Imagine this scene: late at night, you're curled up on the sofa with the old dog that has accompanied you for ten years. The vet has just delivered bad news—cancer. You stare at the ceiling, and a thought flashes through your mind: what if... what if I could design a treatment plan for it myself?
It sounds like science fiction, but on March 28, 2026, this scenario actually happened.
The Beginning of the Story
Paul Conyngham is an ordinary AI user. When his dog was diagnosed with cancer, instead of giving up, he opened a familiar interface—ChatGPT.
Yes, the same ChatGPT you type questions into on your phone.
Paul started conversations with the AI, asking about mRNA vaccine knowledge, the principles of cancer immunotherapy, and personalized medicine design. Like a diligent student, he was guided step by step by the AI—acting as a patient mentor—through complex biomedical concepts.
Eventually, Paul designed a personalized vaccine scheme based on mRNA technology. Whether it ultimately cured his dog, we don't know. But when Sam Altman shared the story, the discussion it sparked went far beyond "the fate of one dog."
Why mRNA?
To understand the significance of this story, we first need to know what mRNA vaccines are.
Imagine your body as a city, and a virus or cancer cell as an invader. A traditional vaccine is like giving the city wanted posters of the invader—letting the body recognize the enemy in advance. An mRNA vaccine is different: it's like handing the city a "manufacturing manual" for the invader. After your body reads the manual, it produces the enemy's "signature fragments" itself, then mounts a targeted defense.
The beauty of this technology lies in its flexibility. Traditional vaccine production—culturing the virus, inactivating it, purifying it—can take months. An mRNA vaccine only needs the gene sequence of the target protein, and can be designed and produced within weeks.
This is exactly what made it possible for Paul to "design" the scheme—he didn't need a lab or cell cultures. He only needed knowledge and imagination.
What Role Did AI Play?
Some might ask: did the AI really do anything special? Isn't all this information already online?
Yes, the information is online. But the ocean of information is chaotic. Imagine walking into a library where shelves stretch from floor to ceiling, every book about cancer treatment. Where do you even begin?
AI's role was that of a guide in this library. It helped Paul:
- Filter the core concepts related to mRNA vaccines from massive amounts of information
- Translate obscure medical terminology into understandable language
- Sort through the pros and cons of different treatment strategies
- Build a logically coherent framework for the plan
- Sam Altman's original share: https://x.com/sama/status/1895206925259778440
This isn't simple information retrieval—it's the structuring and re-creation of knowledge.
The Deeper Meaning: A Regulatory Gap and a Laboratory for Innovation
The most interesting part of this story may not be the technology itself, but where it happened.
In human medicine, mRNA vaccine development is strictly regulated. Clinical trials, approval processes, safety assessments—every step takes years and enormous funding. This system protects human safety, but it also slows the pace of innovation.
In veterinary medicine, regulation is relatively loose. Paul didn't need FDA approval or ethics committee permission. He could directly try his ideas.
This reveals an interesting paradox: sometimes, fields with fewer constraints become the testing grounds for frontier technologies.
Consider a different angle: if Paul's scheme succeeds, it might not only save his dog, but also provide valuable empirical data for human medicine. Veterinary medicine becomes a "sandbox environment" for personalized medicine—where ideas can be validated quickly and successes quickly replicated.
The Feynman Moment: Explaining It in One Sentence
If Richard Feynman were alive today, he might explain it like this:
"Imagine your dog is sick, and you have two options: wait for a doctor to provide a standard plan, or learn enough yourself to design a customized one. AI makes the second option possible—it compresses knowledge that once required ten years of medical education into a few evenings of conversation."
Imagining the Future
Paul's story is an individual case, but it points toward a possible future.
In that future, when someone is diagnosed with a rare disease, a doctor might work with an AI to design a personalized treatment plan based on the patient's genotype within hours. The plan might combine mRNA technology, immunotherapy, and targeted drugs—all components tailor-made.
Of course, this still requires substantial technological maturity and regulatory refinement. But Paul and his dog have already taken the first step.
---
Related links: