The Salamander Factor
Scenario #3: Discovery — The Solve-Everything Decade
In high school, I came across a short story in our “Rapid Reader” series called “The Salamander Factor.”
A gifted surgeon loses the use of his hands. His identity, his purpose, everything he’d built, gone. Then, inspired by the salamander’s extraordinary ability to regenerate lost limbs, he finds a way to rebuild his own capabilities from scratch. The story stayed with me because of the question it planted: what if biology itself wasn’t the ceiling?
That was fiction then. I’m not sure it is anymore.
Paul Conyngham, a tech entrepreneur in Sydney, learned his dog Rosie had cancer. Tumors persisted through chemotherapy and surgery. So he turned to ChatGPT, which pointed him toward immunotherapy and the University of New South Wales. He used AlphaFold to identify mutated proteins. He convinced researchers to help, funded the genomic sequencing himself, and within two months a bespoke mRNA cancer vaccine had been designed and administered. Rosie’s tumors shrank by 75%. Six weeks post-treatment, she jumped a fence chasing a rabbit.
This is the first personalized cancer vaccine ever designed for a dog, and the learnings will inform human medicine.
The reaction from the people building these tools was immediate.
Greg Brockman, President of OpenAI, highlighted the case publicly: AI had empowered Conyngham to create a custom mRNA vaccine when Rosie had only months to live.
Demis Hassabis, CEO of Google DeepMind and Nobel laureate whose AlphaFold modeled the mutated proteins that made the vaccine possible, called it an early example of “digital biology.”
Aravind Srinivas of Perplexity called AlphaFold one of the greatest outcomes to come from AI.
And Kevin Weil, OpenAI’s CPO, has been building what he describes as “the next great scientific instrument: an AI-powered platform that accelerates scientific discovery.” Rosie is exactly the kind of case that platform exists to make possible.
What strikes me is the architecture of how this happened.
A non-medical engineer, armed with AI tools and genuine urgency, walked into a frontier that credentialed institutions hadn’t yet reached. Not because the institutions lacked capability, but because no one had reason to combine these particular pieces in this particular moment for this particular patient.
That’s the signal. AI doesn’t just accelerate what experts already do. It compresses the distance between a motivated human and a breakthrough that would otherwise require years of institutional scaffolding.
The salamander doesn’t wait for permission to regenerate. It just begins.
We are entering a period where the barriers between “patient” and “researcher,” between “problem” and “solution,” between aspiration and capability, are collapsing faster than our frameworks can track. The human premium here isn’t diminished by AI. It’s what activates it. Conyngham’s judgment, his refusal to accept a terminal framing, his willingness to learn and connect and ask, that’s what made the tools matter.
Personalized medicine. Personalized learning. Personalized infrastructure. Each a version of the same idea: intelligence, made specific to you, at the moment you need it.
The salamander factor isn’t fiction anymore.