Daily Podcast #21

Source 🎧

How AI is saving billions of years of human research time | Max Jaderberg - TED Tech

Sentences ✍️

1. Ultimately, to create new knowledge that we can take back out and change the world around us.
   • back out: To take something that was developed or discovered within a system (like research or theory) and apply it outside of that system — into the real world or practical situations.
   • New sentence: I applied my thinking back out into the real world.
2. And then we haven’t worked out how to describe these complex dynamics in coarser and simpler terms that we could write down with maths.
   • coarser: Less detailed, less refined, or less precise — often used when simplifying something complex in order to make it easier to understand or describe.
   • New sentence: I don’t even remember that because I studied it in a coarser way.
3. Isomorphic, because we believe there is an isomorphism, a fundamental symmetry that we can create between the biological world and the world of information science, machine learning and AI. This material may be protected by copyright.
   • isomorphism: A structural similarity or one-to-one correspondence between two different systems, where parts of one system can be directly mapped to parts of another system in a meaningful way.
   • New sentence: Isomorphism is a very difficult concept.

Summarization 👀

Max Jaderberg explains how AI is transforming scientific research by creating virtual models of the real world, called AI analogs. These models allow scientists to experiment and discover new knowledge much faster than traditional methods.

A key example is AlphaFold, an AI model that can predict the 3D structure of proteins in seconds — a process that used to take years in a lab. This breakthrough has saved an estimated billion years of human research time.

Jaderberg believes that this is not a one-time event. With advances in neural networks, computing power, and data collection, AI can continue revolutionizing fields like drug design, materials science, and energy.

In drug discovery, new versions like AlphaFold 3 can model complex interactions between proteins, DNA, and drug molecules. This enables rapid virtual testing and design of new drugs.

Ultimately, he envisions a future where thousands of AI agents design personalized drugs for individual patients, accelerating medical breakthroughs and scientific progress at an unprecedented scale.