Caring in AI times

In the last year, several customers have approached us for help with projects that were written heavily by AI. Some of them were completely broken, some were great. The difference was in caring. In the good ones, you felt very quickly that someone had cared. In the bad ones, there was no care.

Caring is a distinctive trait of living beings, and especially of humans. Computers don’t give a damn. I asked GPT if selling my turds would be a good business idea and it congratulated me. The idea was a wonderful combination of art and entrepreneurship. Then I replied that the idea was stupid and it agreed, and said that I am right again. Computers simply don’t give a damn about anything. Humans do.

The Greeks had a word for this: Eulabeia. It is a kind of relationship that we have with things, that allows them to flourish in their full glory. This concept was so crucial for Greeks that it made its way into our modern languages as “religion” (with major mutations of meaning). You can think of Eulabeia as acting as if the gods were looking at you while doing things. In Homer, Eulabeia is the capacity to listen and the exercise of this talent. Its opposite is Akedeia, to be careless, to be negligent, a mixture of pride and indifference that leads to disaster, to leak your API keys, to model your state in a way that makes impossible further progress, or to paste issues into the LLM and mark them as complete until the project is supposedly “90% done”.

Now, to care you must know. Knowledge is not unique to humans, but it is a requirement for care. You cannot care properly for a garden without knowing about plants. No matter how mechanized your tools are, if you don’t have the knowledge, caring will only mean crying when things go wrong. That is a wrong kind of caring.

Think about Mr. Miyagi and his bonsai trees. He exercised care, he let them flourish in all its glory, but he also knew a great deal about bonsais. The Karate Kid is a movie about Eulabeia and Akedeia (Cobra Kai).

To care doesn’t mean obsessing over details, or refusing to ship until everything is perfect. It doesn’t mean to be slow or to do things by hand. It means being attentive to the needs of the particular project, as a whole, that you are working on. It means listening, and also knowing where to look, where precision matters and where a quick approximation is ok. Being slow when the project requires to you to be fast is not paying attention, not caring. Different things have different needs. Some projects need to be put on the market fast, without too much polishing. Others will have different needs.

Caring, along with the knowledge that is needed to exercise it, is what makes all the difference, as it always has.