Industry Social Media and Updates

Don't Bet on One AI Vendor: Claude Beats ChatGPT, Gemini

The AI tool I called “best in class” just got dethroned - in three months. So when health systems go all-in on one AI vendor, is that a BIG mistake? I was loyal to ChatGPT. Then I became Gemini-pilled. Recently my CTO got me hooked on Claude. Why? I've been running the same prompts across Claude, Gemini, and ChatGPT to compare results… and the differences are stark: → When I asked Claude to redesign our website, it didn't describe what it would do - it just started building a prototype I could see and interact with. Gemini and ChatGPT? They simply wrote text descriptions. → When I asked Claude to draft a strategic plan, it asked clarifying questions before producing something that felt McKinsey-level. Gemini produced something just okay. ChatGPT produced a high school essay with random emojis thrown in. Now Claude isn't perfect - it's made a few mistakes that I've had to catch and correct. But in general? It's SO much better at most things, it’s really not close right now. A month ago I was all-in on Gemini. Today? I'm probably cancelling my ChatGPT subscription and maybe my Gemini one soon too. So yes - I'm all-in on Claude… but how long will THAT last? Which got me thinking about how many health systems are approaching AI right now... What if you'd forced every employee onto a single enterprise ChatGPT license last year - and locked them in for the next three years? You'd have missed the boat with Claude. Health systems are under enormous pressure right now to standardize, lock-in enterprise controls and minimize costs. Pick one enterprise AI scribe, one enterprise AI CDS tool, etc. Roll it out to every clinician. And I get why: standardization feels safe. It's good governance. It feels like the responsible thing to do. But what if we're in the ChatGPT moment for many Healthcare AI tools right now… and we just don't know it yet? What if the tool you're betting the enterprise on today gets lapped in 12 months? What if a clinician who had the freedom to use a different tool (or two) would've made a better call for their patient? Sure, there are real reasons health systems push for standardization: IT governance and security, EHR-integrated workflows, liability, etc. But they risk becoming an excuse to lock in mediocrity at scale. Real question: is there a version of this where we can give clinicians access to a broader set of AI tools - with guardrails - and actually let the best ones surface through real clinical use? I don't think the answer is pure chaos - shadow AI tools are a governance nightmare. But depending on the use case, the answer may not be forcing everyone onto one enterprise tool and hoping it's anywhere close to the best one in a year. That clinicians can easily switch between OpenEvidence and DoximityGPT as consumers is perhaps a good thing. Because ultimately, the patient doesn't care which vendor won the enterprise contract.