AI Just Started Curing Cancer
A sweeping dispatch on how AI is compressing the cancer research timeline — from the first AI-designed drug to shrink a tumor to blood tests that catch cancer years before symptoms — and why the real weapon isn't a cure but a head start.
1 / 5at 3:38
Breast cancer and leukemia have less in common than the flu and a broken bone.”
AI Summary
Her tumor shrank by a third and stayed smaller for over six months after four failed traditional treatments — a milestone the host calls proof that 'the door everyone said was decades away is now open.' The video explains why cancer is so hard to beat: it is not one disease but hundreds, and tumors evolve and develop drug resistance, meaning treatment is effectively a race against fast-forward evolution.
Yank Summary
What matters, in Yank’s words, each linked to where the source says it.
Recursion's AI designed drug REC617 went from idea to patient in under 12 months, requiring only 136 candidate molecules instead of the usual thousands.
What we call cancer is an umbrella name for hundreds of different diseases; even within one patient, a tumour's cells mutate and evolve, allowing resistant cells to survive treatment.
Breast cancer and leukemia have less in common than the flu and a broken bone.
Shatters the assumption that "cancer" names a single enemy, making the failure of past treatments logical rather than scandalous.
In context
people hear the diagnosis. So why is the most funded war in medical history still ongoing? Because we declared war on an enemy that doesn't exist. What we call cancer is not one disease. It's an umbrella name for hundreds of different diseases. Breast cancer and leukemia have less in common than the flu and a broken bone. Different genes drive them and different weaknesses kill them. Cancer is not even one disease inside of one patient. A tumor is a living colony of cells that mutate and evolve. So when a drug wipes
Receipt
The failure rate has not fallen yet. It's just that the learning rate has exploded.
Reframes apparent stagnation as progress: speed of learning now compounds even before survival rates visibly improve.
In context
decade into months. Recursions drug went from idea to patient in under a year. Even if nine out of 10 still fail, which will not be the case in my opinion, the surviving medicine now arrives years earlier and every failure teaches the AI something that no human team could absorb at that speed. The failure rate has not fallen yet. It's just that the learning rate has exploded. Which brings us to the question, can AI cure cancer? The honest answer is much more than a simple yes or no. In 1995, a diagnosis of HIV meant you were dying. It was one of the deadliest
Receipt
For 50 years, cancer has been stealing our time. AI is how we start stealing it back.
Closes the source's central argument by casting AI as humanity's tool for reclaiming the time cancer has historically consumed.
In context
dismantling cancer's only real weapon. And that is the head start. Every scan it reads earlier, every switch it finds, every medicine it designs in months instead of decades takes another piece of that head start away. The woman from the beginning of the video didn't get a miracle. She got a medicine that brought her time. For 50 years, cancer has been stealing our time. AI is how we start stealing it back. The famous futurist Ray Kurszville once said that if you can live long enough, technology will arrive in time to save you. For most of history, that
Receipt
Doctors aren't fighting an illness — they're fighting evolution running on fast forward inside a human body.
Explains why no single drug can defeat all cancers — the target itself evolves inside each patient under treatment pressure.
In context
weaknesses kill them. Cancer is not even one disease inside of one patient. A tumor is a living colony of cells that mutate and evolve. So when a drug wipes out over 99% of it, the resistant 1% survives and grows back immune. That's why doctors are not fighting an illness. They're fighting evolution running on fast forward inside of a human body. That's why the drug from the opening of this video cannot simply be given to everyone. It picks one lock that only some cancers use. If you change the cancer, the lock changes as well. But
Doctors are not aren't fighting an illness — they're fighting evolution running on fast forward inside of a human body.
Receipt
An AI-designed personalized cancer vaccine cut the risk of melanoma returning by 50%, then passed its final stage human trial in 2026 — the first individualized cancer therapy to do so.
In context
a vaccine is made from scratch from that patient alone, teaching their body to hunt their exact cancer. In the melanoma trial, patients who received it alongside standard treatments saw the risk of their cancer coming back drop by 50%. That is half. And in 2026, it passed its final stage human trial. The first individualized cancer therapy in history to pull that off. This is the treatment wave. More than 100 AI designed medicines are now moving through human trials. They're aimed at everything from lung cancer to leukemia.
melanoma trial, patients who received it alongside standard treatments saw An AI-designed personalized cancer vaccine cut the risk of their cancer coming back drop melanoma returning by 50%, That is half. And in 2026, it then passed its final stage human trial in 2026 — the first individualized cancer therapy to do so.
Receipt
Keep going with this source
Create something
Turn this source into a post, a thread or a newsletter, every line checked against it.
See what you can create →Ask a question
Ask about anything in this video. We’ll try to provide an answer.