You wouldn’t put your best surgeon on circulating nurse duty just because they’re available. But that’s exactly what most people do with their AI tools. One model, every task, all the time.
In an OR this isn’t abstract. The team works because everyone has a defined role they’re trained for and trusted with. The anesthesiologist is not doing what the scrub tech does. The surgeon is not doing what the circulating nurse does. The moment someone starts reaching outside their lane because you’re short-staffed or didn’t plan well, the margin for error shrinks fast. Good outcomes in that room come from role clarity, not from finding one person who can handle everything.
But almost nobody applies this thinking to their AI tools. Most people pick one model, get familiar with it, and use it for everything. Drafting a newsletter, pulling research for a brief, restructuring a coaching framework. One tool, all tasks, all the time. That’s not a workflow. That’s a habit that looks like a strategy.
Claude thinks carefully and handles nuanced reasoning well. GPT tends to be fast and strong on structure. Perplexity is built specifically for research and sourcing. These are not the same tool wearing different logos. They have genuinely different strengths, different failure modes, and tasks where they quietly underperform. Treating them as interchangeable is the same mistake as handing a scalpel to the wrong person because they’re smart and available.
There’s also a longer game worth thinking about. Every field that matures follows the same pattern. Medicine started with general practitioners. Then specialties emerged. Then subspecialties within those. Cardiology, then interventional cardiology, then minimally invasive interventional cardiology. The same arc happened in law, engineering, finance, and consulting. Generalists dominate early because nobody knows exactly what the specific problems are yet. Once the landscape clarifies, specialists win.
AI is early. The models today are all trying to do everything, and some of them are remarkably good at a lot of things. But there is no reason to believe AI skips the pattern every other field has followed. The more likely future is that models become more specialized over time, built for specific domains, specific tasks, specific types of reasoning. Plan your workflow with that in mind now, not after the shift happens.
The bigger trap is lock-in. If your entire workflow is built inside one company’s product, you can’t adapt when things change. Switching means losing everything you’ve built up: the context, the history, the accumulated understanding of who you are and what you’re working on. So you stay. Not because it’s the best tool anymore, but because leaving feels too expensive. That’s not loyalty. That’s a structural problem that gets worse the longer you ignore it.
The best workflows run multiple models simultaneously, each doing what it does best, without losing continuity between them. That requires your context to actually travel with you. Not backed up somewhere. Not exportable if you ask nicely. Actually yours, available to whatever tool you reach for next.
That’s the problem Open Brain solves. Your context lives in a layer you control, above any single model. The decisions you’ve made, the projects you’re running, the way you think and communicate. Any model can read from it. I started using it because I kept losing continuity every time I switched tools or hit a context limit, and rebuilding that context every time was overhead I couldn’t afford. Anesthesia residency does not leave a lot of spare hours. This week I drafted this newsletter in Claude using context from Open Brain, then handed the same brief to GPT to restructure it. Took about four minutes. Without Open Brain, the re-explaining alone would have eaten ten. Both models knew who I was and what I was building. No re-explaining. No starting over.
When your memory is portable you can actually build a team. You can put the right model in the right role, switch when something better comes along, and stay agile when the landscape shifts.
The best teams I have been part of worked because everyone knew their role and trusted the people next to them to do theirs. That’s what made them more than the sum of their parts. Your AI workflow deserves the same thinking.
Figure out your depth chart.
The Short Version
Most people run one AI model for everything. That’s like staffing an OR with one person. Different models have different strengths, and if your context is locked inside one of them, you can’t build a real team. Open Brain makes your context portable so you can put each model in the right role.
Try This Now
Pick one task you did with AI this week. Run it through a second model you don’t normally use. When I ran the same research brief through Perplexity instead of Claude, the sourcing was noticeably tighter. It takes ten minutes and it changes how you think about your workflow.
Carlo DelDonno MD, CPT
If this was useful, forward it to one person. That’s the whole ask.