In an opinion piece this week for Veterinary Practice News, Dr. Patty Khuly reflected on how veterinary medicine has shifted from historically avoiding referrals to now, in her view, over-relying on them. She contrasts her early career experience in a practice that discouraged sending cases out, even at the expense of patient care, with today’s environment, where failing to refer can carry legal and professional risk.

Dr. Khuly argues the pendulum may have swung too far. Younger veterinarians are increasingly quick to refer cases they are capable of handling, driven by factors like time pressures, skill insecurity, and training that emphasizes “ideal” care over practical realities.

A few weeks ago we wrote about how to keep younger vets engaged at work, and one of the themes that stood out was mentorship. Khuly echoes that idea, framing a lack of mentorship as the throughline behind rising referral rates. In her view, key drivers include:

Her core concern is that excessive referrals may limit both access to care and the development of clinical skills in general practice. While specialist care remains critical, she argues for a more balanced approach, one that reinforces mentorship and ensures GPs continue building confidence and capability. She worries that excessive referrals may limit both access to care and the development of clinical skills in general practice. While acknowledging the importance of specialist care, she calls for a more balanced approach that preserves learning opportunities for veterinarians while still prioritizing patient outcomes.

As the fantastic team at VIN News reported this week, more veterinarians are hearing a new pitch from corporate groups: instead of just salary or a bonus, why not take a piece of the business? On paper, it is compelling. You get shares in a practice or a larger network, maybe some quarterly distributions, and the sense that you are building something alongside your employer. For associates who have never had ownership, it can feel like a long-awaited seat at the table.

But the reality behind those offers is often less straightforward. Many of these shares are effectively locked boxes. You cannot easily sell them, you may not know what they are truly worth, and whether you ever see a payout depends on events you do not control, like a future sale or recapitalization. Contracts can also include clauses that claw back equity if you leave, or delay ownership through vesting schedules that never quite materialize. Layer in strict noncompetes, sometimes covering wide geographic areas, and what looked like an opportunity can start to limit your mobility and flexibility.

The risks become even more pronounced in joint ventures or heavily leveraged corporate settings, where veterinarians may carry financial exposure without real decision-making power. At the same time, softer practice revenues and rising costs are putting pressure on margins, which can directly impact equity value. There are scenarios where these deals work out well, but most advisors are consistent in their guidance: take as much cash upfront as you can, negotiate every term, and view equity as a high-risk bet, not guaranteed upside. If a group resists transparency or refuses to negotiate, that is often the clearest signal to proceed with caution.

This week, you may have heard the story of Australian tech entrepreneur Paul Conyngham, who used tools including ChatGPT, genomics work, and collaborators at UNSW to create a personalized mRNA cancer vaccine for his dog, Rosie, and her extensive mast cell tumors.

To me, the story is fascinating precisely because the truth of the matter is is not the sensationalized headline it became. While many outlets and tech moguls suggested the dog was “cured by AI”, it misses the reality of the situation and in turn does a disservice to the remarkable, but nuanced, story. Rosie was not simply “cured by AI.” Reporting makes clear that AI helped Conyngham brainstorm and interpret information, but the actual work still required tumor sequencing, expert researchers, vaccine manufacturing, and real-world treatment. Rosie reportedly improved, but one tumor did not respond, and the vaccine was also given alongside other therapies. In other words, this is not “guy with chatbot solves cancer in his garage.” Don’t get me wrong… it is still remarkable but it’s more like, “motivated owner uses AI to help navigate a high-end scientific workflow that still depended on specialists, infrastructure, money, and luck.” This is what a future in vet med might look like. Owners with means seeking personalized care, research progressing in previously unknown or slowly progression ways.

But this idea that AI can “cure” something raises red flags especially in light of other recent stories. There is a growing body of discussion around what some researchers call “AI psychosis,” though they use that as a heuristic label rather than a formal diagnosis. The concern is that chatbots can be overly affirming, reinforce delusional or grandiose thinking in vulnerable users, and give people a false sense that they have cracked problems far beyond their expertise. Nature has reported that chatbots can reinforce delusional beliefs in rare cases, and OpenAI itself acknowledged last year that one ChatGPT update had become overly agreeable or “sycophantic.” The lighter, non-clinical version of that problem is everywhere already: AI gives people just enough fluency to feel like they are the lead author, attending clinician, and biotech founder all at once.

We recently got an interesting look at the future of work from Andrej Karpathy, one of the most influential thinkers in modern AI. Karpathy is a former OpenAI co‑founder, led Tesla’s self‑driving efforts, and now runs Eureka Labs, while frequently publishing exploratory tools and experiments on the side.

Take a moment to explore the US Job Market Visualizer, an interactive treemap of BLS labor data showing 342 occupations sized by employment and colored by metrics like pay, projected growth, education requirements, and AI exposure. The AI exposure scores are generated by an LLM and are explicitly exploratory rather than predictive, but the patterns are still revealing.

In veterinary medicine, roles that are deeply physical score low. Veterinary assistants and lab animal caretakers land at 2/10 for digital AI exposure, veterinarians at 4/10, and veterinary technologists and technicians at 5/10. The logic is intuitive. The more a job involves hands‑on animal handling, anesthesia, surgery prep, emergency response, and in‑room judgment, the harder it is to digitize. The more it involves interpretation, records, imaging, lab data, and administrative work, the more AI can assist.

Importantly, assist is the right word. Karpathy is explicit that exposure does not mean disappearance, and the BLS still projects strong growth for veterinarians, technicians, and assistants over the next decade. The work is not going away, but it is being reshaped.

AI will continue to redistribute work, moving veterinary work away from some of the screen-heavy, repetitive, synthesis-heavy tasks and toward higher-leverage judgment, communication, and hands-on care. What we need to do is to figure out how to safeguard the profession in the interim to ensure that evidence based medicine prevails. When we look at the potential, and some of the concerning trends including AI “psychosis”, it is clear we still need experts, guardrails, and a healthy respect for the difference between “this is promising” and “I have solved medicine.”

The future veterinarian is amplified, and probably still running late with a hectic schedule, because some Labrador ate a sock.

-RBA

New Brunswick plans to wind down its provincial veterinary field service and transition large-animal care to the private sector over the next two years, with equine services ending in 2026, core field services by 2027, and laboratory services by 2028. The government cites cost to taxpayers and says it will support veterinarians in moving to private practice. However, farmers and animal owners are having a cow (pun intended), as many say there was no prior consultation and describe the current system as essential for herd health, emergency response, and routine preventive care.

Producers and veterinarians are raising concerns about access and system capacity under a fully privatized model. In rural regions with dispersed farms, the existing provincial service provides rapid, 24/7 response that many believe private practices will struggle to replicate.

Other provinces have similar systems, like Ontario, but they also act as a case study for some of the challenges: longer wait times, limited after-hours coverage, and geographic barriers. Laboratory services are also a major concern, with stakeholders pointing to implications for disease surveillance, food safety, and timely diagnostics.

This shift in policy affects workforce distribution, access to care, and public health infrastructure, and veterinarians will face both opportunity and risk. While privatization could open doors for new practices, there are questions about economic viability in low-density areas and whether enough food-animal vets will enter or remain in the field. More broadly, reduced access to timely veterinary and lab services could impact animal welfare, outbreak response, and producer confidence, placing additional pressure on an already constrained large-animal workforce.

Here are some of the other stories that caught our eye and we're following this week from around the veterinary world and animal kingdom: