How GitLab's Co-Founder Built His Own Cancer Cure with AI

Let's go back to the year 2022. 

Sid Sijbrandij, co-founder of GitLab, the company that took developer collaboration to the cloud and IPO'd on NASDAQ in 2021, got diagnosed with a tumor. A few months back, he walked into the emergency room at 4 AM with severe chest pain.

Turns out, there was a six-centimeter tumor growing from his spine.

He was diagnosed with osteosarcoma. It's a rare, aggressive bone cancer that has fewer than a thousand cases a year in the US. He got through Radiation surgery, and for two years, the cancer stayed away.

Then in 2024, it came back.

And this time, his oncologist had a very simple message: "You've exhausted standard care. Maybe there's a trial somewhere. Good luck."

No trials existed for his specific situation. The disease was too rare, and he didn't meet the inclusion criteria. The system had run out of roads. So Sid quit his day job as GitLab's CEO and started building his own.

Here's what that looked like:

Run every test available. Don't wait to decide what to do with the data; just collect everything. He generated 25 terabytes of data on his own cancer. Single-cell RNA sequencing, bulk DNA sequencing, and organoid models. These were actual mini-tumors grown from his cells in a lab to test how drugs would respond to "his" biology specifically.

He assembled a team, approached biotech companies and academic researchers, and had multiple treatments under development simultaneously. This includes a personalized mRNA vaccine encoded with mutations unique to his tumor, and a CAR-T therapy engineered to attack his specific cancer without destroying his liver.

Jacob Stern, the geneticist who joined as operational lead for Sid's care, used ChatGPT and custom AI agents to run literature reviews, generate hypotheses from sequencing data, and analyze 600,000 individual cells from Sid's blood. A $20 API query did in 30 minutes what would have taken a specialist team weeks.

From all that data, one thing jumped out: his tumor cells were unusually high in a protein called FAP. A doctor in Germany had an experimental treatment combining FAP with radioactive substances. 

Sid flew there, underwent it twice, and spent days in an isolation ward as he was technically radioactive. The tumor shrank 20%. 60% of the cancerous tissue died. The cancer detached from his spinal membrane.

Today, Sid Sijbrandij has no evidence of disease.

Sid is a billionaire. Well, GitLab's IPO made him one.

He could hire computational biologists, fly to Germany for a treatment that isn't available in the US, or file individual patient applications with the FDA, which is a special pathway when no standard options exist.

Around the same time, a writer named Jake Seliger was battling advanced throat cancer. He didn't have the resources to hire a team, and he couldn't even get enrolled in the clinical trials that might have given him a shot. His wife was a physician at the Mayo Clinic. She was publicly furious at a system that denied a dying man access to experimental treatment he was willing to accept.

Jake Seliger died in 2024.

You see? Same era, same tools theoretically available, and entirely different outcome. The difference wasn't intelligence or will. It was money.

Whole-genome sequencing costs $500 today, RNA sequencing costs $50, and ChatGPT Pro costs $20 a month. The "tools" are getting cheaper, but the "system" around them is not.

AI didn't cure Sid. It gave him the leverage to think faster, ask sharper questions, and operate at a level of complexity that would otherwise have required a hospital system's worth of experts or years of waiting.

What he built was an "information system" applied to medicine with maximum data, parallel testing, rapid iteration, and radical documentation. This exact same architecture makes good companies. And for the first time in history, that architecture could also be used to outpace one's own death if one has the resources to deploy it.

The question isn't whether AI can personalize medicine. It clearly can. The question is whether that future belongs to everyone or only to those who can afford to build it for themselves.

Sid himself is trying to answer that question. He's publishing all 25 terabytes of his data publicly, starting companies to put this process on rails for others, and hiring people specifically to help other cancer patients navigate what he navigated.

But systems don't change because one generous billionaire documented his journey.

They change when the tools become so cheap, and the process so repeatable, that a person without a NASDAQ IPO can still use them. That's the version of this story we're all waiting for.