Insights / Topics
AI at the point of care: diagnostics, clinical decision support, ambient documentation, and the hospital workflows where models meet patients and clinicians. What actually changes practice versus what stays a demo.
These essays cut through the binary of "AI will replace doctors" versus "AI is useless at the bedside" to ask where clinical AI actually works and what it takes to trust it. The throughline: the technology is rarely the bottleneck — clinical inertia, broken systems, missing benchmarks, and the gap between a great product and a viable business are. The real shift is from answers to actions, with data standards and neutral, task-specific evaluation mattering more, and augmenting clinicians beating replacing them.
2 Jun 2026 · 3 min read
At ASCO this week, AI dominated the oncology agenda — but a Penn abstract pointed at a critical blind spot: the AI information cancer patients actually find online is low quality, hard to read, and silent on risks. The model can be clinically right and still fail at the bedside.
5 Mar 2026 · 3 min read
The FDA quietly granted Breakthrough Device Designation to a generative AI chatbot — not a diagnostic, not a scoring tool. An LLM that talks to post-surgical patients twice daily and escalates red flags. No GenAI device has ever received FDA authorization. This becomes the test case.
19 Feb 2026 · 1 min read
The AI-in-healthcare debate keeps swinging between “AI is going to take over everything” and “AI is useless in a medical setting.” Neither is useful. What the field actually needs are external neutral benchmarks on specific clinical tasks. Congrats to Rory Davidson and team for bringing one to market.
20 Jan 2026 · 3 min read
I just completed my first peer review — on a healthcare AI paper — under journal rules that explicitly forbid AI assistance. Peer review is slow, strained, and imperfect. And it now actively forbids the tools that could make it better. Will this model survive the next decade?
13 Jan 2026 · 3 min read
Everyone is framing the new healthcare AI launches as ChatGPT vs Claude. That’s the wrong comparison. The real story is consumer memory vs reimbursement workflows — two completely different bets on where AI lock-in actually gets built.
21 Oct 2025 · 1 min read
Many think GenAI will replace data standards — I argue the opposite: for AI in healthcare to be trustworthy, GenAI makes standards like SNOMED and LOINC more critical than ever.
25 Sept 2025 · 1 min read
The NHS is building AIR-SP, a national AI screening cloud that lets every trust plug into shared, real-world trials — starting with a breast-cancer study involving nearly 700,000 women. I think the overall approach makes sense, and I'd love to see other countries adopt something similar.
16 Sept 2025 · 1 min read
Healthcare AI is moving from answers to actions — and model evaluation must shift from simplistic multiple-choice testing to comprehensive workflow safety and governance. I'm convinced that agentic workflow automation will be a massive opportunity in healthcare and life sciences.
28 Aug 2025 · 1 min read
Consolidation in healthcare AI is inevitable, but the $81M acquisition price of Paige AI by Tempus AI — after raising over $260M and once valued at $600M — is a sobering reminder that building a great product and figuring out how to monetize it are two very different challenges.
8 Jul 2025 · 1 min read
A recent Microsoft study suggests AI outperforms doctors at diagnosis, but only under artificially constrained conditions — the real story is about AI augmenting physicians, not replacing them. The greatest barrier to better patient outcomes isn't physician skill or technology potential; it's the clinical inertia and outdated infrastructure keeping them apart.
12 Jun 2025 · 1 min read
One of the biggest paradoxes in health data is that the most valuable insights are locked within the most sensitive and difficult-to-access information — a natural brake on innovation. I explore how the new Atropos Health and Databricks partnership takes a meaningful step toward solving this.
27 Mar 2025 · 1 min read
Bill Gates predicts AI will replace many doctors and teachers within 10 years — but I think he may be underestimating how hard it is to fix systems that are already deeply broken. The technology will be ready; whether our dysfunctional healthcare systems will truly change is another question.
20 Mar 2025 · 1 min read
We're increasingly relying on AI and LLMs in medical settings — from self-diagnosis to clinical co-pilots — but there are still no good benchmarks for how well these models actually perform in real clinical practice. I think we have a long way to go before LLMs become truly reliable thought partners in our medical journeys.
18 Mar 2025 · 1 min read
AI in healthcare shouldn't aim to replace professionals — its true potential lies in enhancing their ability to deliver quality care. I explore how Wysa's merger with April Health exemplifies the broader industry trend of combining human empathy with AI-enabled tools in mental healthcare.
11 Mar 2025 · 1 min read
Even Epic — not historically known as an innovation powerhouse — is now fully embracing AI agents. I share highlights from an interview with Epic's SVP of R&D and offer my take on what their commitment to AI and open collaboration means for healthcare.
4 Mar 2025 · 1 min read
I've written several times about the power of AI agents, and I remain convicted that they will bring a fundamental shift in the way we work — including in healthcare — in the next five years. Healthcare may be sleeping on this shift, but the time to start exploring Agentic AI is now.
22 Jul 2024 · 1 min read
I've been following the debate around whether enabling patient interactions with GenAI is a net positive or negative — and a recent study suggests machines can sometimes outperform overworked hospital staff on empathy, understandability, and tone.
19 Jul 2024 · 1 min read
CB Insights has just released a report on the AI readiness of different hospital systems in the US, and I'd love to see a similar study done for Europe. From my experience working with startups in the field, France, Spain, and several Scandinavian markets appear ahead of places like Germany or Switzerland.
1 Jul 2024 · 1 min read
AI agents — systems that don't just converse but actually do things for you — are the next frontier after GenAI, and I predict we'll see the first pilots for administrative tasks in healthcare within the next 2–3 years.