Why Socrates and Plato wouldn't like AI tools

Although the initial hype around AI tools has calmed down, we are still witnessing the emergence of new AI-based tools every week. Contrary to a lot of influencers, it’s not really first time in history where we’re in tool revolution like this. Let me show you 3 historical references that may help you see all of the current buzz from different perspective.


Socrates on writing

Let’s roll back the clock to ~370 BC. Plato, Socrates student, writes a dialogue titled Phaedrus. As in other works from that years the subject matter is quite broad. Yet we’ll look closely on the section 274 to 275. In this dialogue, Socrates discusses the invention of writing. Let’s see a few quotes:

Writing, when it is written down, is no longer an aid to memory; it becomes something in itself, an external substance that is not part of the mind.

For this invention will produce forgetfulness in the minds of those who learn to use it. They will not practice using their memories because they will rely on writing, which is an external source.

They will appear to be wise, but they will not actually possess true wisdom. They will have a form of knowledge that lacks understanding.

Does this sound familiar? Replace 'writing' with 'AI/ChatGPT/Copilot,' and you have a modern debate. You may say that they lived in different times. Whole existence of philosophers was based on them walking around and telling wise words. Writing them down seemed endangering to their importance in society. But was it really? Philosophy did not stop after writing became common. And we can learn from that story only because eventually Plato decided to write it down eventually…

Bessemer on steel

Jumping now to Industrial Revolution. Good old times of children in factories, 16h work days and massive pollution. It was also transition from Iron Age to Steel Age. That was assisted by The Bessemer process, developed by Henry Bessemer in the 1850s, was a revolutionary method for producing steel. It decreased the cost from £40 per long ton to £6–7 per long ton, along with greatly increasing the scale and speed of production.

Overview of the Second Industrial Revolution

But how was his discovery received at first? It was generally criticized everywhere, mainly for its quality (The Engineer, 1857) and reliability (The Times, 1856). Later, most of these issues were found to be caused by misuse by certain iron masters. They tried to make steel the same way they made iron before or were negative about the idea from the start. Good old habits + quality concerns = AI Skepticism 101.

Jevon on coal

AI steals jobs vs AI will make us more effective. Which camp are you in currently? Still being in the Industrial Revolution times, the English economist William Stanley Jevons was looking into improvements happening around coal stoves. He observed that technological improvements that increased the efficiency of coal use didn’t decrease consumption of coal. Actually other way around - it increased general consumption across wide range of industries. Same tendencies have been observed elsewhere:

  • Increasing the yield of a crop, for a given area does not reduce the area required to achieve the same total yield - it makes it more profitable to grow clops so area increases

  • More fuel efficient cars do not decrease fuel use. They make it cheaper to ride, so more people choose driving.

  • Invention of the incandescent bulb made lighting much more energy-efficient. This however increased overall energy use as it was more convenient to light up places that previously were kept in dark.

So my point for AI tools would be:

  • The adoption of AI tools will improve workers' efficiency. However, this will not reduce the need for workers. Instead, it may become financially viable to undertake projects, positions, and initiatives that were previously unaffordable.

Jevon's Paradox - Sketchplanations

Bonus: StackOverflow in AI era

Not trying to prove anything here, but for a really long time we had jokes about “programmers who just copy code from StackOverflow”. Since 2021 that portal I constantly loosing traffic, recently is rapid pace! Do you know what happened as well in 2021? Copilot Beta was launched. And at the end of 2022 - ChatGPT.

Will we ever see again rise of SO popularity?

Total Questions on StackOverflow by year

https://x.com/v_lugovsky/status/1746275445228654728

0
Subscribe to my newsletter

Read articles from Krzysztof Przekwas directly inside your inbox. Subscribe to the newsletter, and don't miss out.

Written by

Krzysztof Przekwas
Krzysztof Przekwas