Podcast Summary:

Scrum Master Toolbox Podcast: Agile Storytelling from the Trenches
BONUS Episode: The Future of Seeing—Why AI Vision Will Transform Medicine and Human Perception
Guest: Dr. Daniel Sodickson (Chief Medical Scientist, Function Health)
Host: Vasco Duarte
Air Date: February 19, 2026

Overview

This bonus episode explores the impending revolution in artificial intelligence-powered vision, with a focus on medicine and its transformative impact on human perception. Dr. Daniel Sodickson, a leading figure in AI-driven medical imaging and author of The Future of Seeing, discusses his journey, breakthroughs in imaging technology, and the future possibilities unleashed by combining AI with visual data. The conversation spans from the technical evolution of MRI to philosophical reflections on how machines—and humans—can "see" beyond the visible.

Key Discussion Points and Insights

1. The Origins of AI-Driven Medical Imaging

[03:08–05:55]

• Dr. Sodickson credits serendipity and curiosity, not a pre-determined plan, for his focus on "seeing":
  • He recounts a pivotal moment doodling a new MRI data-collection idea on a napkin.
  • Invented "parallel imaging," allowing faster, multi-line MRI data capture, inspired by how the human eye views scenes holistically.
• Quote:

  "For me, it was a complete and utter accident... I was doodling... on a napkin in a piano bar in Boston, and came up with a way that I thought would work to hit multiple lines at once."
  — Daniel Sodickson [04:40]

2. Imaging Impact on Medicine and Discovery

[06:12–09:06]

• Medical imaging advancements don’t just reveal hidden disease but fundamentally change the questions physicians can ask.
• Imaging leads to routine, not rare, patient-impacting discoveries, especially in early tumor detection and preventive care.
• Draws analogy to the Copernican revolution—each new imaging capability shifts scientific understanding.
• Quote:

  "Every time a new imaging modality... came along, they were instantly adopted because when you can see something you couldn't ever see before, it adds value."
  — Daniel Sodickson [06:33]

3. Transforming Errors into Signal: The Power of Perspective

[09:06–11:13]

• Apparent artifacts or "errors" in new imaging technology (e.g., high-frequency MRI distortions) often represent new sources of information.
• Turning a technical limitation into a new field: using electrical properties for imaging.
• Memorable Moment:

  "...we could make an image of the body's electrical properties... we were able to create maps of new information by turning errors into signal."
  — Daniel Sodickson [10:51]

4. How AI 'Sees' Differently Than Humans

[11:38–13:54]

• AI systems aren't limited to human-style inputs (images) and can analyze raw sensor data.
• Convolutional neural networks are inspired by animal visual systems but surpass them by operating directly on unprocessed data.
• Properly trained AI can "fill in gaps," allowing imaging with far less data—sometimes without a full image at all.
• Quote:

  "We can, in some ways, do imaging without the image."
  — Daniel Sodickson [13:41]

5. Upstream AI: Rethinking Data Collection

[15:17–17:49]

• Traditionally, AI operates downstream (analyzing fully formed data), often as a human competitor.
• Dr. Sodickson advocates for "upstream AI," where data collection and even device design are tailored from the start for AI's strengths.
• Envisions cheap, ubiquitous, embedded sensors that AI can use for longitudinal health monitoring, even if humans can't interpret the data.
• Quote:

  "Why are we limiting ourselves to downstream? Why are we limiting ourselves to tasks that humans can already do...? Why aren't we thinking of tasks that the AI can do like that, as you were saying, no human could ever do?"
  — Daniel Sodickson [16:16]

6. Proactive Health: Applying Agile Principles to Medicine

[17:49–19:43]

• His transition to Function Health reflects a desire to implement continuous, proactive health monitoring—reminiscent of Agile, rapid-iteration product development.
• Focuses on collecting routine, multidimensional data (blood tests, imaging) in healthy individuals for early intervention.
• Memorable Moment:

  "...my mission now is to figure out how we build the AI systems that take all of that proactive data and turn it into kind of a GPS for your health."
  — Daniel Sodickson [19:21]

7. Context Is King: The Philosophy of Seeing

[21:08–23:38]

