Hi Dotikers!

You've probably heard of Demis Hassabis. Chess master at 13. Programmer of the cult classic video game Theme Park at 17. In 2010, he founded DeepMind with a mission that made people smile at the time: "solve intelligence, then use intelligence to solve everything else." Google bought the startup four years later for $500 million.

First breakthrough: AlphaGo beats the world champion at Go in 2016. A game experts thought would remain out of reach for machines for decades.

Second breakthrough: AlphaFold solves a 50-year-old problem in biology in 2020. How does a protein fold in space? A protein's shape determines its function, and predicting that shape from its amino acid sequence was the Holy Grail of biochemistry. AlphaFold cracked it. For virtually every known protein. Nobel Prize in Chemistry 2024.

And now, DeepMind is going after DNA.

AlphaGenome has just been released. The model analyzes sequences of up to one million DNA letters and predicts what they do. Not just the 2% of the genome that codes for proteins, but more importantly the remaining 98%. This part, long dismissed as "junk DNA," actually controls how and when our genes turn on and off.

What's the real-world impact? A single mutation in your DNA can do nothing, or trigger cancer. Telling the difference is one of the great challenges of genomic medicine. AlphaGenome promises to answer this question: is this genetic variation dangerous?

Hassabis's trajectory is crystal clear. AlphaFold told us what proteins look like. AlphaGenome wants to tell us what DNA does. If the model delivers on its promises, a doctor could one day scan your entire genome and pinpoint exactly which variants to watch. Personalized medicine is no longer a concept. It's becoming an engineering problem.

M.