Discussion about this post

User's avatar
Rajasrinivas Adaikalam's avatar

The convergence of AI and synthetic biology is fascinating. But I think the next major challenge goes beyond designing living systems — it is making their behaviour predictable, traceable and verifiable.

Biology may become programmable, but unlike software, living systems evolve, interact with their environment and can produce unexpected outcomes.

We need an architecture that connects biological intent → models → constraints → experiments → evidence → validation, with explicit boundaries between what AI predicts, what experiments establish, and what we can safely authorize.

This is a direction we're exploring with GraphModel: making domain knowledge, constraints, provenance and evidence computationally explicit, rather than treating an AI-generated design as inherently trustworthy.

The real breakthrough will be when we can not only design living systems, but also establish the conditions under which their behaviour can be trusted.

Lesetja R T Mogoba's avatar

A tardigrade edited to have a nitroplast and/or chloroplast

40 more comments...

Ready for more?