Mar 6, 2024 · 50m · mad
AI is now revolutionizing early cancer detection | Emi Gal, CEO of Ezra
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The MAD Podcast, host Matt Turck interviews Emi Gal, co-founder and CEO of Ezra, discussing how artificial intelligence is accelerating full-body MRI scanning to make early cancer detection fast, accurate, and affordable.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 16.6% of the talking time here. How this is scored →
Matt as informed peer 3.7 Guest teaching 3.8 Guest disagreement 0.3 Matt pushing back 0.3
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 25:43 Reframing radiologist replacement
Emi rejects the premise derived from Geoff Hinton's talk that AI will eliminate radiologists, asserting instead that radiologists using AI will replace those who do not.
Hardest push from Matt ▶ 25:43 Challenging radiologist longevity with Geoff Hinton quoteMatt pushes back against Emi's optimistic radiologist partnership model by invoking Geoff Hinton's high-profile prediction that AI makes radiology obsolete within years.
Biggest teaching moment ▶ 16:42 Lecture on quantum mechanics and MRI noiseEmi delivers a detailed explanation of Tesla magnet strengths, proton spin alignment, radio frequency pulses, and resonance noise reduction physics.
Matt holds his own ▶ 25:43 Leveraging deep AI history knowledgeMatt demonstrates high domain familiarity by recalling specific expert commentary from deep learning pioneer Geoff Hinton at a Toronto conference to pressure test Emi's business strategy.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome and FirstMark Portfolio Connection | 3 | 1 | 0 | 0 | Matt introduces the connection between FirstMark and Ezra, demonstrating familiarity with the company's Series B funding history. Emi warmly confirms the investor breakdown and recent Series B round details. | |
| The Honeymoon Epiphany and Founding of Ezra | 1 | 2 | 0 | 0 | Matt prompts Emi on the founding story with basic clarifying questions about his previous startup in Romania. Emi recounts reading research papers on his honeymoon that inspired Ezra. | |
| Clinical Validation and Reaching Product-Market Fit | 2 | 3 | 0 | 0 | Matt asks about clinical validation safety thresholds and product-market fit tipping points. Emi educates on assembling 22 top scientific advisors and tracking early life-saving diagnostic outcomes. | |
| Industry Tailwinds and Y Combinator's Call for Startups | 4 | 2 | 0 | 0 | Matt demonstrates industry knowledge by noting longevity trends and Y Combinator's recent call for MRI startups. Emi details Ezra's three scan tiers and price points. | |
| Financial Accessibility and the $500 Scan Master Plan | 4 | 3 | 0 | 0 | Matt asks a targeted follow-up probing whether cost reductions stem from AI acceleration or patient historical baselines. Emi outlines the master plan to bring full-body MRI scans down to $500. | |
| Software Strategy and the Physics of MRI Noise | 2 | 8 | 0 | 0 | Matt frames the choice between full-stack hardware versus pure software. Emi delivers an extended physics explanation of 3 Tesla magnetic field limits, proton alignment, and image denoising algorithms. | |
| Technical Architecture and Proprietary Data Advantages | 5 | 5 | 0 | 0 | Matt asks specific technical questions about computer vision model architecture and training data. Emi explains their U-Net architecture, convolutional layers, and proprietary longitudinal dataset advantages. | |
| Bootstrapping Data and Radiologist Workflow Assistance | 6 | 5 | 2 | 3 | Matt cites Jeff Hinton's famous prediction that AI will replace radiologists to challenge Ezra's partnership model. Emi reframes the premise, arguing that radiologists using AI will replace those who do not. | |
| Ezra Reporter Architecture, Guardrails, and False Positives | 6 | 4 | 1 | 1 | Matt asks if the Ezra Reporter uses generative GPT models and correctly observes how guardrails constrain the output domain. Emi explains using fine-tuned Llama models restricted to pre-written medical ground truth. | |
| Core Lessons for AI Entrepreneurs | 3 | 6 | 0 | 0 | Matt asks for lessons learned for AI founders. Emi outlines three pillars—team, domain knowledge, and data—explaining why non-transferable domain expertise is critical. | |
| Navigating Healthcare Regulation and VC Fundraising | 4 | 4 | 0 | 0 | Matt asks about healthcare fundraising hurdles and regulatory complexity. Emi describes pitch realities, FDA approval timelines, and meeting with 100 investors to secure a lead. | |
| CEO Health Optimization and Performance Routines | 4 | 3 | 0 | 0 | Matt brings up Emi's public founder routine posts, specifically recalling his reliance on compound weightlifting. Emi breaks down his physical health regimen, supplement stack, and 5-hour weekly workout schedule. |
Statements from this episode (29)
Emi Gal: Ezra raised $21M Series B co-led by FirstMark and Healthier Capital
“We announced last week that we raised twenty-one million co-led by Rick at FirstMark and Amir Dan Rubin from Healthier Capital.”
Emi Gal: Ezra's investors include Allianz, Schwarzman family, and Lord David Pryor
“We added Allianz Insurance Group, or the Allianz Life Ventures Fund, which is part of the Allianz Insurance Group. We added the Schwarzman family of Blackstone, and the former head of the NHS in the UK Lord David Pryor.”
