Autopoiesis Sciences

Let’s talk about one of the biggest weaknesses of today’s AI: it bluffs. That’s why I invested in Autopoiesis Sciences — maybe our best shot at tackling hallucinations for science. A new study by OpenAI and Georgia Institute of Technology shows why language models keep making things up. Not because they are just “badly built,” but because of how we train and test them. Today’s AI often guesses when it doesn’t know the answer. And worse: our benchmarks reward this behavior. Like students punished for leaving blanks, models learn to bluff with confidence instead of admitting uncertainty. For casual use, hallucinations may seem harmless. But in science, they are fatal: a wrong lead in drug discovery can waste years and millions. That’s why many researchers still keep their distance from AI — and they’re right. Skepticism is the foundation of good science. If we want AI to truly accelerate discovery, it must be transparent, rigorous, and able to say: “I don’t know.” That’s what drew me to Autopoiesis. Founded by what WirtschaftsWoche calls a “Wunderkind” from California Joseph Reth (23, serial tech founder and pioneer in self-learning systems), alongside Dr. Eike Gerhardt (former VC investor) and Lawrence Callahan (former senior FDA official), they are building Aristotle: an AI co-scientist that doesn’t bluff, but collaborates with researchers the way a rigorous colleague would. Really looking forward to seeing them grow! Because we need systems like theirs that are designed to embed the scientific method into AI itself and help us move from bluffing to trust.

From breakthrough research at Google to pioneering the future of AI for science at Autopoiesis Sciences. We’re proud to welcome Kavita Kulkarni as our new President and Head of Scientific Superintelligence. As a core leader of Google’s AI Co-Scientist team, she helped transform how AI approaches scientific reasoning. Now, she’ll guide our mission at Autopoiesis to build scientific superintelligence: AI that accelerates scientific discovery beyond human imagination.

Excited to share that we've partnered with Oracle Cloud Infrastructure to power the next phase of development for Aristotle, our AI co-scientist. The most revealing moments in our conversations with scientists happen when they ask whether they can trust AI with work that defines their careers. Real research demands integrity alongside intelligence, and Oracle's AI infrastructure gives us the computational power and enterprise security that serious scientific work requires. Their team recognizes what we've learned: building AI for science means earning the trust of people whose work shapes our understanding of the world. Grateful to work with partners who share our conviction that the future of science depends on getting this right. Dr. Eike Gerhardt Lawrence Callahan Abigail Fulton Jordan Bauer Matt Turner Kyle Sheehy

We’re partnering with Oracle Cloud Infrastructure to power the next phase of our AI co-scientist development. OCI AI infrastructure provides the computational power, enterprise security, and technical support needed to scale AI systems that researchers can trust. Full blog: https://lnkd.in/gg7K2fSC

This will change science forever. We're announcing new funding led by Informed Ventures to accelerate our mission: building the foundation for scientific superintelligence. We've achieved 92.4% accuracy on GPQA Diamond, one of the most rigorous scientific reasoning benchmarks, and 96.1% on SimpleQA, OpenAI's factuality benchmark. Surpassing every major AI system while solving the calibration problem. "Autopoiesis solves a critical barrier for the adoption of AI models by scientists: reliably generating accurate outputs without hallucinating. Scientific superintelligence has the potential to turbocharge scientific discovery in a way not seen since the discovery of the scientific method. Joseph and the team have built the best performing model specifically designed for scientists, and the exceptional performance of Aristotle will drive exponential adoption by frontline scientists looking to accelerate their research with a trusted co-pilot," - Sunny Kumar, MD, Partner at Informed Ventures Full blog: https://lnkd.in/g_ZANzSF

Autopoiesis Sciences is hiring for several engineering and design roles! 1. ML Scientist: https://lnkd.in/gJGWsckt 2. Software Engineer, High Performance ML: https://lnkd.in/geAuGncF 3. Product Designer: https://lnkd.in/g47WgySQ 4. Software Engineer: https://lnkd.in/gekWyvDv TC is competitive for all of our positions. SF & Toronto. Apply above, DM, or visit apois.co/careers for all available positions. Dr. Eike Gerhardt Larry Callahan Anna Butz Tran Huynh #Hiring #MachineLearning #Engineering #Design #SanFrancisco #Toronto