The Tools We Had

By: John F Groom

How much of an accomplishment belongs to the person, and how much belongs to the tools available to them? This article is itself an example of what it is about. Not very long ago, if I wanted to write an article like this, I would have had to do almost all of the work myself. I would think about the idea, struggle with the structure, write a draft, decide that parts of it did not work, rewrite them, move paragraphs around, look up examples, check facts, and eventually produce something that I hoped expressed the original thought clearly. Depending on the subject, that might take hours. Sometimes it might take days.

For this article, I explained the basic idea to an artificial intelligence system. AI did much of the work of turning that idea into finished prose. I read it, criticized it, changed things I did not like, added things that were missing, and decided whether it actually expressed what I wanted to say. Even with several rounds of revision, the amount of human labor required was radically less than it would have been only a few years ago.

Does that make the article less valuable? Not necessarily. But it does change what is impressive about producing it. If I had written an article of similar quality twenty years ago, the finished article would not reveal the many hours of labor behind it. Conversely, someone reading this article today cannot tell merely by looking at the finished product how much of the mechanical work was done by AI. That suggests a useful way of thinking about accomplishment generally. We normally judge people by what they produced. Perhaps we should also ask a second question: What tools did they have available to produce it?

Building Chartres

Consider Chartres Cathedral. Construction of most of the great Gothic cathedral that survives today began after a fire in 1194. Its builders created an enormous stone structure with soaring vaults, flying buttresses, towers, sculpture, and some of the most celebrated stained glass ever produced. More than eight centuries later, people still travel across the world simply to stand inside it.

Now put Chartres next to a modern skyscraper. The skyscraper may be vastly taller, contain enormously more usable space, and incorporate technologies that the builders of Chartres could not possibly have imagined. By almost any objective engineering measure, the modern building is the more technologically advanced object.

But modern builders also possess structural steel, reinforced concrete, tower cranes, electric motors, hydraulic equipment, computer-aided design, laser surveying, standardized components, trucks, elevators, power tools, and centuries of accumulated engineering knowledge. Before construction begins, engineers can model loads, stresses, and wind forces with computers. Materials can be manufactured thousands of miles away to precise specifications and delivered through an elaborate global transportation system. A modern construction site sits on top of an immense technological civilization.

The people who built Chartres had almost none of this. Stone had to be quarried, shaped, and transported using human and animal power supplemented by relatively simple machines. Enormous masses had to be raised high into the air without powered cranes. Measurements were made with comparatively simple instruments.

The builders were working with accumulated knowledge, of course. They were not ignorant people improvising from scratch. But the technological base available to them was tiny compared with ours. This leads to a curious result. A modern skyscraper is a much more technologically sophisticated object than Chartres Cathedral, while building Chartres may nevertheless represent the more astonishing human accomplishment. We are comparing not simply two buildings, but what human beings managed to achieve from two radically different technological starting points.

Shakespeare's Word Processor

The same principle applies to intellectual accomplishment. Imagine Shakespeare sitting down to write a play. There was no word processor, no search function, no copy and paste, no cloud backup, no instantaneous dictionary, no searchable database of historical sources, and certainly no AI system capable of reorganizing a scene or producing ten alternative versions of a troublesome passage in seconds.

Writing meant physically writing. Revision meant physically revising. Moving a passage was not a matter of highlighting it and dragging it somewhere else. Making another copy required another physical copy to be made. Research required books, memory, conversation, or access to someone who possessed the relevant knowledge. Even the physical act of producing and preserving a long manuscript imposed costs that modern writers almost never think about.

Information was also fragile. Today I can write something and create multiple identical copies of it at essentially zero marginal cost. They can exist on my computer and on remote servers in several places. I can lose the physical computer on which I wrote something without necessarily losing a single word. For most of history, losing the physical object could mean losing the work itself. Fire, water, carelessness, theft, or simple deterioration could destroy months or years of intellectual effort. The survival of knowledge depended on physical preservation and, often, on someone taking the trouble to make additional copies.

