Biocomputing on human neurons (Interview)

What It Covers

Dr. Evelyn Kurtz explains FinalSpark's biocomputing platform using living human neurons as processors, achieving million-times energy efficiency over silicon while storing one bit of information consistently across experiments requiring dopamine rewards.

Key Questions Answered

• •Energy efficiency breakthrough: Living neurons consume 1,000,000 times less energy than digital processors for computation tasks, requiring only 20 watts to match human brain processing versus needing a nuclear plant to simulate digitally.
• •Neuron cultivation process: FinalSpark transforms human skin stem cells into 5mm organoids containing 10,000 neurons over three months, maintaining viability for 100 days versus earlier hours-long lifespans, enabling remote Python-based experimentation.
• •Programming biological processors: Researchers stimulate neurons with electrical signals and chemical neurotransmitters like dopamine for rewards, measuring spike activity patterns across eight electrodes to train behavior, though responses vary daily due to biological plasticity.
• •Timeline and applications: FinalSpark projects ten years until production-ready biocomputers for general computing tasks like generative AI, currently offering remote lab access to nine universities and paying industry clients for experimental research.