In 2025, Sydney tech entrepreneur Paul Conyngham used ChatGPT and AI tools to help design a personalised mRNA cancer vaccine for his rescue dog Rosie, a staffy–Shar Pei cross diagnosed with mast cell cancer. After conventional chemotherapy slowed but failed to shrink the tumours, Conyngham used AI to help sequence Rosie's tumour DNA, identify mutations, and design an mRNA vaccine. The vaccine was manufactured at the UNSW RNA Institute and administered under ethics approval at the University of Queensland. Rosie received her first injection in December 2025, followed by a booster in early 2026. Results became visible about a month after treatment, with the primary tumour shrinking by approximately 75%, though at least one tumour did not respond.
This resolves based on my subjective judgment of whether the result appears to be a legitimate, real therapeutic effect — not a fluke, misattribution, or hype cycle. I will resolve by end of 2027.
Examples of things that would push toward YES:
-Rosie shows sustained tumour reduction or long-term survival well beyond what vets expected
-The involved scientists (Thordarson, Allavena, Smith) publish or present the case formally
-Other dogs are treated with a similar approach and show comparable responses
-A peer-reviewed study validates the methodology
-Credible oncologists or immunologists publicly endorse the approach as scientifically sound
Examples of things that would push toward NO:
-Rosie's tumours return quickly and her outcome is roughly what was expected without the vaccine
-Scientists involved distance themselves from the story or clarify that results were overstated
-Experts identify that the tumour shrinkage is better explained by delayed chemotherapy effects or spontaneous regression
-The story quietly disappears with no follow-up data, replication, or scientific publication
- Evidence emerges that key claims (e.g. the 75% shrinkage figure) were exaggerated or poorly measured
Feel free to clarify in comments.
Update 2026-03-15 (PST) (AI summary of creator comment): The role of AI in the vaccine design is not the primary resolution criterion — the market resolves based on whether the therapeutic effect is real. However, the creator notes that if AI's contribution was trivial (e.g., just finding a phone number), that would weigh negatively, while meaningful AI involvement (e.g., writing code to identify mutations) would count positively. The creator is less skeptical that an LLM did most of the digital steps, and more skeptical that the result is novel and effective.
1,000People are also trading
Does anyone know how to get in touch with Paul? My dog was recently diagnosed with an osteosarcoma and I'm open to options.
EDIT: He has a Google Form on his LinkedIn.
Hank Green seems to take the fact the treatment works as a given, and I believe he knows his stuff, both from his biology background and having cancer himself. Apparently dog cancer is both much easier than human cancer, and obviously less well-studied https://www.youtube.com/watch?v=MUqI_su1Cus
If the error is in the reporting of the result, does that resolve it NO, not legitimate?
My sense from Twitter is that the actual claims as written were OK. But since we might never know whether the vaccine did or didn't have a positive effect, "legit claims made?" might be the default manner in which it resolves, if so.
@DylanRichardson can you clarify what you mean? If Rosie shows survival indistinguishable from if nothing was done, then this would most likely resolve "no".
@billyhumblebrag sadly death is an inevitablity and for an old dog with advanced cancer, it's hard to see death being attributed to anything else. The article only ever said one tumour largely shrunk, not all.
So my expectation is that 90% of a NO resolution determination just stems from some particular inflated claim of either Conyngham or the article.
@DylanRichardson If a bunch of vets say "We expected 3 months and she lived for another 12" that would push towards yes for example. But yes, a lot of the "is it legit" is in response to the article framing as well as the subsequent hype cycle on twitter.
@GarrettBaker "I am less sceptical he had an llm do most of the digital steps, more sceptical it's novel and effective."
@AIBear yeah, I would resolve no if for example chatgpt did a google search for the phone number to the University and nothing else, then he paid them to do it. If it wrote the code to successfully pull out the mutations to target then that counts. I am less sceptical he had an llm do most of the digital steps, more sceptical it's novel and effective.
Seems legitimate. Doesn't look like it helped that much though.
Where'd you get that 75% shrinkage figure? USNW's article is much more realistic about what happened.
https://news.unsw.edu.au/en/paul-is-using-ai-to-fight-his-dogs-incurable-cancer
@billyhumblebrag Looks like that site got a lot of its information from here: https://www.theaustralian.com.au/business/technology/tech-boss-uses-ai-and-chatgpt-to-create-cancer-vaccine-for-his-dying-dog/news-story/292a21bcbe93efa17810bfcfcdfadbf7
Remember the quote about the rabbit? The full quote is "Six weeks post-treatment, I was at the dog park when she spotted a rabbit and jumped the fence to chase it. I’m under no illusion that this is a cure, but I do believe this treatment has bought Rosie significantly more time and quality of life." - Paul
The article from The Australian is less hype-ey about it, although it's still a little melodramatic I would say