• Human vision and AI alike use context and memory, not just current snapshots, to interpret incomplete data robustly.
• AI's potential lies in comparing new data with a rich history—detecting even minute changes relevant to health.
• Quote:

  "What AI systems can do... is to detect from today's signal whether you look like yourself or whether you've changed in any worrisome way from everything we know about you."
  — Daniel Sodickson [22:11]

8. Beyond Medicine: New Frontiers for AI Vision

[24:39–27:40]

• AI vision techniques have broad applications: satellites, environmental monitoring, surveillance, infrastructure safety, astrophysics, and more.
• The power is in giving machines "memory"—the ability to synthesize data in space and time.
• Memorable Thought Experiment:

  "Could we combine all of the cell phones on Earth into a massive radio telescope...? If we can figure out how to interlace all of the signals... could we be...collectively looking out at the universe with all of that individual power marshaled in coordination?"
  — Daniel Sodickson [26:52]

9. Seeing vs. Knowing: The Human Side

[28:30–31:11]

• Decades in imaging have made Dr. Sodickson hyper-aware of the process of perception.
• He describes almost meditative moments when his scientific and poetic sides merge:
• Quote:

  "Sometimes...everything I'm seeing just sort of fades away. And what I see instead is how I'm seeing. I sort of imagine light bouncing off of things and landing in my eye...tracing the sight lines..."
  — Daniel Sodickson [29:29]

Notable Quotes

| Timestamp | Speaker | Quote | |-----------|----------------|-------------------------------------------------------------------------------------------------------------------------| | 04:40 | Daniel Sodickson | "I was doodling... on a napkin... and came up with a way that I thought would work to hit multiple lines at once." | | 06:33 | Daniel Sodickson | "...when you can see something you couldn't ever see before, it adds value." | | 13:41 | Daniel Sodickson | "We can, in some ways, do imaging without the image." | | 16:16 | Daniel Sodickson | "Why are we limiting ourselves to downstream? ...Why aren't we thinking of tasks that the AI can do like that, as you were saying, no human could ever do?" | | 19:21 | Daniel Sodickson | "...my mission now is to figure out how we build the AI systems that take all of that proactive data and turn it into kind of a GPS for your health." | | 22:11 | Daniel Sodickson | "...to detect from today's signal whether you look like yourself or whether you've changed in any worrisome way..." | | 26:52 | Daniel Sodickson | "...could we be...collectively looking out at the universe with all of that individual power marshaled in coordination?" | | 29:29 | Daniel Sodickson | "Everything I'm seeing just sort of fades away. And what I see instead is how I'm seeing. I sort of imagine light bouncing off of things and landing in my eye..."|

Timestamps for Segment Highlights

• Dr. Sodickson’s accidental entry into imaging: [03:08–05:55]
• Transformative power and philosophy of new imaging capabilities: [06:12–09:06]
• Turning errors into opportunities: [09:06–11:13]
• AI’s unique modes of "seeing": [11:38–13:54]
• Reimagining devices for AI (Upstream AI): [15:17–17:49]
• Agile, proactive health monitoring (Function Health): [17:49–19:43]
• Context, memory, and the evolution of vision paradigms: [21:08–23:38]
• Wider consequences of AI vision (society, science): [24:39–27:40]
• Personal transformation—how imaging changed Daniel’s perception: [28:30–31:11]

Recommended Resources

• The Future of Seeing by Daniel Sodickson
• An Immense World by Ed Yong (on animal senses)
• "Deep Learning" (Nature paper) by Yann LeCun, Yoshua Bengio, and Geoffrey Hinton
• "A Path Towards Autonomous Machine Intelligence" by Yann LeCun

Where to Learn More

• Daniel Sodickson’s book: The Future of Seeing
• Academic papers from Daniel’s lab
• Follow advancements from Function Health on proactive, AI-enabled health monitoring

Final Note:
Dr. Sodickson’s interview offers a compelling exploration of how AI-driven vision is about much more than machines mimicking our eyes. It’s about context, memory, and the leap from passive observation to proactive, holistic understanding—across medicine and all domains where “seeing” is being redefined.