The Podcast

Anjali Dhiman Hansen started out in computational biology, using neural nets and statistical models to find biomarkers for drug discovery. From there she moved into pharma, working on oncology launches in breast cancer and melanoma, and then into digital health and medical devices. Very few people I've talked to have seen innovation from the lab bench, the commercial launch, and the manufacturing floor. Anjali has real respect for all three, and her point about operations stuck with me: when a production line goes down, you can't sell anything.

The part of this conversation I pushed on hardest was experimentation. I talk constantly about being willing to try weird things, and early on it sounded like Anjali was arguing for less of it. She wasn't. Her argument is that experimentation in a regulated industry has to be disciplined. Every experiment needs to tie back to a clinical need you can put a number on and to a strategy leadership is actually funding. Otherwise the team burns its budget engineering a solution to a problem nobody has. In medtech, where you can prototype and iterate, that discipline keeps the handoff from discovery to product development from losing time. In pharma, where every step runs through clinical trials, skipping it costs even more.

Her framing of the two industries is the line I keep coming back to. Pharma struggles with speed and flexibility. Medtech struggles with scale and consistency. The future belongs to whoever can combine the strengths of both. We also get into AI, where her view is that a model is only as good as the data systems and accountability underneath it, and that AI in healthcare goes nowhere if clinicians don't trust it. If you run an innovation team inside a big organization, stick around for the last third. That's where she explains how a team keeps its funding after the executive who championed it leaves.

Takeaways

Here’s a teaser…

Find Anjali on Linkedin or reach out via email.

What You Missed on Sunday

Here’s what we covered in Sunday’s newsletter edition…

Amazon is shutting down Mechanical Turk on September 30. For 21 years, Mechanical Turk was where you went to buy human judgment in bulk at a few cents a task, and the people building modern AI were among its heaviest users. The image dataset that kicked off the deep learning boom was labeled by roughly 49,000 Mechanical Turk workers.

This week: what Amazon was actually selling for 21 years, why researchers called Mechanical Turk a market for lemons two years before ChatGPT arrived, and why the only grade the platform kept ended up rewarding the thing that killed it.

Here’s what you’ll find: