A few further thoughts about NVIDIA CEO Jensen Huang’s farcical description of the impact AI has “already” had in completely revamping the field of radiology.

Huang presents radiology as the “evidence” of what the near-future impact of AI on the workforce will be. I’ll include the quote again in this post for completeness:

One thing that I will say, give you some evidence, is that, and I was just telling Elon about this earlier, radiology, for example, has largely been converted to AI-driven radiology. And there’s some really great companies doing that. And the surprising thing is the prediction that all radiologists would be the first jobs to go was exactly the opposite. The trend shows that there are more radiologists being hired now as a result of AI.

And the reason for that, if you take a step back, it’s because the goal of a radiologist is not to study the images. The goal of a radiologist is to diagnose a disease. Now the studying of the images became so productive they could study more images, study more modalities, spend more time with the patients, and as a result, they were actually accepting more patients. We’re doing more radiology all around the world, we’re doing a better job with diagnosing disease.

Again, none of that is even merely an exaggeration. It’s just wholly untrue as to the current use of AI in radiology, let alone the impact far enough in the past to have percolated through and already changed how the whole field practices.

This is a man personally worth almost $200 billion in charge of a company with a market cap over $4 trillion. One imagines he has access to reality if desired.

I’ve seen the clip shared countless places by credulous people who don’t know any better. This is not to say these things won’t happen, but the use of the past tense is a real problem. I think it’s worth putting the fantasy in context.

Radiology Isn’t Illustrating Jevons’ Paradox

Jevons Paradox is the observation that technological advancements that increase resource efficiency can counterintuitively lead to an overall increase in resource consumption. Jevons’ original formulation concerned improving coal technology leading to paradoxically increased demand/consumption of coal. As in, when it’s cheaper, you can buy more and do more.

Huang is parroting a recently commonly invoked human radiology analogy: AI makes radiologists more efficient in interpreting scans, more scans somehow get done, and voila demand magically increases.

The problem is this doesn’t reflect reality, at least not currently. Imaging volumes have been increasing steadily for decades. Scan acquisition time has nothing to do with scan interpretation time (the former actually ironically is benefiting from machine acceleration in some situations). Interpreting efficiency has barely moved, if at all, and turnaround times are actually lengthening amidst unmanageable volumes. Increased demand for radiologists is just the secular shortage of qualified radiologists struggling to keep up with organically increasing volumes. AI has nothing to do with it.

To further compound how much we are not in Jevons’ land yet: reimbursement for radiologist professional services is stable to falling from payors (again, not related to efficiency), but radiologists have actually cost the system more recently as they demand stipends from hospital systems to provide continued access to services. More simple supply/demand at work. Those scans, even the “more efficient” ones, do not cost less for patients or payors yet. More efficient MRIs and radiologists are both paid the same on a per-study basis.

There are some AI tools for radiologists with reasonable market penetration: variations on list triage (enabling potentially positive scans to jump the queue for interpretation) and generative draft impressions derived from the findings section of the report. Meaningful computer vision is frankly not in particularly widespread use despite what is parroted by tech companies using the news as free marketing, and so far limited to narrow pathologies like brain bleeds, blood clots, fracture detection. Any purported efficiency benefits have been unconvincing. A mediocre tool allows careless people to move faster but gives careful radiologists just another thing to review. So far, the state of the art has been mostly a wash.

Again, this isn’t one of those “AI could never do what my amazing organic brain can do!”; it’s an “AI really certainly hasn’t yet in real practice, and nothing on the market so far has really moved the needle.”

Greater efficiency could trickle down to depress prices in the future and/or eventually lead to increased demand for imaging, but it hasn’t happened yet, and it’s unclear if it did that it would lead to increased demand for the human component. Healthcare is not really a free market. No one in healthcare is the equivalent of coal.

Is it possible that one day AI will lead to automated scan acquisition and instantaneous scan interpretation a la Star Trek, and we will suddenly not just want but need all scans all the time, and everyone will be building more and more machines to scan more and more people because interpretive costs are down, and physical throughput is the real bottleneck? Sure. We can just put a conveyor belt at the entrance to the ER!

Is it also possible that—instead of a Jevons process—professional fees could crash as radiologists become box-checking liability-operators, but that as the field contracts, the smaller number of remaining radiologists will enjoy persistent high wages in the form of obscolence rents? Also, sure.

Is any seismic AI change, if it occurs, going to be limited to radiology? I doubt it.

Ultimately, humans are not quite a natural resource. The tortured metaphor may or may not hold in the future, but we should all be able to agree that it sure hasn’t happened in the past.

The Obsession with the Jevons

So why is the world—and everyone contributing to the AI bubble—so obsessed with Jevons’ Paradox?

Because it’s comforting.

Instead of worrying that the economic value of most humans is going to trend toward zero and that our entire society will have to contend with massive disruption, it’s far easier to believe that AI efficiency will unlock the magic of productivity. That somehow, this machine intelligence revolution will set us free to do our best, most magical, most human work.

A narrative where increased efficiency leads to increased demand soothes the uncertain soul. In that version of the story, the highest-skilled humans won’t be made obsolete—they’ll be unleashed.

I don’t know if that logic will hold in real life over the coming years or not. What I do know is that invoking it right now is a marketing fantasy.

To reiterate: I don’t know what will happen to radiology or to any other field. The future is uncertain, and I certainly don’t have a crystal ball. We have plenty of bullshit jobs already, even without the help of AI.

What I do know is that all attempts to pretend radiology has already seen the fruits of AI and absorbed them—that AI is responsible for the current surge in demand for radiologist services—are lies.

But we can appreciate why this specific variation of the lie is being told:

In 2016, AI Godfather Geoffrey Hinton famously said:

People should stop training radiologists now—it’s just completely obvious within 5 years deep learning is going to do better than radiologists. It might be 10 years, but we’ve got plenty of radiologists already.

So the fact that we’re still here and in more demand than ever is supposed to be comforting to other humans.

Don’t fear AI! Even the radiologists are thriving!

I’m sure that in advance of an earnings call, a narrative where AI unlocks human potential is so much more compelling than one with a zero-sum game where the short-term economic value goes to tech companies but the long-term impact is to potentially destroy human enterprises and sink the economic value of previously high-training, high-skill, once-human tasks.

If we think further ahead, no one is going to pay for computer work the same as we did for comparable human work for a sustained period. The RVU system in healthcare already attempts to account for effort, liability, etc. We don’t get to hold even that one system’s current balance fixed while the world changes. Causes yield effects, and consequences themselves have consequences.

Anyone who imagines a perpetual money-printing machine generating revenues to match the humans they once replaced is naive, so maximum AI utopianism (and the value tied up in these companies) doesn’t envision that world. The devastating disruption fears may or may not be valid, but Huang and others in the tech world clearly feel compelled to address them.

It seems to me that Huang is trying to wave away generalized replacement fears by pretending that radiology is the canary in the coal mine and we’re still here, therefore rainbows and unicorns.

Maybe that “doomer” path to darkness is wrong, but that doesn’t mean radiology is the example to light the way.

I believe in the importance of thriving independent private practices for the field of radiology. True private practice—where doctors control the organization, are responsible to their peers and patients, and earn the full fruits of their labor—is the benchmark that sets the market and provides the anchor against exploitation from unscrupulous employers.

My group, like most groups in this market, is hiring. Here are several more 100%-independent radiologist-owned private practices across the US that are recruiting. If you’re in the market for a new position, consider reaching out with your CV. (Click the triangles for more information.)

//

Iowa Radiology

(Des Moines, Iowa)

34-radiologist subspecialized collegial private practice with long-term stability providing coverage to 30 locations across Central and North Central Iowa. Ownership of several outpatient imaging locations yields distributions from technical fees, and Des Moines is one of the fastest-growing areas in the country with a low cost of living, excellent schools, and low taxes.

Partnership track:

• 2-year partnership track with 12 weeks vacation
• On-site daytime coverage; hybrid night and weekend call coverage
• Greatest needs: Breast, Neuro, General. All subspecialties considered for the right candidate.

Employee track:

• Fully remote, hybrid, or on-site options
• Excellent salary and benefits package, including up to 10 weeks of vacation depending on role/shift
• Immediate Need: 3^rd Shift Radiologist
  • Nine-hour shift, 7 on/14 off, 100% teleradiology
  • 90-120 cases/shift; ~40 of those CT
  • Base compensation $450,000 w/ additional compensation opportunities
    • Additional comp consideration based on subspecialty (for example, can final read neuro cases)
    • Moonlighting available
    • RVU productivity incentives
  • Full benefits package including 401k, health/dental/vision, pension, life insurance, etc. ($83k benefit value annually)
  • Malpractice and tail paid
• Other Available positions: General, Breast, Flex Shift, Overnight
  • Overnight shift is 10p-8a CST, 7 on/14 off, 100% teleradiology with average 90-120 total cases per shift
  • Evening shift is 4p-12a CST, alternating weeks M-F only (26 weeks and no holidays/weekends)

Learn more at http://www.iowaradiology.com
Contact: recruit@iowarad.com

//

RAPA

(Central Arkansas [Little Rock, Conway, Searcy, and Pine Bluff] & Northwest Arkansas)

40+ radiologists covering 20+ sites (hospitals + clinics) with a multi-subspecialty mix of inpatient, outpatient, and ED imaging.

• Robust reading room assistants and IT presence to promote peak efficiency
• Unified Clario reading list, Inteleviewer PACS, Powerscribe dictation (no switching stations or PACS to read other sites)
• RadAI automated impressions to improve efficiency.

Partnership track – Central Arkansas (Little Rock, Conway, Searcy, and Pine Bluff) and Northwest Arkansas:

• 1 year to partnership
• 10 weeks vacation + 1/2 day off per week on average for partners
• Highly competitive compensation combined with low cost of living
• Internal moonlighting options to boost income
• Robust CME allowance, signing and moving bonuses, full benefits
• Many work-from-home shift options
• Minimal after-hours requirements – overnight shifts are fully staffed with telerads

100% Remote Nighthawk, partnership track:

• 7 on/14 off, 10 pm-7:45 am CST
• 1 year to partnership
• Competitive compensation, robust CME allowance, signing bonus, full benefits

100% Remote Neuro/General:

• Daytime tele neuro/general radiology approximately 50/50 split.
• Option for employed or partner track. Partner track includes call weekend shifts, approximately 1 in 5 weekends.

Employee track:

• General radiology, mammography, and other options available
• Fully remote, hybrid, or on-site options available
• Flexible scheduling
• Competitive salaries
• Robust CME allowance, signing and moving bonuses, full benefits

Greatest needs are IR, mammography, body imaging, MSK, and nuclear medicine, but all subspecialties and general radiologists are welcome.

Learn more at http://rapaxray.com/
Contact: recruitment@rapaxray.com and Dr. Brandon Kelly at bkelly@rapaxray.com

//

Eastern Radiologists

(Greenville & Coastal North Carolina)

Eastern Radiologists is a private practice in Eastern North Carolina that is physician-owned and operated and seeking to hire multiple radiologists due to growth and retirement. The well-established private practice serves a large geographic region through 14 area hospitals and several state-of-the-art imaging centers. Support is provided by nearly 70 subspecialty radiologists.

