I got Gemini 3.1 to do this in 1 prompt.
ChatGPT is a bit behind on image generation.
Gemini and Grok are the leaders with images (for now, always changing).
“I Tried AI Once. It Was Useless.”
Good. Let’s talk about that.
If you tried AI once and gave up, you’re probably right about your first experience. Most physicians tested AI with a shallow prompt and no data. That produces generic advice, which is correctly ignored.
Latest post discussing how to start using AI in a small medical practice. open.substack.com/pub/superint...
It’s early, but clearly here (since GPT-5/Opus 4.5)
www.lesswrong.com/posts/XcrgeM...
the original Moore's law was every 18 months
How I Use AI to Appeal Prior Authorization Denials:
The prompt I use to quickly turn denial letters into structured, evidence-based appeals open.substack.com/pub/superint...
Is it time for a new standard of care when it comes to mammograms in the era of AI?
erictopol.substack.com/p/why-all-ma...
My writeup on how to build a custom AI scribe template, with examples from my rheumatology practice:
open.substack.com/pub/superint...
Medscape article I'm quoted in...
How Rheumatologists Are Making Smart Use of AI Platforms
www.medscape.com/viewarticle/...
AI Scribes: What the Evidence Shows So Far. open.substack.com/pub/superint...
The largest randomized trial of medical A.I.
—Over 100,000 women in Sweden
—radiologist + AI vs 2 radiologists, in follow-up
—AI added led to 29% more cancer detected, 44% reduced workload, and
—Less cancer dx in subsequent 2 years, and, when found, less aggressive
thelancet.com/journals/lan...
Latest post on Superintelligent Medicine:
Background Agents and Prior Authorizations in Medicine: Clawdbot gives us a glimpse of what might be possible
open.substack.com/pub/superint...
Latest post on Superintelligent Medicine:
How One State Is Testing Whether AI Can Actually Reduce Healthcare Costs
Utah just approved the first autonomous AI to legally prescribe medications.
open.substack.com/pub/superint...
Latest post in Superintelligent Medicine:
Understanding AI: Context Windows — AI's Working Memory
AI's working memory went from 6 pages to 1,500
open.substack.com/pub/superint...
Amazing! I also use Claude to do almost everything on my Mac now (some email, calendar, task lists, etc). I did s Dropbox cleanup too!
I gave Claude Coworker access to my computer 😬 and it cleaned up 15 years of files and found a ton of space on my computer. I feel so organized!
New post on Superintelligent Medicine:
Google Releases MedGemma 1.5: An Open Source AI That Can Read 3D Imaging.
And you can run it on your laptop. No data center required.
Full article:
Latest post on Superintelligent Medicine:
What OpenAI and Claude for Healthcare Actually Mean for Practicing Physicians open.substack.com/pub/superint...
New post: How to use AI to optimize your clinic template.
Most physicians have never analyzed their schedule.
I used AI to analyze mine and found ways to better align productivity and patient demand.
Full post + prompt: open.substack.com/pub/superint...
If you use Claude + OmniFocus, I *highly* suggest adding this connector/MCP server:
github.com/themotionmac...
Ask Claude to read everything in OmniFocus and make suggestions/organize things for you.
My next post on Superintelligent Medicine will be about how I used AI to optimize my clinic template.
I'll include the exact prompt you can use to do this on your own.
Subscribe: superintelligentmedicine.substack.com
Latest post on Superintelligent Medicine:
Guide to AI Prompting for Physicians
Link: open.substack.com/pub/superint...
Subscribe: superintelligentmedicine.substack.com
Just started a Substack about AI in medicine: Superintelligent Medicine.
First posts:
1. The Widening Gap Between AI Capability and Physician Adoption
2. How AI Works Without Actually Knowing Anything
Next post is on improving your AI prompting
Subscribe: superintelligentmedicine.substack.com