None of this is why Shakespeare was Shakespeare. Giving a mediocre writer a quill does not make him a genius, and giving Shakespeare a laptop would not somehow diminish Hamlet. His accomplishment rests primarily on what he was able to see and express about human beings. But understanding the tools with which he worked adds another dimension to our appreciation of what he accomplished.

The Tools of Survival

The idea extends beyond the accomplishments that end up in museums and history books. Some of the most important examples involve things ordinary people have always done. Consider childbirth. Giving birth to another human being is a momentous event in any century, and childbirth can still be dangerous today. But a woman giving birth in a modern developed society normally does so within an extraordinary technological system that has become so familiar that it is almost invisible.

Antisepsis, antibiotics, blood transfusions, anesthesia, fetal monitoring, laboratory testing, ultrasound, trained specialists, operating rooms, and emergency Caesarean sections all stand behind an event that human beings have been experiencing since our species began. For most of human history, those tools did not exist. Pregnancy therefore meant accepting a level of physical risk that is difficult for people living in a modern wealthy society to intuitively appreciate. A hemorrhage, infection, obstructed labor, or other complication that can often be treated today could once kill the mother, the child, or both.

The underlying biological event has not changed very much. The technological environment surrounding it has changed enormously. A woman giving birth hundreds of years ago was doing something recognizably similar to what a woman does today, but under a very different set of risks and constraints. If we want to understand the human experience rather than merely describe the event, the tools matter.

Technology Changes the Difficulty Level

We usually think about technology by asking what new things it allows us to do. But technology also changes the difficulty level of things humans have always tried to do. Building a structure, writing down an idea, crossing an ocean, giving birth, finding food, communicating with someone far away, calculating a number, preserving information, and finding your way home are ancient human problems. The objectives can remain surprisingly constant while the difficulty of accomplishing them falls by orders of magnitude.

A medieval traveler and a modern airline passenger can both travel from London to Rome. If we record only the starting point and destination, they appear to have accomplished the same thing. But almost everything about the actual accomplishment is different. One journey might consume weeks, expose the traveler to disease, robbery, bad weather, and unreliable transportation, and require substantial knowledge and planning. The other may require buying a ticket online and spending a few hours sitting in an airplane.

This creates a problem whenever we compare people across time. We tend to compare outputs without adjusting for tools. A person who calculates something in seconds with a computer may produce a more accurate answer than a mathematician who spent days doing calculations by hand. A modern photographer can take thousands of technically excellent photographs at almost no incremental cost. A modern writer can research, draft, revise, typeset, and distribute an article around the world without leaving a desk.

The modern output may be better. But better output and greater accomplishment are not necessarily the same thing. Perhaps we need something analogous to adjusting money for inflation. Comparing $1,000 in 1920 with $1,000 today is misleading because the units have the same name but represent different purchasing power. We therefore adjust the numbers before making comparisons.

We might think about tool-adjusted accomplishment in a similar way. Instead of asking only, "What did this person produce?" we would also ask, "What did this person produce given the tools, knowledge, infrastructure, and resources available to them?" This would change some historical comparisons dramatically. It would also prevent us from making the opposite mistake. The point is not to romanticize primitive technology or claim that everything people did in the past was more impressive. Modern humans attempt things that earlier civilizations could not attempt at all.

Designing an advanced semiconductor, operating a global communications network, or landing a spacecraft on another planet requires forms of accumulated knowledge and coordination that would have been inconceivable to the builders of Chartres. The important distinction is between the accomplishment and the technological platform supporting it. Every human being begins with a technological inheritance. None of us invented the world in which we work.

AI Changes the Calculation

Artificial intelligence makes this question much more urgent because AI is beginning to alter the difficulty of cognitive work in much the same way that machines altered the difficulty of physical work. For centuries, better tools primarily amplified muscles. A crane could lift more than a person. A locomotive could pull more than a horse. A bulldozer could move more earth than hundreds of laborers. Computers then dramatically amplified our ability to calculate, store information, and retrieve it.