Currently seeking candidates with the following training:

• Pediatric Imaging
• Breast Imaging
• Neuroradiology
• Body Imaging
• Nighthawk/Overnight
• General Diagnostic Radiology
• Vascular/Interventional

Details:

• Positions are available in Greenville and other coastal communities in Eastern North Carolina.
• Most are partnership track positions which include evening and weekend call responsibilities, but other employment options can be considered. Employment positions can be customized to fit desired schedule/income. Internal moonlighting is available for extra income.
• Great benefits, competitive salary including profit sharing and bonus, generous vacation and paid family leave, and business/educational discretionary account.
• Sign-on bonus

Learn more at https://www.easternrad.com.
Contact: Erica Askew at easkew@easternrad.com

//

Midwest Radiology

(Minneapolis – St. Paul, Minnesota)

Large 175+ subspecialized practice with 17 weeks of vacation. Two-year shareholder track for new graduates and a one-year track with experience.

Positions (On-site):

Positions (Remote):

Learn more at www.midwestradiology.com.
Contact: Barry.Lindo@MidwestRadiology.com

//

Mecklenburg Radiology Associates

(Charlotte, North Carolina)

Established in 1917 and fiercely independent, MRA is a well-respected subspecialized practice of 60 radiologists and growing. They especially take pride in their group culture (seriously, there are some really nice people in that group including one of my old senior residents).

Partnership track:

• 3-year partnership track with 8 weeks of vacation per year
  $400K starting salary and an additional $75K sign-on bonus ($125K with 2+ years of experience)
• On-site daytime hospital and imaging center coverage. Hybrid evening and weekend call coverage, including the option to read from home.
• Available positions: Breast, Body, Cardiovascular, Neuro, Light IR, MSK, Nuclear Medicine, and General

Employee track:

• Fully remote, hybrid, or on-site options
• Competitive salary
• Available positions: MSK, General, Overnight ER, Swing Shift ER, Neuro, Body, and Cardiovascular

Learn more at www.meckrad.com/recruitment.
Contact: Charlene Eichinger at ceichinger@meckrad.com

//

Radiology of Huntsville

(Huntsville, Alabama)

ROH is 60+ physician, high volume, private practice with a 2-year partnership track and nominal buy-in. Hiring for a remote partnership-eligible overnight position as well as onsite/remote openings in most subspecialites. Huntsville is a tremendously livable, high-tech, and growing mid-sized city (a #1 Best Place to Live in the U.S. by U.S. News and World Report).

Learn more at www.radiologyofhuntsville.com.
Contact: Brandy McCown at bmccown@radhunt.com or call 256.713.0621.

//

Radiology and Imaging Specialists

(Central Florida: Lakeland/Winter Haven, SW Orlando, Bradenton)

Independent and long-standing group covering a diverse mix of financially sound hospital partners, outpatient imaging centers, an ASC/OBL, clinics, and a multi-specialty group. 50+ radiologists and 10+ midlevel providers. In-house and experienced IT, Credentialing, HR, and Admin team including dedicated Radiology Liaison support 24/7/365.

Recruiting for Body, Breast, Cardiac, Neuro, IR, and General.

Partnership Track:

• 1-year track for diagnostic radiologists, 2-year track for interventional radiologists
• 10 weeks of vacation
• Sign-on bonus for diagnostic radiologists
• Call compensation for interventional radiologists
• Internal moonlighting available but not required
• Full benefits including CME allocation

Employee or Contractor:

• Flexible scheduling: Hospital (ED/IP) or outpatient coverage options available
• Fully remote, hybrid, or on-site depending on location(s) and coverage schedule desired
• Competitive compensation models (including benefits if employed/full-time)

Learn more at http://risimaging.com.
Contact: Alice Varnadore, Executive Assistant at avarnadore@risimaging.com

//

South Texas Radiology Group

(San Antonio, Texas)

STRG is a well-established, expanding, independent subspecialty practice of 70+ radiologists serving 15 hospitals, multiple freestanding ED facilities, and a large outpatient imaging practice (STRIC). Both onsite and remote positions are available.

STRG is a forward-thinking practice with a strong and cohesive culture. Multiple AI projects are in place and in progress to improve efficiency and quality. Internal moonlighting opportunities are available.

Shareholder Track:

• Flexible associate period as short as 18-months.
• Leads to full partnership with board of directors membership.
• Equity ownership within a large expanding outpatient imaging practice (STRIC)
• Competitive salary with up to $800K starting salary for experienced applicants. Up to $600K starting salary for inexperienced applicants.
• Excellent benefits package
• Onsite and work-from-home options for daily work and call
• Hiring across all specialties, with a current emphasis on Body Imaging and Cardiothoracic/Body

Emergency Radiology Track:

• 1 week on/1 week off (7/7) and 1 week on/2 weeks off (7/14) positions available
• Competitive salary with excellent benefits package available.
• Fully remote and onsite options available.

Employee Mammography Position (Onsite):

• Competitive base salary with excellent benefits package
• No call. No nights or weekends.
• Remote opportunities available for general radiology or body imaging work if interested.

Employee Position – All Subspecialties (Onsite, Hybrid, or Remote ):

• Competitive base salary with excellent benefits package
• 5.5 weekends (day shift) per year. No evening or overnight call.

Learn more at https://stric.com/.
Contact: Waynea Finley at wfinley@strg-pa.com

//

Reno Radiological Associates

(Reno, Nevada)

Serving Northern Nevada since 1947, RRA is a strong, highly subspecialized, democratic, extremely collegial, high-earning group of 25 radiologists working in perhaps the best city in the country for raising a family and enjoying unparalleled access to natural beauty, endless outdoor activities (perhaps the best ski-city in the US), and incredible weather (jokingly referred to as “the cheapest city in California” for its proximity to Lake Tahoe (<30 min), Napa Valley, national parks, and more without the super high cost of living and no state income tax).

