درباره این اپیزود
AI can produce more in an hour than a nurse can properly read in a shift. So what happens when the output grows faster than the people checking it?
Garrett Hillman has spent his career in healthcare IT. He started as a Cerner consultant working across the US and the Middle East, and worked on the smart hospital projects where he and Russell first met in around 2014. In late 2022 he taught himself SQL by asking ChatGPT. Four years later he runs scheduled AI agents that work in the background while he gets on with his day.
In this episode, Russell and Garrett talk about why healthcare still doesn't trust AI and why the black box is the real barrier. They cover the extra person an AI quietly added to a colourised photograph, and why that should worry anyone deploying agents in clinical settings. Garrett explains the review bottleneck: AI works around the clock, and clinicians don't.
Then they get practical. Garrett shares the small agents that give him time back: school calendars, email follow-ups, and a news digest every three days. He explains what actually makes something an agent, how to think about memory before you build, and the Obsidian setup that keeps a profile of everyone he deals with.
His advice to his 16-year-old self: AI is here to stay, so learn to use it well, and "don't use the tools to just cheat your way through."
Guest: Garrett Hillman, AI Engineer
Hosted by Russell Mayne
رونوشت 🔗
00:00:19.760 --> 00:00:22.879
Hello and welcome to McClear Podcast.
00:00:23.120 --> 00:00:37.759
Today I've got uh Garrett Hillman, and um my Claude Research said that he is a CERNA-trained smart hospital engineer and turned industry agnostic data governance advisor, which is obviously 100% correct.
00:00:37.920 --> 00:00:40.640
But I've known Garrett uh and his wife for a long time.
00:00:40.799 --> 00:00:52.960
We met uh in the CERN office, I think in about 2014, when you when you walk…