Tzadok Weinberg on the Budget Cuts and AI Hype Reshaping American Healthcare

American healthcare is being squeezed from two directions at once. Neither is blinking.

The first is money, or, more precisely, its removal. The One Big Beautiful Bill Act, signed into law on July 4, 2025, cuts federal Medicaid spending by roughly $1 trillion over 10 years, the largest rollback of federal health support in the program's history. The nonpartisan Congressional Budget Office projects the law will result in 11.8 million more Americans going without insurance by 2034, accounting for Medicaid losses and ACA marketplace changes combined. KFF estimates the total could reach 17 million once all provisions are taken into account.

The second is artificial intelligence. The technology has produced real results: early-stage dementia detected from blood markers years before symptoms emerge, sepsis risk identified from patient data before clinical deterioration, and administrative documentation handled by AI systems that once absorbed hours of physician time each day. The hype has kept pace with the actual progress.

Tzadok Weinberg has been watching both. As an executive at Eastern Healthcare, which operates skilled nursing facilities in Virginia, and owner of Peerstar, a behavioral health company serving communities across Pennsylvania, he has a direct financial stake in how this plays out. His view is unsentimental. Neither force will rescue operators who haven't already made a more fundamental shift in their thinking about care.

The Math Operators Are Trying Not to Do

Medicaid represents roughly 20% of hospital revenue on average, rising to 40–50% for some rural facilities already operating at the financial edge, according to KFF. Under federal law, hospitals must treat patients who arrive at emergency departments regardless of their coverage status. When coverage disappears, the cost doesn't.

Weinberg has written about what that means in practice. "Medicaid cuts alone could strip coverage from millions, leaving hospitals drowning in uncompensated care and forcing rural facilities to shutter their doors." The facilities most exposed are those serving populations with the fewest alternatives: low-income elderly patients, people with behavioral health conditions, and rural communities where the nearest hospital may be hours away.

The legislation also blocks implementation of national minimum staffing requirements for nursing homes, compounding the financial pressure on skilled nursing operators with an operational constraint.

The political path from legislation to realized cuts is rarely clean, and advocacy organizations are already mounting legal challenges to several provisions. But the direction of pressure is clear. Operators planning around an assumption that federal health spending will stabilize are taking a gamble that most of them haven't fully priced in.

What AI Can Do. What It Can't.

Weinberg is neither an AI booster nor a skeptic. His position is more useful than either.

The administrative burden on American physicians has reached a well-documented level. A UCLA Health study published in NEJM AI found that physicians spend roughly two hours on paperwork for every hour of direct patient care, a figure that tracks with the 41.9% of physicians who reported at least one burnout symptom in 2025, according to the American Medical Association, with electronic health records consistently cited as a primary driver. AI scribe tools tested across 238 physicians and 72,000 patient encounters produced measurable reductions in documentation time, pointing toward real relief for a real problem. Machine learning excels at pattern recognition: reading imaging, flagging anomalies, and identifying early risk before clinical symptoms appear.

But the limits are real, too.

"Many physicians rightly argue that no algorithm can replace human healing," Weinberg has written, noting the concerns practitioners raise: accountability when the AI is wrong, patient data privacy, training data that may not reflect the actual patient population, and the opacity of many systems' reasoning.

Trust calibration is a separate problem. "Even physicians who embrace AI don't necessarily know how to use it to their patients' benefit," he has noted. "A doctor may use an AI tool dozens of times and still have no clear sense of when to trust it and when to override it."

Then there is the infrastructure problem. Many healthcare organizations run on older, siloed health record systems that were never built with external integration in mind. Retrofitting those systems to support AI tools demands capital, and budget pressure is simultaneously eroding. The organizations with the greatest clinical need may be the last to reach technological readiness.

The System AI Is Being Asked to Fix

The most pointed part of Weinberg's argument about this moment is not about the technology or the cuts. It is about what both are being applied to.

"Here's what I've learned running three healthcare businesses," he has written. "The technology question is secondary. The real problem is that our entire system is built on the wrong model."

American healthcare is built around reaction. Patients arrive sick. The system responds. Financial incentives reward treatment volume, not the quality of outcomes. Data collection is episodic. The feedback loops that could catch problems early are broken at almost every level.

"We wait for the heart attack. We wait for the stroke. We wait for the cancer diagnosis," he has written. "This is not healthcare. This is sick care."

That distinction shapes how operators should approach the current moment. AI deployed into a reactive system accelerates reactions rather than changing the model. What Weinberg describes is something prior to any technology decision: a shift in philosophy from treating illness to preventing it.

The Playbook

Weinberg describes a response for providers navigating both pressures simultaneously.

Patient education comes first. Every clinical encounter, he argues, should include some investment in teaching: sleep, movement, nutrition, and stress management. Published research has found that approximately 80% of chronic disease is preventable through foundational lifestyle factors. A system oriented toward prevention doesn't just produce better outcomes. It reduces the volume of expensive acute interventions that budget cuts make harder to absorb.

The second element is building lower-cost touchpoints that maintain continuity of care. Virtual consultations, nurse-practitioner-led monitoring, and automated follow-up reminders keep providers connected to patients between the acute visits that dominate the current model. Early intervention at low cost is fundamentally different from emergency treatment after a condition has progressed.

The third is connecting patients to their own data. Wearables already generate substantial health information. Interpretation is the gap. Providers who help patients understand what their resting heart rate, sleep patterns, and activity levels actually mean can build a continuous feedback loop that the current episodic model structurally prevents.

"The providers who will thrive," Weinberg has written, "are those who integrate these tools into their care models, helping patients select appropriate technologies, interpret the data, and take action before problems become crises."

Budget cuts and AI will reshape the industry regardless of whether any individual operator is ready. The question Weinberg keeps returning to is simpler and harder than either: when did healthcare decide it was acceptable to wait until people got sick?