Shareholder Track (onsite):

• 1 year to 75% partner, 2 years to 100%
• Top 90+% comp nationally, maximally generous benefits
• 8/9/10+ weeks of vacation
• Remote shifts + opportunities for extra income, $100 buy-in

Employee Track (remote or onsite):

• High comp, Maximally generous benefits, opportunities for extra income
• Onsite preferred, fully remote night available (8p-3a PST)

Independent Contractor:

• Flexible shift options, high compensation

Learn more at http://www.renorad.com/
Contact: CEO Anthony Dispenziere at adispenziere@renorad.com

//

Vantage Radiology & Diagnostic Services

(Near Seattle, Washington)

~35 subspecialized radiologists practicing near Seattle and enjoying a quality lifestyle focused on work-life balance, located minutes from urban amenities, and with easy access to abundant outdoor activities in the natural beauty of the Pacific Northwest. We’ve provided professional services for two hospitals since 1970 and are co-owners of their affiliated imaging centers.

• Competitive salary ($370-570k) with a signing bonus for a 4-day work week and 8 weeks of vacation.
• Benefits include a CME allowance, memberships to professional societies, profit-sharing, malpractice and disability insurance, 401K contributions, and a commitment to associate mentorship.
• All partners and partner-track associates share in general call responsibilities (many from a practice-provided remote home workstation). Overnight call is covered by our internal nighthawk team.

Openings:

Learn more at https://www.vrads.com/
Contact: recruiting@vrads.com

//

Radiology Consultants of Little Rock

(Little Rock, Arkansas)

Founded in 1961, RCLR is a 25+ radiologist, subspecialized, independent private practice serving the largest healthcare system in Arkansas as well as the largest urology group in the state.

Learn more at https://www.radconlr.com
Contact: Terrie Threlkeld at terrie.threlkeld@radconlr.com

//

Inland Imaging

(Washington, Oregon, Montana)

Inland Imaging is a 120+ radiologist-owned private practice proudly serving multiple outpatient, clinic, and hospital sites throughout the Inland Northwest region. We own and operate one of the first outpatient interventional labs in the West in addition to 7 outpatient imaging centers.

Our practice:

• Sub-specialty interpretations available across all locations.
• All studies are accessible on a common PACS system, one common voice recognition system, and one common worklist
• Internal Nighthawk System providing final interpretations on all ER, STAT, and Inpatient Exams. 24/7 Neuroradiology Coverage and IR Call Coverage.
• 2-year Partnership Track.
• Full-Time Radiologist Average Time Off = 15 weeks with the ability to decrease or increase total time off.
• Average after-hours obligations = 9 weekends per year and 22 evening/swing Shifts.
• Salary Range – $450,000 – $600,000
• Full Benefit Package including health, life, vision, dental, disability, and life.
• 401k match, profit-sharing contribution, and cash balance plan.
• Hybrid Work Options

Current Openings:

• Remote overnight ER
• Neuroradiology (choice of Spokane, Seattle, Walla Walla, Tri-Cities, or Missoula)
• Washington
  • Tri-Cities (with $75k signing bonus)
    • Breast
    • Neuro
    • MSK
  • Spokane – Pediatrics
  • Moses Lake – General/Breast (with $90k signing bonus, $20k relocation benefit, and annual retention bonus)
  • Colville – General/Breast
• Montana
  • Missoula – IR
• Oregon
  • Pendleton – General/Breast

More information about these positions can be found on our Independent Radiology listing.

Learn more at https://inlandimaging.com/careers
Contact: Sarah Russell, CEO at srussell@inlandimaging.com

//

Charleston Radiologists

(Charleston, South Carolina)

Charleston Radiologists, PA (CRPA) is a private 18-physician radiology group providing care to Charleston and the surrounding area for over 50 years.

Partnership track:

• Open to any fellowship-trained radiologist willing to do a mix of general & subspecialty work
• $500K+ starting salary based on experience, followed by
• 90% of partner salary in year 2
• 2-year partnership track
• Sign-on bonus based on experience
• 8+ weeks vacation
• Nighthawk coverage beginning January 2026
• Benefits include health insurance, malpractice insurance, HSA, 401K match, $7,500 CME allowance, relocation stipend (over $78,000 in annual benefits)
• On-site daytime coverage. Hybrid evening and weekend call coverage, including the option to read from home.

Employee track:

• Fully remote, hybrid, or on-site options
• Competitive salary based on experience

Contact: Tonya McGue at tonyamcgue@gmail.com

//

Grand Traverse Radiologists

(Traverse City, Michigan)

Established in 1938, GTR is a 100% physician-owned and operated private practice in a four-season resort town on the bays of Lake Michigan.

• High compensation and 12 weeks vacation
• Post-call weekend days off
• Internal moonlighting available
• Positions/partnerships are on-site, but multiple shifts can be done from home.
• Collegial and youthful group with an average age under 40. All partners are equal.
• Hiring for growth. Lists are caught up!

Needs: Body, Breast, ER/General, and IR

Learn more at https://www.grandtraverseradiologists.com
Contact: Dr. Anthony Livorine at alivorine@gtradiologists.com

//

If you’re a group looking to advertise, get in touch. The monthly post here is limited to just a handful of groups at a time, but last year, I launched Independent Radiology as a resource for the broader community, a dedicated private practice radiology job board featuring 140+ groups. If you’re in the market, please also check it out for your job-hunting needs.

