With the healthcare landscape being reshaped by labor disputes, soaring operational costs, and rapid technological advancements, we sit down with Dominic Jainy, an IT professional with deep expertise in AI, machine learning, and blockchain. Today’s discussion will explore the intricate dynamics of the ongoing nurses’ strike in New York, the long-term financial and operational repercussions for hospitals, and the revolutionary potential of new AI “world models” in healthcare. We’ll also delve into the practical challenges of implementing predictive tools in clinical settings and examine the forces driving major investments in specialized fields like pediatric care.
With nurses striking over staffing levels, pay, and security, how do these three issues typically influence one another during negotiations? Please describe the operational and financial challenges hospitals face when trying to address these interconnected demands simultaneously.
These three issues—staffing, pay, and security—are fundamentally intertwined, creating a complex knot for hospitals to untangle. You can’t really solve one without addressing the others. For example, low pay and unsafe working conditions, like the reported violence at Mount Sinai, directly lead to poor staffing because nurses leave for better, safer jobs. This then puts more pressure on the remaining staff, making their jobs even more dangerous and stressful, which in turn fuels demands for higher pay as compensation for the risk and workload. From a hospital’s perspective, this is a financial tightrope walk. Addressing staffing by hiring more nurses directly increases payroll costs. Improving security requires investment in personnel and infrastructure. And raising pay across the board is, of course, a massive budget item. When you have nearly 15,000 workers striking, the pressure to meet all these demands at once is immense, and it forces a difficult conversation about resource allocation, especially when they’re already facing external pressures like federal subsidy cuts.
Hospitals are reportedly spending over $100 million on temporary staff and hotels during the strike. Beyond the immediate financial cost, what are the long-term impacts of such spending on hospital budgets, patient care quality, and the morale of non-striking staff?
That $100 million figure is a staggering immediate cost, but the ripple effects go much deeper and last much longer. Financially, that’s a massive, unplanned expenditure that has to come from somewhere. It likely means delaying investments in new medical equipment, facility upgrades, or innovative patient care programs for years to come. In terms of patient care, bringing in temporary nurses who are unfamiliar with a hospital’s specific protocols, its electronic records system, and the unique patient population can introduce risks and inefficiencies. It disrupts the continuity of care that is so critical for good outcomes. Then you have the morale issue. The non-striking staff, who are already under pressure, see these vast sums being spent on temporary replacements, which can create a sense of resentment and devalue their own loyalty and hard work, potentially leading to a new wave of resignations after the strike ends.
Yann LeCun’s new AI venture is focusing on “world models” that learn from video and spatial data. How might this technology move beyond current AI, like transcribing doctor visits, to solve more fundamental problems in diagnostics or patient monitoring? Please provide a specific example.
This is a really exciting leap forward. Current AI, like the kind used by Nabla to transcribe visits, is fantastic at processing language and text. But “world models” aim to build a much deeper, more intuitive understanding of reality by learning from visual and spatial data, much like a human does. In healthcare, this could be revolutionary. Imagine an AI that doesn’t just read a radiologist’s report but has learned from thousands of hours of surgical videos and dynamic MRI scans. It could potentially identify subtle, early-stage signs of a disease like pancreatic cancer not just from a static image, but by recognizing minute, abnormal changes in tissue movement or blood flow over time—patterns a human eye might miss. This technology moves from simply documenting what a doctor says to actively assisting in the diagnostic process by understanding the physical, three-dimensional reality of the human body.
Epic’s ‘Curiosity’ tool aims to improve hospital operations by predicting patient discharge dates. What are the practical, step-by-step challenges in implementing such a predictive system, and how can clinicians be trained to trust and effectively use its outputs in their daily workflow?
Implementing a tool like ‘Curiosity’ is far more complex than just flipping a switch. The first major hurdle is data integrity. The system is trained on 100 billion patient events, but for it to work accurately at a specific hospital, it needs clean, consistent, and correctly entered local data. That’s a huge operational challenge. The next step is workflow integration. You can’t just add another screen for a busy doctor or nurse to check; the prediction needs to appear seamlessly within their existing charting or patient management system at the exact moment it’s useful. The biggest challenge, though, is building trust. Clinicians are trained to rely on their own judgment. To get them on board, you have to start with low-stakes, operational predictions, just as Epic plans. By accurately predicting discharge dates for a 600-bed hospital and demonstrating how that smooths out patient flow, you build credibility. It starts to “seem like Epic got smarter,” and over time, that trust allows you to introduce more critical clinical predictions that they’ll be willing to consider.
Zarminali Pediatrics, a tech-enabled pediatric care company, recently raised $110 million. What specific technology or service model do you think is attracting such significant investment in pediatrics, and how might it change the way families access care for their children?
The $110 million investment in Zarminali points to a huge demand for a more integrated and accessible model of pediatric care. The “tech-enabled” part is key. This isn’t just about telehealth visits; it’s about creating a seamless ecosystem for parents. Think of a platform that combines virtual urgent care, direct messaging with pediatricians for quick questions, scheduling for in-person visits, and management of complex conditions, all in one place. The story of the CEO starting the company after struggling with his own daughter’s care is very telling—it highlights a personal, deeply felt need in the market. This model is attracting investors because it addresses major pain points for modern families: convenience, speed, and coordinated care. It could fundamentally change access by reducing the reliance on inconvenient office visits for every minor issue and providing parents with a trusted, on-demand resource, which is invaluable.
Considering the labor disputes in New York and the potential for a large strike in California, how do these actions reflect broader, systemic pressures within the U.S. healthcare industry? Please share your thoughts on what this might mean for provider-patient relationships in the coming years.
These strikes, both the one in New York and the potential action by 31,000 workers in California, are symptoms of a system under immense strain. They aren’t isolated incidents but a reflection of years of mounting pressure on healthcare workers from understaffing, burnout, and the feeling that the business side of healthcare is prioritizing profits over people—both patients and providers. This escalating tension is bound to affect the provider-patient relationship. When providers are stretched thin and feel unsupported, the quality of their interactions can suffer, not from a lack of caring, but from sheer exhaustion and moral injury. Patients, in turn, may feel the effects through longer wait times and rushed appointments. In the long run, if these systemic issues aren’t addressed, we could see a further erosion of trust in the healthcare system as a whole.
What is your forecast for the adoption of predictive AI in clinical settings over the next five years?
Over the next five years, I predict we’ll see a two-tiered adoption of predictive AI. The first tier, which will become widespread, will be operational AI, much like what Epic is starting with. Tools that predict patient flow, optimize surgical schedules, and manage bed capacity will become standard because they demonstrate a clear and immediate return on investment for hospital administrators. The second tier, true clinical decision-support AI—the kind that predicts pancreatic cancer or suggests a specific test for an earache—will see a slower, more cautious adoption. It will likely begin in specialized academic centers where it can be rigorously validated. Full, mainstream adoption will depend heavily on solving challenges around regulation, data privacy, and, most importantly, earning the trust of the physicians who will ultimately be responsible for acting on its predictions.