The CEOs selling you AGI are using cancer patients as props in a funding pitch.
The Summary
- Emilia Javorsky, director of the Futures program at the Future of Life Institute, published an essay challenging the narrative that we need artificial superintelligence to cure cancer, arguing that intelligence isn't the bottleneck in cancer treatment.
- Over a trillion dollars has been invested in AI, with Meta, OpenAI, and others explicitly citing cancer cures as justification for pursuing AGI/ASI.
- AI can't analyze patient data that was never collected, and breakthrough treatments mean nothing if patients go bankrupt accessing them.
- Current AI is already making meaningful contributions to cancer care, but the real barriers are systemic, not computational.
The Signal
Javorsky, who brings the rare trifecta of doctor, scientist, and entrepreneur to this conversation, is calling out a convenient lie: that cancer remains unsolved because we haven't built smart enough computers yet. It's the kind of techno-solutionism that sounds profound in a keynote but crumbles under clinical reality.
The cancer-cure promise has become the go-to rhetorical move for AGI advocates. It's emotionally bulletproof. Who's going to argue against curing cancer? But Javorsky's reframe cuts through the performance: the problem isn't that we lack intelligence, artificial or otherwise. The problem is that we lack data infrastructure, equitable access, regulatory frameworks that move faster than glaciers, and economic models that don't bankrupt patients.
"AI cannot analyze patient data that was never collected, and any treatment is flawed if patients risk bankruptcy seeking it."
Consider what actually blocks cancer progress today:
- Fragmented health records that don't talk to each other
- Clinical trials that exclude 97% of cancer patients due to eligibility criteria
- Treatments that cost $150,000 per year that insurance may or may not cover
- Research incentives that reward publication over patient outcomes
None of these problems get solved by making GPT-7 better at reasoning. They're systems problems, political problems, economic problems. An ASI could theoretically design a revolutionary cancer treatment tomorrow. Then what? It sits in regulatory review for eight years while patients die. It gets priced at $2 million per treatment. It requires genetic data that only exists for 3% of the population, all of them white and wealthy.
Javorsky's essay also highlights what current AI is actually accomplishing in oncology right now, without waiting for superintelligence. Image recognition models that spot tumors radiologists miss. Predictive models that identify which patients will respond to immunotherapy. Natural language processing that pulls signal from millions of unstructured clinical notes. These tools aren't sexy. They don't get TED talks. But they're helping real patients today.
The AGI pitch relies on a sleight of hand: conflating "hard problem" with "problem that requires more intelligence." Cancer is extraordinarily hard. But so is getting the FDA to approve a drug in less than a decade. So is convincing insurance companies to cover novel treatments. So is building electronic health record systems that actually share data. None of those are intelligence problems.
The Implication
If you're building in the agent space, this matters. The cancer-cure narrative isn't just marketing copy. It's shaping where capital flows, what regulators prioritize, and which problems get attention. When trillion-dollar companies promise that ASI will solve our hardest challenges, they're implicitly arguing that incremental progress with today's tools doesn't matter. That's dangerous.
The real opportunity isn't waiting for superintelligence. It's deploying narrow AI against the unglamorous infrastructure problems that actually bottleneck progress. Agents that coordinate patient data across hospital systems. Models that accelerate regulatory review by flagging safety signals earlier. Tools that make clinical trials accessible to the 97% currently excluded. Build for the cancer patients who exist now, not the hypothetical beneficiaries of ASI in 2035.