This week, Elon Musk and Nvidia’s Jensen Huang discussed AI and the future of technology at the U.S.-Saudi Investment Forum. Here is Jensen Huang discussing radiology:

One thing that I will say, give you some evidence, is that, and I was just telling Elon about this earlier, radiology, for example, has largely been converted to AI-driven radiology. And there’s some really great companies doing that. And the surprising thing is the prediction that all radiologists would be the first jobs to go was exactly the opposite. The trend shows that there are more radiologists being hired now as a result of AI.

This is sheer unadulterated fiction. Leave aside the fuzziness of what “AI-driven radiology” might mean; AI simply doesn’t drive a meaningful part of the radiology workflow. Some AI list triage and a few algorithms to detect intracranial blood or fractures has not changed the game in even the slightest of ways. The only thing that has been in meaningful if still limited use over the past few years that has arguably driven even small efficiency gains is generative AI for drafting impressions based on dictated findings.

There are, of course, more things in rare use and plenty of things announced that might matter, but that is beside the point here: Jensen is wrong.

More radiologists are being hired because there is a shortage of radiologists due to steady, some would say incessant, imaging volume growth. The heavy utilization of CT and MRI in US healthcare has literally—and I do mean literally—nothing to do with AI.

The only thing AI could potentially have to do with hiring more radiologists is faster MRI scanning with some vendors/machines/protocols allowing for more patient throughput, impressive but still in its infancy and with limited market penetration.

And the reason for that, if you take a step back, it’s because the goal of a radiologist is not to study the images. The goal of a radiologist is to diagnose a disease. Now the studying of the images became so productive they could study more images, study more modalities, spend more time with the patients, and as a result, they were actually accepting more patients. We’re doing more radiology all around the world, we’re doing a better job with diagnosing disease.

Sure, the goal of a radiologist is to diagnose disease. But, let’s not pretend that radiology hasn’t, for the past century since its inception, essentially been the art and science of diagnosing disease through studying images and putting imaging findings in context. What Jensen is saying is just nonsense.

Even if we accept that the fraction of radiologists using clot- and fracture-detection tools is doing a better job diagnosing disease (very unclear), we are not, as a field “study[ing] more modalities” (what?!) or “spend[ing] more time with the patients” (less than ever thanks to heavy volumes, long turn-around-times, and the explosion of teleradiology). Current computer vision tools do not make radiologists significantly more efficient unless they inappropriately trust them enough to stop looking at the images.

And so that’s kind of the near term outcome of AI and productivity.

I don’t know if Jensen actually doesn’t know anything about radiology (everyone’s favorite white-collar AI-replacement use-case) or if this is a cynical don’t-fear-the-future puff angle. But either way, he’s wrong across the board.

I am reminded of Michael Crichton’s Gell-Mann Amnesia, where you realize how useless many perspectives and most news are only when confronted with obviously incorrect information in contexts for which you are a subject matter expert. Almost every single news article about AI and radiology is entirely wrong. They’re acting like what the world could look like in the coming years is what has already happened: that we are awash in game-changing, useful AI that has rapidly been deployed across the field and fundamentally altered the practice of radiology. And that’s not true. Ironically, clinicians using LLMs for documentation is probably far more ubiquitous and impactful so far.

Now being wrong about the present doesn’t necessarily mean being wrong about the future. Things are changing fast, and the future is always largely unknowable.

But the reality distortion is just so damn irritating.

This week, the ABR quietly dropped a big change in their long-term plans for the new oral board version of the Certifying Exam. After the very first administration in early 2028 during fellowship for the class of 2027, subsequent administrations will occur at the end of residency:

That’s the email I got as a program director.

As in, in 2028, diagnostic radiology, as a field, will again be graduating board-certified (not “board eligible”) radiologists.

The decision to change the (useless, duplicative) Certifying Exam was first announced back in February 2023. In April 2023, they then announced their intention to bring back the oral boards.

The original plan was to keep the timing the same despite the change in format, so that residents would take the exam during the calendar year after graduating from residency, typically a few months into their first post-fellowship attending job. Despite the reality that orals would be much harder to prepare for outside of the residency training environment than a written exam, the ABR referred to this timing as “the least bad choice.”

In that “Backwards to the Future” article, I wrote:

This exam needs to be at the end of residency like it used to be. If anything, it might help combat the post-Core senioritis that many fourth-years struggle with, particularly when rotating through services outside of their chosen specialty. I appreciate that many program directors don’t want this during residency because in the past seniors used to disappear from service (and especially the call pool) before Orals just like they do now before the Core Exam. It’s easier to run a residency with only one class preparing for one big test at a time. But convenience shouldn’t be our primary metric.

Time will tell. I think I had it right in 2023, and clearly enough stakeholders agreed that the ABR has changed its plan before even doing it a single time.

In order to prevent two classes disappearing concurrently in June for their respective boards, the Core Exam has been pushed back into early/fall R4 year so that the senior year will now contain both board exams. Even with that scheduling mitigation, residencies have a lot of work to do to make this happen.

Not all radiology jobs are created equal in part because not all radiologists with the same job are treated equally. In a field divided between democratic groups, corporate employers, and academic institutions, the meaning of fairness and the value of “specialness” vary wildly.

Compensation, autonomy, and respect are all on the table.

For different kinds of radiology jobs, there are different kinds of radiologists.

The Democratic “Ideal”

One thing about any traditional private practice is that, in most cases, all partners are equal. They share in the work and share in the profits. Typically, any differences in compensation—if there are any—are a reflection of differing schedules (like buying and selling of weekend call shifts or vacation) or a reflection of a productivity incentive component, where the radiologist earns additional income for RVUs generated in excess of some predetermined benchmark (because while pay is often equal, production often is not). An external entity can help support necessary admin time through stipends/directorships, but this usually comes from outside of the practice.

If things aren’t fair and transparent, something is wrong.

When Someone Else Holds the Keys

Radiologists working for a third party—like a PE-owned entity or a hospital/university medical center—are in a different situation.

