Google’s DeepMind has solved protein folding, a problem that has vexed scientists since 1972.
Proteins are made up of long strings of amino acids that fold into unique shapes. These shapes determine how a protein functions. By knowing the shape of particular proteins, scientists can figure out how a molecule will bind to it, leading to disease research and treatment improvements.
Their predictions have an average error of 1.6 Angstroms, or about the width of an atom.
DeepMind rose to fame when AlphaGo beat a human champion in the complex game of Go in 2016. The company took a similar approach to both challenges: training a neural network with vast amounts of information about the problem.
Turning this into a story, what if there’s a future world where all human scientists wait for artificial intelligence’s solution to their research problems? Instead of embarking on years of research, they simply plug the data into the neural networks and take action on the results. There could be multiple versions from competing companies.
Then, someone finds a way to upload their own consciousness into this world of artificial intelligences. “The Wizard of Oz” meets “Tron.”
The challenges from “The Wizard of Oz” are recreated in a computer simulation world that shares many characteristics with the digital world in Neuromancer by William Gibson.
The initial question the uploaded person wants to be answered is the meaning of life, a nod to The Hitchhikers Guide to the Galaxy. They meet allies, go through various challenges, and at the end, discover the scary, powerful Oz-type character is, in reality, an old man, like the Architect in “The Matrix.”
The battle of the story could be returning back to the real world.
A second book in the series could be about protecting the digital world from businesses who want to monopolize it for their own private use. A third could be about a new AI system introduced to the digital world that can’t coexist with the others, leading to war inside the system.