Jeff Sutherland

Spitball Sunday: $7B to Reverse Disease With AI. You can see that as either a spitball OR reality via ACCESS While not once in the RFA or rate document does CMMI mention AI or LLMs, CMMI leadership was quoted yesterday saying something that made it’s intentions clear. "Develop NOT for RPM economics, but leverage LLMs to help patients self-manage. If you've already got something built, think about the automation and self-management side." They want AI-powered self-management. At scale. For $1.15 a day. Think that's the worse idea ever, well as Pryce Ancona and Brendan Keeler have said - AI is the worse it's ever going to be. And 3 weeks ago, in the course of a few hours, I was able to download 10 years of health data, generate 3 of comprehensive health eval and optimization plans for myself tapping into 4 different world reknowned physicians. The reports blew away anything I have received from tens of thousands of $$ spent trying to get to the bottom of my health. Inclulding traditional MDs, functional docs and concierge docs. As for the ACCESS rates, it’s as if ŌURA sat down with Oz over dinner and sketched out the economics they needed to participate in access – Initial year - $200-$250 to cover the cost of the hardware and double the current $5.99/month to cover the cost of an amped up version of their current nudging. Follow-up years – no more hardware cost, just the cost of nudging. ACCESS is many things: A System C Design Breakthrough Pot of Gold An Opportunity to Capture Payment from Medicare for Direct to Consumer Disease Reversal Solutions Not Reimbursed Today by Medicare A Golden Opportunity to Show Consumer Latent Demand for a Desire to Reverse Disease NOT Simply Treat It “Better” An Opportunity to Bridge the Gap Between the Grocery Store and Healthcare NOT through FIM but through Food Is Health – Leveraging the Assets that Exist Today and Harnessing the Tech that Healthcare (rightly so) Fears and Views as Perilous A System B Slap in the Face A Tech Bros Best Day Ever An Expensive Dumpster Fire (and Possibly the Biggest One in CMMI History) I see it as a design challenge for the Good Guys. For CKM, it’s $420/month x 21.5 million beneficiaries with a CKM condition. The incumbent healthcare delivery system sees that as a gnat on the elephants ass, but others don’t. This is actually an unexpected opportunity for CPG and retail grocery to take a new position in healthcare with their customers. AND get paid for it. If the rates hit a nerve. Good. We started mapping the use cases for this last week in Denver at our first ever Food Is Health Unconference (Carter Williams). We had farmers, CPGs, retail grocery, healthcare and more designing how we could end the chronic disease epidemic. If you didn't hear about it, you must not be over on our foodishealth substack. It's GO TIME for what Carter Williams and I call System C. The food industry has the opportunity to take back a bunch of marketshare from healthcare.

24

12 Comments

Quaternion Process Theory Process-Oriented Thinking: A QPT Approach - Transcript Description This is a deep-dive discussion between two hosts exploring revolutionary concepts from Carlos e Perez's work on Quaternion Process Theory (R-M-Q-P-T). The conversation challenges conventional "noun thinking" and introduces a dynamic, verb-based approach to understanding reality and cognition. The discussion presents process-oriented thinking as both a philosophical framework and practical tool for enhanced creativity, problem-solving, and cognitive flexibility. https://lnkd.in/eH_kb27n

1

For everyone who is playing around and trying to get agentic frameworks and systems in production, here is a great cheatsheet on Context engineering. https://lnkd.in/eqjTNbNU And the 12 Factor Agents. Patterns let me remember cohesive concepts, so I really appreciate them https://lnkd.in/e7GHqYUq

11

"I have only made this letter longer because I have not had the time to make it shorter." - Blaise Pascal'' This also applies to software demos. We've tightened the onboarding workflow and now have a 60 second demo of the journey from a blank slate in Claude Code to a deployed preview application on VeilStream.

1

Claude Cowork applies the same agentic architecture that powers Claude Code, but inside Claude Desktop, so you can delegate multi-step knowledge work without living in the terminal. This article explores what Cowork is, how it differs from chat and Claude Code, and how to extend and use it responsibly. The full article is in the first comment below.

11

1 Comment

New blog post about subsetting a ligature-heavy font https://lnkd.in/gkNdkmSt Almost always we subset based on text. But in my case, a more fine-tuned glyph-based approach saves 91% of bytes as opposed to 74% in the usual way. Worth it?

21

Anthropic researchers have unveiled a new gauge of AI job-destruction risk called "observed exposure," which Axios describes as "an early-warning system" for potential disruption, particularly among white-collar workers. So far, Anthropic maintains there is "limited evidence" that artificial intelligence is a drag on employment. But it notes those already affected tended to be "older, female, more educated and higher-paid." Separate research shows employment levels decreasing for younger workers, which Anthropic attributes to a slowdown in hiring for early career professionals in AI-exposed fields.

5

4 Comments

In healthcare, almost all of the LLM performance evaluation research has been dedicated to studying clinical correctness. But that leaves another important domain out the picture: interoperability performance. At FHIR DevDays Flexpa launched an LLM Eval for FHIR, where we focused on key interoperability tasks like generating and transforming FHIR resources. LLMs are, perhaps surprisingly to some, extremely adept at understanding and manipulating FHIR, separate from the question of terminology accuracy. With access to the $validator operation, an agent can generate de novo FHIR resources from short natural language descriptions of healthcare events.

18

NYTimes Opinion - About the impact of AI on Critical Thinking. https://lnkd.in/g3ZAyPWT My concern is broader: the use of AI and basic expectations of rational thought. It seems that the magic of AI and non-sequitur magical thinking go hand in hand.

13

4 Comments