Barton Smith

I spend more time worrying about AI than almost anyone I know. Philosophers, economists, lawyers, artists, researchers are genuinely frightened by where this is going. Many of those conversations keep me up at night. Yet I'm still one of the most optimistic people in the room. Why: 1/ Energy. The world simply does not have enough fossil fuel to power what AI demands. That scarcity is a forcing function. It is accelerating fusion timelines in a way nothing else has. Not just because AI enables the plasma simulation complexity that was previously impossible, but because the capital environment it creates finally makes fusion fundable at the scale the physics requires. I've spent real time studying this. The engineering challenges are hard but not mysterious. The teams working on it are world class. We will crack it, alongside other new forms of energy and storage. 2/ Healthcare. I'm about as privileged as it gets on personal healthcare. Top private insurance, best doctors in London and New York, annual full body scans at Mayo Clinic. I had skin cancer before so I'm obsessive about it. And yet it was a £299 scan from Neko Health that caught a spot my dermatologists missed. Something that needed urgent removal and biopsy. That experience changed how I think about what "good healthcare" even means. 14% of Neko's early London scans required unexpected follow-up. That is a public health intervention at consumer prices, something that can truly only be done with AI. 3/ Drug discovery. Rentosertib is the first drug fully designed by AI. Target, molecule, everything. It treats a serious lung disease and in Phase IIa trials it showed real improvement in lung function after just 12 weeks. A process that used to take 15 years and billions of dollars, compressed into months. Most curable diseases will be solved this way. I have very little doubt about that. 4/ Materials science. Researchers at the University of New Hampshire used AI to build a database of 67,000 magnetic materials and found 25 novel ones that stay magnetic at high temperatures without rare-earth elements. Cheaper electric motors, better clean energy hardware, less dependency on problematic mining supply chains. This is one example from one university from one year. Multiply that across every lab on earth operating with these tools. Better solar cells, new catalysts for clean fuel, bio-based packaging that actually competes with plastics, low-cost water filtration systems, and new solvents for carbon capture. 5/ Food. AI-driven precision agriculture gives farmers satellite, drone, and sensor data telling them exactly where water, fertilizer, and pesticides are needed and where they are not. AI-based irrigation controllers on rice farms cut water wastage by 30% and lifted yields by 20%. More food, less of everything it takes to grow it. 6/ Environmental protection. Groups like Capacity are building AI systems that detect, report, and support enforcement against illegal deforestation in the Amazon in real time. They are currently protecting 13 million hectares of indigenous land alongside six indigenous communities. Illegal logging operations that were essentially invisible to enforcement agencies are now getting caught. This is AI doing something no other tool has been capable of doing at this scale. 7/ Education. Tools like Khanmigo are providing 24/7 AI tutoring in math and science to students who cannot afford private tutors, with measurable score improvements and better retention. In Afghanistan, girls banned from classrooms are accessing AI-powered lessons in local languages on cheap phones, some offline. Not a pilot program. Happening now. Compressing centuries of unequal access to knowledge. The risks are real. They may be existential. I am not dismissing them and I am not naive about them. Concentration of power, displacement, weaponization, the speed of change outpacing the institutions meant to govern it. All of it is serious and deserves serious work. But I am genuinely, substantively optimistic. The good stuff is truly extraordinary.

c) Make a mediocre product and be the best at marketing it Historically this hasn't worked in software because making software was so difficult that few could do it. But as barriers fall and it becomes trivial for anyone to make "good enough" software, the core challenge shifts away from execution and toward discovery. This is already prevalent on over-saturated platforms like the App Store where many great apps sit undiscovered and mediocre ones succeed based on aggressive marketing strategies. No surprise, this is the playbook we see in other industries that have undergone similar transformations (DTC, apparel, food). Like Rasmus, I hope to see a lot more take route (b). But if we see a massive flood of new software, don't be surprised to see a lot of people find success with option (c).

I struggle to understand the argument that everyone will just create their own software. Spending on paid services for everyday tasks continues to grow every year, and knowing what you want and how to articulate it is a difficult skill that we may be underestimating.

Meet Sprout. Today, we’re releasing a new kind of robotics platform. One designed to move out of the lab and into the real world, closer to the people who will shape what robots become next. [1/4]