For decades, computers have gotten faster. Exponentially so. They crunch numbers in milliseconds, render high-definition video in minutes, and analyse vast datasets with ease.
But what remains the slowest operation in computing?
Clicking the button.
Not the literal act of clicking, but what it represents—the moment a human has to think, make a decision, and figure out what needs to be done.
The Human Bottleneck
Computers execute tasks at blinding speeds. But most work isn’t about raw execution. It’s about deciding what to execute.
Writing code isn’t slow because of the compiler—it’s slow because you have to design the logic, debug errors, and make architectural choices.
Spreadsheet analysis isn’t slow because of Excel—it’s slow because you have to think through the numbers, their meaning, and what insights they provide.
Graphic design isn’t slow because of rendering—it’s slow because creativity takes time, iteration, and refinement.
Computers wait on us.
Automation vs. Thinking
If a task is repetitive, we can remove the human. We write scripts. Automate workflows. Let the machine handle the drudgery.
But most of our work isn’t repetitive. It’s one-off.
Every month, a script can reformat sales figures in Excel to generate a report. But what happens when you need to analyze a new dataset, answer an unexpected question, or make a strategic decision? Automation doesn’t help. Thinking is required.
And thinking doesn’t scale exponentially.
The Limits of Speed
AI models like ChatGPT can generate text in seconds, but only after a human crafts the right prompt, evaluates the output, and decides what to refine.
Cloud computing can process terabytes of data, but a human still has to decide which data matters and why the analysis is being done in the first place.
GPUs render stunning game graphics instantly, but designing a compelling game world is a slow, human-driven creative process.
Technology accelerates execution. It doesn’t accelerate thought.
The Real Challenge
No innovation—no AI, no cloud computing, no hardware breakthrough—will remove the fundamental bottleneck of human cognition.
The best we can do is build tools that support thinking, structure decisions, and reduce unnecessary friction. But ultimately, the speed of work is limited not by how fast a machine can execute, but by how fast a human can decide what needs executing.
And that will always be the slowest part of computing.