Obviously, in some cases, people can be paid and valued similarly. But a third party holding the keys creates more opportunity for sweetheart deals and special treatment.

This isn’t a knock on those models, because ultimately while flexibility can be used poorly—by undervaluing people, rewarding friends, or exploting those who don’t negotiate—it can also be a powerful business tool (for “good”?) in the sense that you can flexibly pay what the market demands for a given in-demand skill set, even if it doesn’t seem “fair.”

If you’re trying to grow a service line (or keep one on life support) and you need someone with specific skills, you can choose to invest in that person in a way that can be challenging, if not impossible, in a democratic group.

In a world where some radiologists are attempting to optimize for $/RVU, we shouldn’t pretend that democracy always works or that “fairness” always feels fair.

In my practice, a 20-year veteran doesn’t make more for the same work as a new partner. In the academic center I trained at, some senior physicians earned more while doing less.

Now, for those special radiologists who are in demand (like breast imagers in recent years), the current shortage has again enabled a lot of offers—sometimes with high compensation or cush schedules available even for remote work—for the right kind of person for the right kind of job.

What may feel arbitrary or unfair may just be a necessary, intentional response to market forces in order to avoid operational insolvency.

Merit & Loyalty

There is also an important distinction between loyalty to an institution or a platonic ideal, loyalty to a deserving person, or nepotism. The classic academic notion of paying your dues and enjoying better pay with more respect and a better schedule merely through seniority is perhaps not the best way to create a well-functioning meritocratic enterprise.

Academic radiologists need to believe in both the mission and the institution to invest over the long term. Rapid turnover, bad governance, and obvious disparities can easily sabotage what should be the strongest cultures in healthcare.

I once knew an outstanding attending who left their institution because the new junior faculty (including some she helped train) were getting a better deal, including higher compensation. The market had moved, but the institution wasn’t willing to revisit established faculty salaries. The department isn’t a democracy, but this radiologist was worth more by all metrics.

If the academy can’t figure out how to balance specialness with fairness, it’s going to continue to exist in a no-man’s land between democratic private practice and commodified but well-paying corporate work. Many doctors have figured out that the institution often doesn’t love you back.

Rational Actors, Systemic Consequences?

What is best for the individual in the short term may be at odds with what is best for the community in the short term and/or the field in the long term.

This is just another reflection of the tragedy of the radiology commons that plagues all sectors of healthcare:

Those individual choices are logical. The “right” move having downstream effects doesn’t make it a bad choice, especially if the negative consequences are hypothetical or only occur if others pile on (and even then over a long time horizon); that’s why it’s called a tragedy: it’s mostly reasonable people doing reasonable things. Whether those individuals will find that their new opportunities are worth it—or live up to the anticipation—is, of course, unknowable.

Assuredly, sometimes they do. The radiology gig economy is growing precisely because there are a lot of people optimizing for compensation-per-effort and/or flexibility, and some are clearly very satisfied.

But, sometimes, the reality doesn’t quite live up to the expectation. Certainly, some groups that sold to RP over the past decade have regretted the decision. And I see no reason to assume that trends toward commodified pay-per-widget work in a consolidated world will lead to maximum radiologist utility over a long-term time horizon.

Rates per RVU are awesome—but only when they’re high. In the long run, commodification doesn’t care how special you used to be.

A previous reader question:

What do you think is a fair compensation ratio for pre-partner to partner pay? It seems like a lot of jobs offer a 50 to 100 percent pay bump. Is there a threshold that should be a red flag?

I don’t think there is a red flag number.

These numbers mostly reflect supply and demand (and in some cases the impact of technical fees from center ownership after a buy-in).

Part of what will feel acceptable will depend on how long the track is. So if a group has a one-year track, you can tolerate a pretty big differential, but if they have a 5-year one, that might be unconscionable. If someone has a seven-year track, I probably wouldn’t want a big differential—that’s a long time to be paid less. (Given the number of unknowns over almost a decade, I  also think it would be very hard to know if you’re working toward a healthy return on that sweaty equity over such a long period).

Part of it will also depend on how high partner pay is. If a group has truly incredible contracts or an amazing real estate portfolio, it may make sense to accept a large temporary differential to enjoy potential massive returns on that time over the long term.

So I don’t think there’s a set number for it. I think it’s more a matter, unfortunately, of the holistic view.

The reality is that if you look at private practices over the past five years, everyone has shortened their track and bumped associate pay. I think most practices, especially in competitive areas, are largely doing what they can to balance recruiting desirability and providing perks to partnership.

Especially when hiring fresh graduates, it’s also not uncommon for a practice to lose money on its new hires for a while until they get up to speed. The reality is: the practice is often investing in you upfront. Partners also take risks that associates don’t, so there have to be some benefits to being a partner.

So again to summarize: it’s all supply and demand. These are businesses, and fairness is in the eye of the beholder. Years ago, when the job market was tight, we had long tracks and big differences. We are in a different era. Tracks and pay are what they have to be to recruit, and the better the offer relative to a partner, the more desperate the need to recruit or the more challenging the competition for recruitment is. The increasingly nationwide market for teleradiologists isn’t finished having its ripple effects.

So I am entirely unwilling to say there’s a rule of thumb here. Everything is local, but even then, sometimes things are good on paper because they have to be to be competitive in the market, which might mean they’re not competitive in some other way that’s harder to measure.

There are few shortcuts to evaluating jobs, few true red flags, and no ways to entirely de-risk the big decision of where to work.

A reader question:

A lot of my attendings recommend my first job should be somewhere like academics or a hospital system where I have support if there’s a complicated case or someone to help me. Do you feel like you have that in private practice?

