Here are four stories about artificial intelligence expanding what people can understand, invent and improve. A few of them are several years old, but still very important. I'll use this space to hilight the advances that AI is providing to humanity.
From the hidden shape of proteins to the mathematics inside modern computing, these are evidence-led accounts of AI moving ideas into the world.
Read this week’s discoveries ↓One system discovers a new algorithm in pure mathematics. Another offers a clinician practical guidance in real time. Together, they show that AI’s scientific value is not confined to one kind of problem.
Original research, translated for interested readers.
From an amino-acid sequence, AlphaFold can predict a protein’s three-dimensional structure with accuracy that transformed structural biology. It's creation was the reason the Nobel Prize in Chemistry was awared to Demis Hassabis and John Jumper in 2024.
AI is used to explor possible compounds, thus expanding the map of stable inorganic crystals that researchers can investigate.
A deep-learning system measures mucosa touch rate during a procedure and gives immediate visual feedback to the clinician.
Deep reinforcement learning reframed algorithm discovery as a game—and found new, provably correct ways to multiply matrices.
Each entry begins with original research or a primary project source. We explain what changed, why it matters and what remains uncertain—because genuine progress deserves both enthusiasm and precision.