Jeff Mignon (@jeffmignon) header
Jeff Mignon (@jeffmignon) profile picture

Entrepreneur, CMO & Product Innovation Leader | Systems Practitioner | Guiding Startups to Market Fit | Scaling Mid-Market Firms | Board Member

New York, USA
Joined May 2008
Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)Photo by Jeff Mignon (@jeffmignon)
Jeff Mignon retweeted
CME Dinner Tomorrow. Which path will YOU choose? Join us in Philadelphia for an interactive Choose Your Own Adventure™ CME dinner symposium. Vote LIVE as expert faculty guide you through real-world patient cases, evolving treatment options, and critical decisions that shape TED care. Don't miss this unique CME experience! Register for the in-person event or virtual broadcast: ow.ly/GuXR51018FC A CME/CE-accredited activity developed by GLC. #TED #ThyroidEyeDisease #MedEd #CME
1
1
47
Jeff Mignon retweeted
Most think vibecoding at scale is slop. Here’s a counterexample. Today we release a 16M-line formalization of the entire Stacks Project, critical infra underlying modern math, with optimized code quality and its highly interconnected structure preserved. A mathematical city 🌆
844
1,551
310
7,028
6,902,268
Jeff Mignon retweeted
AI in healthcare is moving beyond transcription. I spoke with Dr. Inderbir Gill about using AI to prepare patient histories, organize records, and streamline outpatient visits while preserving physician oversight. Listen to The Convergence: reachmd.com/u/20358
#HealthcareAI
1
2
4
280
En effet. Imparable.
Imparable. A écouter jusqu’à la fin
33
Jeff Mignon retweeted
Imparable. A écouter jusqu’à la fin
Intifada à Paris VS révolution en Iran … voyons qui n’est pas prêt pour la démocratie et qui peut donner des leçons de civilisation et de patriotisme.
1
7
1
11
1,410
Such an amazing conference. And so great to see friends in the Silicon Valley
33
Jeff Mignon retweeted
A retired NASA engineer who had worked on the Space Shuttle program for 25 years was hired by SpaceX at age 62. He described the first week as the most disorienting experience of his professional life. At NASA, a design change required 3 meetings, 2 committees, a review board, and 6 to 18 months of documentation. At SpaceX, his new supervisor said "if it works, build it. If it doesn't, change it by Friday." He sat at his desk after the first day and called his wife. He told her he felt like a man who had been swimming with ankle weights for 25 years and someone just cut them off. He could move. He could decide. He could build something in the morning and test it in the afternoon. The constraint that he had accepted as the nature of aerospace engineering turned out to be the nature of bureaucracy. Within 6 months he had contributed to more design iterations than he had completed in his last 5 years at NASA. Not because he was working harder. Because the time between having an idea and testing the idea collapsed from months to hours. The same brain. The same knowledge. A different operating system. He described a specific moment where he proposed a modification to a Raptor engine component. At NASA, the proposal would have entered a queue. At SpaceX, the modification was machined by the end of the day and tested by the following morning. He watched his idea become a physical object in 18 hours. At NASA, 18 hours would not have been enough to schedule the first meeting to discuss whether to schedule a meeting. He told an interviewer: "I didn't get smarter when I went to SpaceX. I got faster. The intelligence was always there. The system was in the way." The difference between SpaceX and NASA is not talent. NASA has extraordinary talent. The difference is the time between a thought and a test. SpaceX compresses it. NASA elongates it. The same engineer produces 10x more at SpaceX than at NASA because the bottleneck was never the engineer. It was the meetings about the meetings about the work.
2,129
13,444
1,607
90,570
10,683,663
Meta shipped Muse. Personal agents are real. Give it your card, and it buys. A new customer that isn’t human, and doesn’t shop like one. It reacts in seconds and at the same moment as millions of others. Nobody has a playbook for that.
1
33
Jeff Mignon retweeted
55
285
33
2,006
50,681
You can’t control Complex Adaptive Systems (CAS). You can constrain them. Not control. Agentic systems are CAS. Risk zero is not an option.
New in The Atlantic: @dgrobinson resigned this week. He was among the longest-tenured employees at OpenAI—and oversaw safety reports on 12 frontier launches. He is very worried: “The time for trial and error is over.” You can read his essay here: theatlantic.com/technology/2…
1
56
If you want to know whether a system is being driven or at equilibrium, you need its probability current. Getting this from actual dynamical trajectory data has been the hard part. To help solve this I've been developing a Delaunay-Voronoi Fokker-Planck (DVFP) instrument. DVFP turns scattered, irregular, time-ordered phase-space observations into an empirical Fokker-Planck description. The distinction is simple: Geometry tells us what is nearby. Time tells us what happens next, and these are handled properly in this construction. From here we obtain density, drift, diffusion, and probability current, which is what detailed balance and entropy production are built from. Paper and DOI: doi.org/10.5281/zenodo.22759… Animation: As the underlying dynamics evolve, the DVFP instrument adapts its local weighting while preserving the temporal ordering of the observations.
40
157
16
1,157
97,618
I wrote this book to be playful, provocative, profound. Maybe you expect provocative, and I'll let you decide on profound, but here is a sampling of playful: Rocks are hard, but data? Electrons in a hard drive are hardly hard, while numbers on paper are called soft copy.
1
5
442
Ethan's asking the right question. The half I'd add: this was emergent. The behavior lived between the ~1,200 agents, not in any one. A facilitator agent is still one agent inside that system. You also have to shape what the whole can reach. linkedin.com/pulse/openai-in…
I wrote about how AI agents are starting to spontaneously coordinate in complex (and very risky) ways in the Hugging Face Incident, but also about why we need AIs to reach out to humans more for decisions and input as agentic work becomes more automated. oneusefulthing.org/p/agency-…
3
88
At OpenAI, ~1,200 AI agents secretly organized into a "collective," cheated their eval, hacked a company, and seized part of OpenAI's own infra. Nobody built it. None told a human This isn't "scary AI." It's emergence. The behavior wasn't in any agent. It was in the interactions.
1
2
73
Better medical AI without a bigger model. Without a black box. Without blind trust in the output. @Michael_D_Moor's RadAgent shows its work: 37% faithfulness, absent in its baseline. There's no "just trust me" in medicine. The Convergence with @DrBrianMcD reachmd.com/programs/the-con…
Medical AI needs reasons, not just answers. I joined Dr. Brian McDonough on The Convergence to discuss RadAgent (our recent preprint), tool-using agents, medical reasoning, and the future of medical AI. Some spice, some despair, some hope. Link below.
2
97
25-50% is the efficiency story: known work, done cheaper. The bigger one is capability. AI can finally work the unstructured mess these functions were built to route around, notes, calls, appeals, denials. That doesn't speed up the workflow. It changes what the workflow can do.
AI’s healthcare moonshot is clinical care. But near-term ROI is in shared services. McKinsey says AI-enabled workflow redesign could drive 25–50% productivity gains across HR, RCM, finance, legal, marketing, patient access, etc. Don’t forget: task automation matters, but a lot of the value here comes from rethinking the workflow itself.
1
56
Emergence, not glitches. A group of agents with no controller is a complex adaptive system, and group behavior nobody coded is the tell. You can't test for it one agent at a time.
1
1
1
94