So I personally had/have that. Does everybody? No, it depends on the practice. I originally thought most people do, but the number of people I hear from on their second job search has informed me that this is certainly not universal.

But, overall, yes. I think the idea that academia has a monopoly on support is totally inaccurate. People can make you feel inept or give you a hard time for your inevitable mistakes in any environment (I often noticed more attending-on-attending cattiness when I was a trainee).

One key support-related question: Is there a way for you to ask people for help when you have a tough case?

There are plenty of practices now that have built-in instant messaging/case sharing features in their PACS. In this setting, even teleradiologists can share cases with their colleagues back and forth all the time so long as people are generally responsive and sufficiently pleasant.

(Call is always a bit of a different story when there are fewer people working, but this varies too. It’s often a lonelier one-person job. Texting or phoning a friend is always an option, but it’s certainly easier if people are on the outpatient list moonlighting etc and able to provide some support as needed when you’re stuck on a tough case. Being comfortable asking a colleague is, of course, a really helpful place to be psychologically.)

Yes, being in a big, vibrant, distracting reading room is probably going to feel more supportive and lively for most people. One question to answer for yourself when considering an academic job is whether that environment still actually exists. With demands for remote work and expansion of academic medical centers, even large institutions sometimes have their rads increasingly scattered to the winds. (Then, you have to ask yourself if you’ll actually feel more comfortable asking in person, potentially in front of additional attendings and trainees.)

Related and important: Do people share your mistakes with you in a way that’s not going to make you feel too bad, but still let you learn from it? Or do people roll their eyes when you have a miss but don’t tell you, potentially mocking you in front of others but robbing you of the chance to learn from it? Again, that can happen anywhere (including academics).

Ultimately, I think support has more to do with the specific job and less with the model. Every practice is  “collegial” in its job postings, regardless of the reality, and plenty of radiologists in all environments take pride in their work and want new hires to learn and achieve high performance.

I think there’s a certain bubble doctors get into due to the nature of medical education, where we think academia is where the good work happens, and the outside hospital is where the bad work happens. My perception between my experience in academia, my current privademic model, and seeing the work of other practices working in our health system, is that there is no consistent relationship between overall model and quality. Subspecialization to extent, but there are good and bad radiologists and good and bad versions of every model, including in the academy.

I do think being 100 percent teleradiology is probably overall harder to feel supported. Certainly not impossible, but just those interactions won’t all feel the same if no one knows who you are and you don’t really know anybody. Asking a name on a chat list you’ve never met before doesn’t feel the same as asking a friend or a colleague in the same room or one you’ve had dinner with.

How “supported” you feel in that setting may have just as much to do with you and your needs as what the practice provides, but I’ve seen enough young radiologists on the market to know that many people discount how isolating even local radiology can be.

When I was in training in the 2010s, there was a big push for sub-specialization. It was felt to be the future of radiology (and of course, everyone absolutely needed to do a fellowship). Observers opined that the days of the general radiologist were numbered because people needed fancier skills to deal with the increasingly complex and increasingly high-volume of complex imaging.

When the ABR ditched the original oral boards in favor of exclusively multiple-choice examinations, they pushed the final “Certifying Exam” until after fellowship and gave examinees the ability to select a portion of their testing content precisely because the idea was that everybody would be increasingly specialized, and therefore the test should accommodate that increasing specialization. (Never mind that the test was duplicative and useless—that tailoring was at least part of the attempt.)

The Flaw

One flaw in that logic is that increasing imaging volumes have increased imaging across the board. Yes, MRI and CT have disproportionately increased, but there are still plenty of plain films and ultrasounds and DEXA scans, and plenty of CTs are bread-and-butter work well within the skillset of the majority of radiologists. If everybody is so specialized and reads only in their fellowship—doing magical high-end imaging—then no one is left except the aging, near-retirement boomers to read a huge swath of high-volume, often low-RVU work. That is obviously not sustainable. The approach was inherently flawed for our times and has certainly contributed to the current shortage.

The Spectrum

Many discussions of generalist vs specialist are a false dichotomy in the sense that being generalized or specialized is more of a continuum than a binary. There are varying degrees of everything, and the shifting nature of radiology and the expectations of any given job mean that basic foundational skills can end up being important—even if they seem superfluous based on a very narrowly defined position that some radiologists, particularly in academia, find themselves in.

All points on the subspecialization continuum are available. 100% cross-sectional neuro-only? Yes. 100% subspecialized during regular weekday shifts with general radiology only on call (like evenings and weekends)? You bet. Mostly subspecialized with a daily shared pool of things like plain films? Totally. Mostly generalized with carve-outs for things like specific surgeon requests, small joint MRI, certain kinds of procedures, or breast imaging? That too. “General” may include breast imaging, or it may not.

Whatever way you think things are always done, you’re wrong. We have multiple ways to work in part because we have many different employers across 50 states, all trying to solve the question of how to best provide radiological care for patients. The fewer/larger employers we have, the fewer models we’ll continue to enjoy. (That’s one reason I like to support independent practices.)

Back to That Push for Subspecialization

There are several good reasons for increasing specialization. One is that proposed by the ivory tower: complex imaging demands greater skill, and people with more training and focus can theoretically (at least on average) provide higher-value and higher-quality care in those cases. It’s easier, on average, to be better at doing a small subset of the same things over and over again than trying to maintain a broad skillset as a jack of all trades. That narrow skillset can be brittle (all those body parts are squeezed into some tight real estate after all), but there are plenty of surgeons out there who essentially operate on one joint for the same reason.

Obviously, not every case requires marshaling our greatest diagnostic powers, but the reality is that you never know prospectively which cases do—or how to get them to the right person (please, please don’t invoke AI case assignment right now). And in many cases, retrospectively, we don’t know either. Plenty of subtle findings are missed for this reason. Radiology is the easiest field to Monday morning quarterback because the pictures are always there.

