Changelog Master Feed Biocomputing on human neurons (Changelog Interviews #654)
7 snips
Aug 14, 2025 Ewelina Kurtys, a scientist and entrepreneur with a PhD in neuroscience, leads exciting biocomputing research at FinalSpark. She shares groundbreaking ideas about using lab-grown human brain organoids for computing, significantly improving energy efficiency. The discussion dives into the challenges of encoding information in neurons, parallels between neuronal and human learning, and the transformative potential of AI in neuroscience. Ewelina’s insights open up a brave new world, highlighting a future where biology and computing intricately collaborate.
AI Snips
Chapters
Transcript
Episode notes
Remote Access To Living Neurons
- FinalSpark gives remote, programmatic access to living human neurons via a browser API.
- They use neurons as potential energy-efficient processors but remain in R&D without full image/sound processing yet.
Huge Energy Efficiency Potential
- Neurons are estimated to be about one million times more energy efficient than silicon.
- This efficiency motivates using neurons as a post-silicon processor for energy-constrained AI tasks.
Neurons Grown From Reprogrammed Skin
- FinalSpark uses human neurons derived from reprogrammed skin cells rather than growing whole brains.
- They transform skin cells into stem cells and then into neurons for experiments.