Biocomputing push uses human brain cells as alternative AI hardware

The organoid‑intelligence story is a reminder that the race to build smarter machines is not limited to bigger GPUs and better transformers. By treating living neural tissue as a computational substrate, these projects are probing whether biology’s efficiency and plasticity can be harnessed where conventional AI and silicon struggle—especially in energy use and continual learning. The early results are extremely primitive, but they hint at a very different scaling path: instead of stacking more FLOPs, you might grow more tissue or couple organoids with classical compute in tight loops.([psypost.org](https://www.psypost.org/how-scientists-are-growing-computers-from-human-brain-cells-and-why-they-want-to-keep-doing-it?utm_source=openai))

From an AGI perspective, the work is less about near‑term capability jumps and more about hedging the hardware stack. If today’s GPU‑driven approach hits economic or physical limits, bio‑hybrid computing could offer alternate routes with different trade‑offs in power, adaptability and opacity. It also forces the field to confront ethical questions much earlier than expected: at what point does a learning organoid used for computation warrant moral consideration, and how do we regulate a “brain‑in‑the‑loop” system that behaves like a device but might one day have proto‑cognitive properties? Even if organoid intelligence never competes on raw performance, the conceptual and regulatory groundwork being laid now will matter if we ever seriously consider biological components in AGI architectures.