So we trade breadth for depth. This approach was once common only in academia but is now increasingly available in the broader market for several reasons—but in large part because people want it.

1. In a tight job market, many practices have had to offer more subspecialization in order to land candidates. For one simple example, an academic neuroradiologist who hasn’t read a chest x-ray in 20 years may not be willing to fill your practice’s neuro needs if you make them start reading the other stuff. So the easiest way to recruit people who are already subspecialized is to offer subspecialization.
2. Even many young people like the idea of specializing. When you spend a year of fellowship doing one thing over and over again, it’s easier to envision spending the rest of your career in a similar fashion. This can feel natural, especially since many people train in an academic environment where most attendings are similarly siloed.
3. Certainly, to an extent, a job can be “easier” in many ways because you develop and evolve your crystallized skillset faster when you’re doing the same thing in higher volume. There’s comfort there—especially when we live in a world with productivity incentives and productivity metrics, where it’s easier to hit production numbers or deal with high call volumes if you’re able to work efficiently.
4. Increasingly common productivity compensation models (e.g. flat $/RVU) encourage subspecialization because it’s easier to be fast and reasonably accurate doing a smaller number of things. This is especially true when your niche involves reading things that are higher-value, like mammograms, and you can make yourself immune to routine plain films and ultrasound. Yes, internal RVUs can mitigate some of the workload “benefits” of subspecialization, but that doesn’t change the true reimbursement value or the general nationwide trend.

Bigger Pie, Easier to Slice

Another nuance is that—thanks to regulatory demands, payor shenanigans, increasing workloads, quality bureaucracy, and recruiting/retention challenges—the increasing consolidation in the radiology space has itself enabled greater subspecialization.

A small group sharing a call burden means that everyone working alone on the weekend has to read whatever the hospital throws at them. But if multiple hospitals are consolidated into a shared worklist, then there’s enough volume and enough people working to divide out the work by subspecialty in ways that would previously have only been possible within academia.

Whereas previously fellowship training meant that the complicated cases (or the postoperative cases, or the MRIs, etc) went to the person who had done fellowship training and everything else was just shared equally, now it might mean that most if not all cases can be spread similarly.

People operating at the peak of their efficiency—which is, in many cases, more likely to occur when people have a narrow work focus—means that these large corporations, larger companies, and larger groups can also probably get more bang for their buck working with that strategy. Given the workforce shortage, any edge to getting the work done can be a big deal (also, it’s easier to squeeze a juicier fruit). For those rads in the gig economy, it’s also easier to earn a higher hourly rate when you’re reading what you can crank on.

All of this is why “body” imaging and general radiology are in such incredibly high demand—because we need people to do general radiology, especially when many radiologists have opted out.

Making General Work Pay

Long-term, this has some problems, not just because people want to practice at the “height” of their license and training, but because it’s easier to do a “full day’s work” (as measured in RVUs) reading MRIs than it is reading plain films. Adjusting the internal work values to account for the desirability of cases that nobody wants to do—the low-reimbursement, high-frustration, often tedious work of plain films and DEXA and ultrasounds—is one solution. But any change, even internally, means winners and losers. And everyone hates to lose.

The economic and spiritual degradation of general radiology has also meant that with fewer and fewer people really focusing on certain exam types, the quality of those interpretations has gone down, leaving the door open for mid-level encroachment or AI replacement of many tasks.

What Next?

The status quo isn’t going to last.

But the reality is, long-term, it’s impossible to know exactly where things will go, in part because we are at the jagged frontier of AI in radiology. It may be that the need for general radiology will continue to grow as people increasingly subspecialize and opt out of maintaining broad skills from training, older radiologists retire, and imaging volumes continue to explode.

Or, perhaps the hot job market (and fear of being inflexible in the coming AI world) will encourage some people to forgo fellowship and enough others to maintain broad skills to alleviate this pressing issue.

Or, it may be that those tasks—like ultrasounds and plain films—will be the easiest to satisfactorially offload and/or preliminary pre-draft reports from AI tools, such that we can better account for relatively low reimbursement while meeting the already acceptably low quality of those interpretations.

That being said, there’s no way to know how these tools and techniques will percolate through the broad swath of radiology tasks and radiology practices, and what radiologists’ responses to those changes will be, and what the payors responses to that utilization will be, and what the regulators will do when bad outcomes make the news, and so on and so on and so on—and therefore it’s impossible to know the ripple effects in the day to day or the broader workforce (and even later on, the radiology training pipeline).

Predictions are hard.

I would argue that, regardless of individual desires or quality differences, there are several regulatory and market forces that have pushed us toward consolidation that will be difficult to undo. And in a world of increasing consolidation, it is relatively easy to silo people into discrete boxes in ways that are not possible for small groups, especially when those people want to be siloed.

If small groups continue to thrive despite market pressures, then the model of general radiology will continue to survive.

Lastly, Fighting Automation Bias

One related question: as AI tools become more helpful, do we end up in a world where human beings must be extremely skilled in order to add value and countermand automation bias? If so, that may be the strongest and potentially most durable argument for sub-specialization.

A person who reads mostly normal brain MRIs here and there may not be able to function as an effective “liability operator” (or “sin eater“) for AI tools the same way that a subspecialized neuroradiologist could be. We’ve already seen in early trials that susceptibility to AI mistakes is experience-mediated.

So it does depend on how that dance plays out and how regulation plays a role in the implementation of AI tools going forward. There are several plausible outcomes (not to mention midlevel involvement if we can’t get our act together).

But, in the meantime, the willingness to do full-spectrum radiology is and will remain a desirable and valuable skill.