One man's chemo diary became a case study for AI's real superpower — not curing disease, but giving patients their agency back.
The Summary
- Russ Read-Barrow, 43, uses Claude to build a symptom tracker for his stage 4 colon cancer treatment, logging what helps during chemo cycles he repeats every three months
- He prompts AI to generate medical history summaries for new doctors, cutting the cognitive load of retelling his story while exhausted
- His side project FcancerwithAI documents the tools he's built, as colon cancer becomes the deadliest cancer for people under 50
The Signal
Read-Barrow isn't asking AI to diagnose him or replace his oncologist. He's using it to solve a problem most healthy people never think about: when you're sick for years, memory becomes unreliable. Which foods settled your stomach during round three of chemo? What time of day did the nausea peak? Did that anti-nausea med actually work, or did you just think it did?
For someone cycling through chemotherapy every three months since 2021, those aren't academic questions. They're the difference between functional days and lost ones. So he built a symptom tracker in Claude. Not a fancy app. Just structured prompts that help him log patterns and recall what worked when the fog rolls in again.
"I'm saying it's going to help me remember what my consultant said, spot patterns in my side effects, and maybe give me back some control in a situation where control is basically an illusion."
The second use case hits different. When you're meeting a new specialist or switching hospitals, you have to summarize five years of cancer treatment while exhausted, scared, or both. Read-Barrow prompts AI to generate those summaries from his logs. It's not about accuracy, it's about cognitive load. The AI doesn't forget. It doesn't get tired. It doesn't leave out the detail that might matter.
The third angle is harder to measure but maybe more important: he's building in public. His FcancerwithAI blog documents the tools and prompts he's created. Colon cancer diagnoses in people under 50 have climbed since the mid-1990s. It's now the leading cancer killer in that age group. Early detection changes survival rates dramatically, but if you're already in treatment, you're playing a different game.
What Read-Barrow is doing isn't about frontier AI or AGI. It's about something more mundane and more powerful: turning an LLM into a memory prosthetic. Cancer treatment isn't a sprint. It's years of incremental decisions where small optimizations compound. Did this help last time? What did the doctor say six months ago? What pattern am I missing?
The Implication
The agent economy isn't just about automating code or customer service. It's about giving people tools to manage complexity they shouldn't have to carry alone. Read-Barrow's trackers and summaries are early examples of personal AI infrastructure, the kind of thing that becomes invisible once it works.
If you're building AI tools, this is the bar: did you give someone back time, clarity, or agency they didn't have before? The cancer won't care. But the person living with it might get a few more good days. That's the only metric that matters.