The great danger of the AI era is replacing depth with breadth.

I notice it in my day-to-day—spinning up multiple tasks at once for different agents to work on. Build this app. Look up this info. Draft a message for this person.

Unparalleled ability to parallelize. Very cool. But the central piece is still missing: throughput via focus.

Context switching creates entropy, and a new bottleneck emerges—the need to evaluate. Did the agent actually finish the work you assigned? Did it do so well? Now multiply that across five parallel streams.

Maybe it's worth the price. You get five tasks 80% done and spend 40% of the effort at the evaluation stage. The math adds up.

But the hidden cost of attention drain adds up too. We lose the flow of deep, focused work. When we think we can do everything at once, our attention is inevitably fractured. Multitasking was never a real thing anyway—your brain can switch quickly between tasks, but it can only focus on one thing at a time.

This is an era of experimentation, with opportunities to solve these bottlenecks. So experiment away. A new way of focusing will emerge—if we seek it.