The Cost of Switching: An AI Lock-In Experiment (Post 4 of 6)
Fragmentation vs. Convergence: Too Many Choices
By this point in my switching experiment, I had a new appreciation for just how fragmented the AI landscape really is. On paper, we talk about “ChatGPT vs. Claude vs. Gemini” as if that’s the whole story. In reality, those are just three generalist tools sitting at the top of an ecosystem that’s exploding with options.
The AI-Native Office Suite research from Andreessen Horowitz maps it well: there are horizontal generalists that try to do everything, ChatGPT, Claude, Gemini, Manus, Genspark and vertical specialists that go deep into a single workflow like slides (Gamma), notes (Notion, Mem, Granola), spreadsheets (Shortcut, Julius), or email (Serif, Jace) .
Each comes with tradeoffs. Generalists give you breadth, but often feel rough around the edges. Specialists polish a single workflow, but then you’re stuck hopping between tools. It’s a buffet that looks great until you realize you can’t possibly eat everything.
This is where the paradox of choice kicks in. Behavioral economics backs this up. In Predictably Irrational, Dan Ariely cites the famous jam experiment by Sheena Iyengar and Mark Lepper: shoppers were presented with 24 varieties of jam at one table and 6 at another. More people stopped at the table with 24, but far fewer actually made a purchase. The overwhelming number of options created paralysis. With fewer choices, people were more decisive.
AI tools feel exactly like that jam table. With so many options, the cognitive load of evaluating them becomes overwhelming. At some point, you settle for “good enough.” For me, that meant Gemini had some neat features, but not enough to justify retraining myself when ChatGPT already felt “good enough.”
And here’s the irony: the abundance of options doesn’t actually make switching easier. It makes it harder. Because once you find something workable, the inertia to stick with it only grows stronger.
When it comes to AI tools, do you find yourself chasing the newest option or do you settle into “good enough” to avoid the fatigue of constant switching?
