In this video, I break down what I think will be remembered as a turning point: 2026 being the year AI stopped just “assisting” scientists and started making real discoveries. I cover 15 standout examples across medicine, materials science, physics, energy, biology, and even math—cases where AI systems found patterns, mechanisms, or candidate solutions that humans either missed for decades or couldn’t realistically search for at scale. The big theme is that these aren’t small optimizations; they’re paradigm-shifting leaps that change what research looks like when the discovery engine is automated. I also explain why so many of these breakthroughs happened quietly. A lot of AI-driven discovery doesn’t arrive as a viral headline—it shows up as a new method, a surprising material property, a new drug lead, or a weird-but-true relationship in data that later becomes a “law” people build on. My main takeaway: once AI can propose hypotheses, test them in simulation, and iterate faster than any lab team, the pace of innovation doesn’t just speed up—it changes shape. If you’re wondering whether AI can make discoveries without humans, this video shows how close we already are, and why the acceleration is only getting more intense.