Emi Gal: Cancer cannot currently be detected early in most organs
“Early detection is really what makes the difference. But the problem is you can't find cancer early in most organs in the body.”
Gal: MRI is expensive because scanning takes a long time
“Whilst MRI is great, the problem is it's expensive as a modality because it's really slow”
Ezra recruited 22 top scientific advisors before launching its AI scanner
“We brought on board 22 scientific advisors Before we even launched, and these were heads of body MRI at Memorial Sloan Kettering Cancer Center Siddhartha Mukherjee, a Pulitzer Prize winner oncologist at Columbia, the chair of oncology at Columbia some other im…”
Ezra took four years after launch to reach product-market fit
“It was not until last year, so four years in, that we saw Escape Velocity and started getting, like I felt like Promarket Fit was pulling us into great growth as opposed to trying to extract growth and trying to kind of convince people to do this scan that we …”
Emi Gal: Starting a full-body MRI startup now would be very difficult
“Anyone who would want to start a full body MRI company right now would find it very difficult because there's like there are a few companies that are doing a great job, and we've solved most problems that there are in, in this space, and we're growing very fas…”
Emi Gal: Y Combinator's Request for Startups included faster, accurate MRIs
“So YC this year, that was every year they post these kind of 20 ideas that they want people to startup founders to invest in, to spend time on. And one of the ideas this year was MRI and making MRI faster. Making MRI more accurate making, basically doing all o…”
Gal: Ezra offers the fastest full body MRI scan in the world
“We have a cancer focused scan, which is a 30 minute fastest full body MRI in the world.”
Ezra buys MRI time from existing facilities rather than owning centers
“We don't own and operate facilities, we partner with existing facilities, buy MRI scanning time from them, and run our own protocols, AIs, software on their magnets.”
Ezra expects to offer $500, 10-minute full-body MRIs within three years
“And we're currently at the 13 50 scan, and within two to three years, With a lot more AI we think we're gonna get to a 500 dollar 10 minute full body MRI.”
Emi Gal: Most MRI innovation over next decade will be software-driven
“So in an MRI particularly over the next decade, most innovation will come from software.”
Ezra's AI erases noise to enable faster, reduced-repeat MRI scans
“So what we do at Ezra, or one of the things we do is we are able to do the scan fewer times, which results in a noisier image, but we have trained AIs to essentially learn what noise looks like in an MRI machine and just, like, erase the noise.”
Emi Gal: Ezra has scanned thousands of people, yielding millions of MRI slices
“We have thousands of people that we've scanned which results in millions of MRI slices that we can use to train.”
Emi Gal: No existing MRI datasets combine healthy and diseased longitudinal data
“There are no data sets out there with a blend of healthy people and diseased people with longitudinal measures over time”
Ezra gathered 1,000 initial full-body MRI scans using non-AI protocols
“When we launched our first full body scan, it was about 75 minutes. It was not yet using any AI, and we were just selling it as a high quality full body protocol without AI in order to build the data set to be able to train the eyes. And then we got our first …”
Gal: AI already outperforms humans at detecting cancer in medical imaging
“AI can already, in many cases, do a better job at finding potential cancer in Medical imaging.”
Radiologists won't be replaced by AI, but by radiologists using AI
“I think radiologists will not be replaced by AI. They will be replaced by radiologists using AI.”
AI will increase demand for human radiologists over the next decade
“If I am to fast forward five to 10 years, We will need more radiologists than we have now because we will need more medical imaging than we do now.”
Gal: Ezra will never deliver scan reports without human doctor review
“So I don't see a future in which we're delivering a report to a member, Ezra member, without that report having been passed through a human.”
Ezra's AI cuts physician report creation time from 90 to 5 minutes
“Our doctors, we have a team of primary care physicians internally who used to take the radiology report, spend 90 minutes per report to generate an Ezra report, which is on average about seven pages, to describe to the member what each finding means. We built …”
Gal: Generative AI in healthcare requires constrained output and doctor review
“We think that's the correct way to apply gen AI type things into healthcare because it's such, it's so, such a kind of critical domain.”
AI engineering skills do not easily transfer across narrow domains
“I have learned the hard way that it's not that transferable. You actually need, if you're trying to push the state of the art, you need people who have proven experience in the very, very specific narrow scope of problem that you're trying to solve.”
Academic AI researchers rarely possess the skills to ship commercial products
“It's very hard to find an academic who is able to ship a product”
Gal: Unique proprietary data is the only real edge in healthcare AI
“In healthcare especially, the limiting factor in building AIs is data. If you don't have access to a really unique proprietary type of data, you don't have edge.”
Emi Gal: Healthcare startups must add six to nine months for regulatory work
“And so you need to add in six, nine months of regulatory work to any endeavor because it's going to take time to get that done.”
Emi Gal pitched 100 VCs during Ezra's fundraising round
“And for the Ezra raise, I had a hundred meetings. I pitched a hundred VCs, and feel incredibly fortunate to have landed My ideal profile”