AI goes further. It can participate in activities that we have traditionally treated as evidence of human intellectual effort: writing, programming, analysis, research, design, and problem solving. That means that many outputs traditionally regarded as accomplishments will become much easier to produce.

An attractive illustration can already be produced in minutes. Increasingly, so can a competent computer program, a business analysis, a marketing plan, a research summary, or an article like this one. None of those things becomes worthless simply because AI helped create it. But the accomplishment represented by producing it has changed.

That pushes human value toward a different part of the process. If AI can turn an idea into polished prose in minutes, typing and revising the prose become less important. Having an idea worth expressing becomes more important. So do recognizing whether the AI has expressed it correctly, identifying what is missing, distinguishing an interesting answer from a banal one, and deciding what should be created in the first place. In other words, better tools do not necessarily lower the standard for human accomplishment. They may raise it.

What Would Shakespeare Do With AI?

There is a tempting historical thought experiment here. What would Shakespeare have produced if he had possessed a word processor? But that may already be too modest a question. What would Shakespeare have done with AI? We obviously cannot know. Perhaps he would simply have produced plays faster. Perhaps effortless revision would have changed his writing process. Perhaps instantaneous access to history, literature, and information would have expanded the range of subjects he explored. Perhaps he would have ignored most of the technology and continued doing the part at which he was unusually good.

But if we believe that Shakespeare's extraordinary ability was not his capacity to move a quill across paper but his ability to see things about human beings that other people did not see, then better tools might have amplified rather than diminished him. The tool could take over some of the mechanical work while leaving the scarce capability, the extraordinary human perception, untouched.

That is also one of the more optimistic ways to think about AI. The greatest human achievements have never resulted from tools alone. Tools multiply capabilities. They do not necessarily provide purpose, judgment, originality, courage, curiosity, or taste. A medieval master builder with a crane is still a master builder. Shakespeare with a word processor is still Shakespeare. And Shakespeare with AI would presumably still have to decide what was worth saying.

What Did You Do With the Tools You Had?

Thinking in terms of tool-adjusted accomplishment gives us a better way to admire the past without disparaging the present. The builders of Chartres did something astonishing with the tools they had. Shakespeare did something astonishing with the tools he had. Millions of women gave birth and raised children under conditions that would seem extraordinarily dangerous today. Scientists calculated by hand. Sailors crossed oceans without satellites. Engineers built bridges without computers. Artists worked for months or years on objects that could be destroyed by a fire in an afternoon.

Their limitations are part of their accomplishments. But the same standard has to apply to us. We possess tools that would appear almost magical to nearly everyone who has ever lived. That should not make us embarrassed about using them. Refusing to use a good tool is rarely an accomplishment in itself. Instead, better tools should make us more ambitious about what we expect to accomplish.

I did not spend two days laboriously writing this article. I supplied the central idea, the argument, examples, and direction, and AI performed much of what I once would have considered the work. I then judged the result and continued directing it until it expressed what I wanted to express. That is a fundamentally different production process from the one I would have used even a few years ago.

The important question is not whether using AI somehow makes the accomplishment illegitimate. The more interesting question is what I do with the time and capability the tool has given back to me. If something that once required two days now requires two hours, I have not necessarily accomplished the same thing more impressively. I have been given twenty-two additional hours.

Every generation inherits a different technological starting point, and it is unreasonable to judge people for failing to use tools that did not exist. We cannot ask the builders of Chartres why they did not build a hundred-story skyscraper, or Shakespeare why he did not produce twice as many plays with the assistance of a computer.

Future generations, however, will be entitled to ask something of us. We had computers, instantaneous global communication, access to much of accumulated human knowledge, machines capable of performing enormous amounts of physical labor, and eventually artificial intelligence capable of helping us think, write, design, analyze, and create. Given those tools, what did we accomplish?