What do you most want to learn about applying AI in healthcare? Don't miss Chris Lovejoy, MD and I chat about it next week! x.com/i/spaces/1vOxw…

The trouble with evals is the “theory of change” most companies have is “whatever the CEO thinks today” There’s cases where this works great. But 99%+ CEOs are not nearly dialed in enough for that. They change direction because their tummy hurts after a breakfast burrito

Don't forget to look at your data tomorrow.

Doctors don't want your noisy AI alerts. So what do they want? Don't miss Chris Lovejoy, MD and I chat about it next week! x.com/i/spaces/1vOxw…

Two hours away... Don't miss Chris Lovejoy, MD and I chat about it next week! x.com/i/spaces/1vOxw…

Happening in a few mins! x.com/i/spaces/1vOxw…

x.com/i/spaces/1vOxw…

80% of my clients implement a complex opensearch solution with multiple retrieval passes just for me to show them bm25 outperforms in quality and speed

Me: tab Cursor: "oh you want vomit?" Me: <disables cursor tab> Cursor: <disables tab> yeah sorry Cursor we goin' back to Visual Studio Code today.

The kids are not alright

But also if everything is a priority, nothing is 🤷

sorry nerds, gpt5 is actually pretty good. skill issue.