Friction Was the Point — Hisham Thabet

Copy post Most people think AI will replace workers directly. The image is familiar: a machine becomes intelligent enough to perform a job, companies automate roles, and human workers are gradually pushed aside. It is a compelling narrative because it is visible. You can point to a customer support department, a design agency, or a software team and identify where automation might fit.

I think something more subtle is happening. AI is not primarily replacing humans through competition. It is replacing us through attrition. Not the attrition of jobs, but the attrition of skills. The attrition of patience. The attrition of curiosity. The attrition of the cognitive processes that allow human beings to become capable in the first place.

The greatest threat posed by AI may not be that machines become smarter than people. It may be that people gradually become less willing to endure the discomfort required to become smart.

The compression of learning

For most of human history, learning was slow. Knowledge was passed through apprenticeships, books, mentors, trial and error, and lived experience. Developing expertise required years of repetition and exposure. Questions often remained unanswered for days, weeks, or even years. Understanding emerged through persistence rather than convenience.

Over the last century, we have steadily compressed the distance between question and answer. Books reduced the need for apprenticeships. Courses reduced the need for books. YouTube reduced the need for courses. Short-form content reduced the need for long-form content. Now AI reduces the need for almost everything that came before it.

Human progress has increasingly become a process of removing cognitive latency. Every generation gains faster access to information than the last. At first glance, this appears entirely positive. Yet hidden within this trend is an assumption that information and understanding are the same thing. They are not. Knowing the answer and understanding the answer are fundamentally different experiences. One can be acquired in seconds. The other often requires prolonged exposure to uncertainty, confusion, and failure.

Friction is the mechanism

Modern technology treats friction as an enemy. Slow websites, long queues, unnecessary paperwork, and inefficient systems are all rightly viewed as problems to be solved. However, when it comes to learning, friction serves a different purpose. In many cases, friction is the mechanism through which understanding is created.

Confusion forces attention. Repetition creates memory. Failure creates adaptation. Boredom builds endurance. Struggle creates abstraction. These experiences are not unfortunate side effects of learning; they are learning itself.

A programmer who spends eight hours tracking down a difficult bug is not merely fixing software. They are building mental models that will allow them to solve similar problems in the future. A musician practicing scales for thousands of hours is not simply repeating movements. They are developing intuition. A mathematician wrestling with an unsolved problem is not wasting time. The struggle is the education. The modern world increasingly views these processes as inefficiencies. AI takes this idea further than any technology before it.

Relief on demand

Unlike a search engine, AI does not merely provide information. It provides relief. The moment uncertainty appears, AI can dissolve it. The moment confusion emerges, AI can explain it. The moment effort becomes uncomfortable, AI can often complete the task on your behalf.

This creates a powerful psychological loop. A question arises. Uncertainty creates discomfort. The AI provides an answer. Relief follows. The brain receives a reward. The cycle repeats. In this sense, AI is not merely a productivity tool. It is personalized cognitive relief available on demand.

Social media provides entertainment on demand. Streaming services provide stimulation on demand. Food delivery provides convenience on demand. AI provides answers on demand.

The danger is not that the answers are wrong. The danger is that the brain begins to associate effort itself with unnecessary suffering. Why wrestle with a difficult concept when the machine can explain it instantly? Why spend months developing a skill when the output can be generated in seconds? Why tolerate confusion when relief is only a prompt away? These questions are rational, which is precisely why they are dangerous.

Output without understanding

AI allows people to produce things that previously required significant expertise. This is both extraordinary and deceptive. A junior developer can create applications that would have been impossible a few years ago. A student can produce essays that appear polished and sophisticated. An entrepreneur can build marketing campaigns, websites, and business plans without specialist knowledge.

Output has exploded. Understanding has not...