my feed has one message on repeat: AI is bad.

‍I work in this industry. I run a podcast about network engineering. I get the skepticism, and a lot of it is earned. The hype is exhausting. The job-loss fear is real. Half the products with "AI-powered" on the box are a search bar with a markup.

‍But "AI is bad" is not an argument. It is a mood. And "AI will save us all" is not an argument either. It is a sales pitch. Neither one survives contact with evidence.‍ ‍

Why bother?

Two camps are forming, and both are getting louder. On one side, the AI zealots, certain the technology can do no wrong and anyone who hesitates is a dinosaur. On the other, the "AI is bad" crowd, certain it is all theft, hype, and destroyed careers, and anyone who sees value in it has been conned. They are not debating. They are yelling past each other, each dug into its own bias, each sure the other side is the entire problem.

‍Meanwhile the technology is not waiting for them to sort it out. AI is here, and it is staying. We are at the front edge of a shift in society and the economy that is arriving faster than the Industrial Revolution did. That is not a reason to panic. It is a reason to pay attention, because a change moving this fast is the worst possible thing to be confused about.

‍If the two loudest camps just keep screaming across the aisle, the only thing they produce is a more divided room. And a divided population cannot cooperate. A population that cannot cooperate cannot solve the hard problems this shift is about to hand all of us.

‍I do not think this is binary. You do not have to be a true believer or a hater. The useful position is the boring one: look at the evidence, discuss it respectfully, and let the data move you.

That is what this series is. Not opinions, receipts. Real, verifiable cases of AI helping people in ways that matter, each one linked to a primary source you can check yourself. No hype decks, no vendor marketing, and where the science is early or contested, I will say so. I am leaning toward what is working, because in my corner of the world that is the half getting drowned out. The goal is not to win a fight for one team. It is to think clearly, together, about something that is going to touch every one of us.‍ ‍

Let’s start with the biggest one.

For half a century, biology had a problem it could not crack.

‍Proteins are the machines that run your body. They build tissue, carry signals, fight infection, and drive nearly every chemical reaction that keeps you alive. What a given protein actually does depends almost entirely on how it folds into a three-dimensional shape. Work out the shape, and you can design drugs, vaccines, and treatments around it.‍ ‍

Predicting that shape from the underlying genetic sequence was staggeringly hard. Determining a single protein structure through lab work could take years, and plenty of proteins were never solved at all. Researchers had been stuck below roughly 40 percent accuracy for decades.

In 2020, an AI system called AlphaFold, built by Google DeepMind, broke the problem open. It went on to predict the structures of nearly every known protein on Earth, more than 200 million of them, and the team made the entire database free for anyone to use.‍ ‍

The results are not subtle. AlphaFold has been used by more than two million researchers across 190 countries, and work that took years now takes minutes. In 2024, the creators won the Nobel Prize in Chemistry for it.

Sit with what happened. This was not a tool that replaced biologists. It handed every biologist on the planet a 200-million-entry head start, for free, and then accelerated the pace of an entire field. Drug discovery, vaccine design, and disease research all got faster because one hard problem stopped being a bottleneck.‍ ‍

That is what a force multiplier looks like. Keep that phrase in mind, because it is going to come up in every post in this series. ‍

Next up: machine learning that designs new antibiotics to fight the superbugs our current drugs can no longer beat. If you think I am wrong along the way, good. Bring data. Let’s have a dialogue. That’s the entire point.

‍/Andy

Sources

• The Nobel Prize in Chemistry 2024, press release — NobelPrize.org

• The Nobel Prize in Chemistry 2024, popular science background — NobelPrize.org (2M+ users, 190 countries, years to minutes)

• AlphaFold Protein Structure Database in 2024 — Nucleic Acids Research (214M+ structures)

• Demis Hassabis and John Jumper awarded Nobel Prize in Chemistry — Google DeepMind (free, open access)