Thinking Technologies

Around 370 BCE, Socrates argued that writing would destroy human memory. People would stop exercising their minds because they could rely on external marks instead of internal recall. They would appear wise without actually being wise.

In the 1470s, the Venetian humanist Hieronimo Squarciafico warned that the printing press would make people less studious. The abundance of books would enfeeble the mind by relieving it of too much work.

In the 1970s and 1980s, educators worried that calculators would prevent students from developing mathematical reasoning. They would become dependent on devices. Basic skills would atrophy.

The fear followed the same structure each time. The new technology would destroy the very capacity it was designed to support. Memory, studiousness, reasoning. Each time, the technology did not destroy the capacity. It enabled something that had not existed before.

Before writing, knowledge passed from person to person through speech. What one generation learned had to be memorized by the next, or it was lost. The amount of knowledge any society could maintain was limited by what its members could hold in their heads.

Socrates was correct that writing changed memory. People did rely less on internal recall when external records became available. The elaborate memory techniques that educated people once trained, the memory palaces and mnemonic systems, became less essential.

What Socrates did not anticipate was what writing made possible. A scholar could now engage with the thinking of someone who died centuries earlier. Ideas could accumulate across generations in ways that oral transmission could not support. Contradictions could be identified by comparing texts. Arguments could be refined over time.

Writing also changed the nature of thinking itself. The act of putting words on a surface forced a kind of clarity that speech did not require. You could see your own reasoning laid out, identify gaps, revise. People began to write in order to understand what they thought. The technology did not replace thinking. It created a new form of thinking that had not existed before.

Before the printing press, books were copied by hand. A single book might take months to produce. The largest European library in 1300 held about 300 manuscripts. Errors crept in with each copy. No two versions of the same text were identical.

Squarciafico was correct that printing reduced the need for memorization. When books were scarce, committing texts to memory was essential. When books became widely available, that particular skill became less critical. By the 1490s, a printed copy of Cicero cost about a month's wages for a schoolteacher.

What Squarciafico did not anticipate was what printing made possible. For the first time, scholars across Europe could reference identical editions of the same text. Scientific observations could be compared reliably. Errors could be identified and corrected. The standardization of knowledge that printing enabled was essential for the development of modern science.

Printing also transformed who could participate in intellectual life. Learning shifted from the lecture hall to the page. Book learning, previously the domain of monasteries and universities, became part of ordinary childhood. Ideas that authorities wanted suppressed became impossible to contain. Every time the Church published a list of banned books, printers knew exactly what would sell.

Before affordable calculators, mathematical computation required either extensive training or tedious manual work. Complex calculations were bottlenecks. Many practical problems that had mathematical solutions remained unsolved simply because the calculations were too time-consuming.

The educators of the 1970s were correct that calculators changed which skills students practiced. Less time went to long division and manual computation. The hours that students once spent on arithmetic drills decreased.

What those educators did not fully anticipate was what calculators made possible. Students could now engage with mathematical concepts without getting lost in computational mechanics. Problems that would have taken hours by hand could be explored in minutes. The focus could shift from executing calculations to understanding when and why to calculate.

Research from the 1980s and 1990s found that calculator use did not damage mathematical reasoning. The Cockcroft Report of 1982 concluded that calculators in no way reduced the need for mathematical understanding. If anything, they revealed understanding more clearly. A student who did not understand the problem could not use the calculator effectively. The tool amplified capacity rather than replacing it.

The fears about AI follow the same structure. People will stop thinking because machines can think for them. They will become dependent. The capacities that made them valuable will atrophy.

If the pattern holds, these fears are looking at the wrong thing. Writing, printing, and calculators did not destroy the capacities they augmented. They enabled transformations that were difficult to anticipate at the time. Writing enabled the accumulation of thought across generations. Printing enabled the scientific revolution. Calculators enabled mathematics that would have been impractical by hand.

What might AI enable? The honest answer is that we do not yet know. The transformations that previous technologies made possible were not fully visible to the people who first encountered them. Socrates could not have anticipated the scholarly traditions that writing would create. Squarciafico could not have imagined the scientific method that standardized texts would support.

What we can observe is what AI handles well and what it does not. AI is remarkably good at processing information, recognizing patterns, and executing tasks that can be systematized. It is less capable when situations require judgment that cannot be fully specified in advance, when the right question matters more than quick answers, when context and values must be weighed in ways that resist standardization.

Previous technologies shifted value toward what they could not do. Writing could not synthesize across sources or recognize what mattered. Printing could not analyze or interpret. Calculators could not determine what to calculate or whether the result made sense. The human role shifted toward the capacities each technology lacked.

If AI follows this pattern, it may shift value toward judgment, creativity, and the kinds of understanding that come from experience rather than data. It may enable humans to spend less time on what can be systematized and more time on what cannot. What that actually looks like remains to be seen.