GOTO - The Brightest Minds in Tech State of the Art of Biological Computing • Ewelina Kurtys & Charles Humble
Mar 17, 2026
Ewelina Kurtys, strategy advisor and scientist-turned-entrepreneur working on bio-inspired computing, discusses building computers from living neurons. She talks about why living neurons could be vastly more energy efficient, the technical challenges of encoding and plasticity, organoid lifespans and lab constraints. The conversation covers ethics around stem cells, remote-access neuroprototypes, and the potential impact on AI costs.
AI Snips
Chapters
Books
Transcript
Episode notes
Choosing Living Cells Over Bioinspired Algorithms
- Biocomputing differs from bio-inspired algorithms by using living cells as the computation substrate rather than only copying biological ideas.
- FinalSpark chose the substrate route believing actual neurons offer qualitatively larger gains than algorithmic or analog-hardware tweaks.
Biocomputers Trade Speed For Complex Problem Efficiency
- Biological computers will likely trade raw speed and memory for energy-efficient complex problem solving; they mirror human brains: slow but powerful for high-level tasks.
- Ewelina expects biocomputers to excel at generative AI and complex problems, not high-throughput numeric tasks.
Neural Encoding Uses Time And Space Not Bits
- Encoding in brains is fundamentally different from binary: information is represented in time and space (when and where neurons fire), not simple zeros and ones.
- Kurtys calls encoding the biggest challenge because we cannot yet reliably decode neural signals to know 'what they mean.'
