MMost of today’s discussions about generative AI (GenAI) and large-scale language models (LLM) focus on their promise, shortcomings, and the speed of their evolution. For many users, writing, summarizing, translating, brainstorming, coding support, answering questions, and more are already part of their daily routine. These are not marginal uses. They are reshaping the way many of us think and work. For example, OpenAI CEO Sam Altman reportedly said that AI will increase programmer productivity by up to 10 times. But there are also stories that suggest something more important than simple productivity gains. They suggest that in the right hands, these systems can serve as bridges between disciplines. Otherwise it will remain inaccessible.
Paul Conyngham’s story and his dog Rosie is one of those cases.
Conyngham was not a doctor, biologist, or chemist. He was an Australian AI consultant and entrepreneur who was faced with a challenge that was far removed from his formal training. Their dog Rosie was diagnosed with terminal cancer. Rosie underwent multiple surgeries, chemotherapy and immunotherapy, but these treatments only slowed the progression of the disease, according to the University of New South Wales, which was supported by Conyngham researchers, and subsequent scientific reporting. Veterinarians told Conyngham that time was limited, with estimates ranging from one to six months.
But rather than treating that boundary as final, he decides to look beyond it and into AI, motivated by his deep personal bond with Rosie. What happened next wasn’t a story of “AI solving cancer.” But the series of human decisions supported by AI proceeded through real institutions, real laboratories, and real scientists.
After sequencing Rosie’s tumor, Conyngham used ChatGPT and AlphaFold to design a neoantigen and, through collaboration with scientists, ultimately developed Rosie’s personalized mRNA cancer vaccine. Ultimately, Rosie received a vaccine combined with an immune checkpoint inhibitor. Although the result was not a cure, her response was very good. Her largest tumor has shrunk and her mobility has improved.
That’s why this case deserves attention. What is impressive is not that AI has replaced scientists, but that one determined person was able to use AI as scientific support, connect with experts who could carry out critical steps, and keep the entire process moving forward even under pressure. The key interactions were not human versus machine, but human versus GenAI/LLM, human versus sequence, human versus AlphaFold, and human versus scientist.
This also helps clarify things that we often overlook. We typically think of expertise as deep domain knowledge, and rightly so. However, there is another form of expertise that is very important in the real world. It’s the ability to quickly coordinate knowledge, tools, institutions, data flows, payments, and decisions to take action. Public reporting tends to show only the visible path. It doesn’t show the failed attempts, wrong turns, interactions with scientists and research centers, emails, sequencing costs, software workflows, lab logistics, or the ethical and administrative steps needed to go from idea to actual treatment.
Also, the most important constraint, time, is not shown.
It was not a situation that people could think about forever. Analysis must be quick, but not careless. Decisions must be timely and evidence-based. In such environments, a combination of human judgment and machine-supported knowledge can make a big difference. Contributes to persistence, responsibility, coordination, and action in the real world. The other contributes to speed, analytical support, and access to vast knowledge. One does not replace the other. Each amplifies the other.
That’s what makes Conyngham and Rosie’s story so memorable. It’s not just about technology. This is a story about what is possible when human initiative meets strong scientific support of humans, machines, and organizations to achieve meaningful goals.
