Process and criteria

Making things requires you to think about every single aspect. The methods and the materials, purpose, goal, your practice. And then maybe you will also think about how the tools should work, or you start to wonder about how it was done before. When and why is it good enough? This is curiosity in its most essential form.

As you get involved with the process, so does your understanding of its possibilities and limits. Inevitably, while you work, plans form on how the work could be better. And you learn to accept if you discover some of it is as good as it gets. You let these thoughts inform the way you go about the next time you engage with this task. The making gets better, the work gets more interesting and more importantly your understanding of the work and everything connected to it grows. And you will have learned why this current iteration will do for now. Your criteria are embedded in, and represented by the process.

Hey now these are your friends.

On the other hand

Generative AI allows you to cause things that you otherwise would not have been able to. You feel there is a connection between writing the prompt and observing the outcome. Look, I made this happen! You assume this is the result of the unique way in which you tossed the dice. Post hoc ergo fucking propter hoc.

But as there was no process, your mind was not challenged to think about steps and ponder alternatives. You did not examine possibilities, you did not negotiate with the limits of tools and task.
Maybe the generated outcome was very straightforward. Maybe it overstepped a boundary somewhere. Maybe it was mediocre or just plain wrong. You do not really know. You did not learn anything from making it because there was no making. You did not develop any criteria.

Not knowing how to judge the outcome is not an uncomfortable side effect of generative AI use, it is the main attraction. Not learning (about how something could be done) is exactly where you save the most time. By reaching for generative tools, you implicitly say that you do not wish to learn from the process.

Teach

The services offering generative AI already know that everyone who signs up is willing, no, expecting to be fooled. It is easy to satisfy the hasty, uninterested mind and so the thinnest, most mediocre generated thing will do.

There is no reason to assume generative AI services will remain free or cheap or even accessible. The industry runs at an unprecedented loss, subsidised by speculation. There are certainly applications in which machine learning can do useful things, fold a protein, cure cancer, sure please.

But the loss of critical thinking skills is not an abstract thing we will lament in an abstract way..
If education fails to transfer these skills, it is near impossible to reconstruct them later by looking at artefacts.

As generative AI invites you to step outside the boundaries of what you know, the results will look believable. In this context, consider Knoll’s Law of Media Accuracy: “Everything you read in the newspapers is absolutely true except for that rare story of which you happen to have firsthand knowledge.” [wiktionary]

Not mentioned here, but no less important: the AI industry’s energy and water consumption. The appropriation of intellectual property. Billionaires as failures of policy. The active support of fascist politics.

Typo.social post, 30 September, 2026.