March 3, 2026

On Stupidity

Computers are not smarter than humans. They are just able to collect all the knowledge lore that humans have created, better than any expert, because they have access to more sources. They also are equipped to generate text automatically containing a high level synthesis of information they have gathered.

So far, search engines returned lists of links to sources. Now, the AI engines swallow the content of those sources, find what is relevant to the prompt, and return a more precise and well-written answer.

Creating these syntheses of accumulated knowledge also create knowledge. Many scientific works, many dissertations, are compilations of existing works. Therefore AI can facilitate these types of work and replace the people who are doing them.

AI only creates things based on algorithms, rules, and reasonings that have been built-in. Some AI companies now pretend that AI is able to generate their own rules, but this is only a circular way to perform the same thing. It means that rules can be created inside other rules, and can lead to complex results, that may produce an interesting output.

Sometimes, sources contain contradictory assessments. A person can express a statement on something that may be entirely different, even contradictory with another person's statement on the same subject. A sophisticated AI agent may report both and utters the fact that diverse opinions exist over that subject.

AI is based on statistical analysis. The more the merrier. This is where biases can be introduced. A number of chatbots and troll farms exist for the sole purpose of multiplying a message to falsify the statistics. There is no reliable way for an AI engine to evaluate whether a message has been generated by a bot or a by a genuine human being. In other words, it is impossible to know whether this is a genuine original writing, or not. We, as humans have values, beliefs, and we can express doubt on the veracity of certain assertions, because they are at odds with what we know.

The end of the quantitative era?

The availability of big data and the digitization of everything is culminating with AI. Not only letters are numbers (character codes) but text is a vector database, containing computed correlations. The reporting is made by numbers. Corporations love numbers, graphs, curves, trends. Money is just a number. Make more money is adding numbers. The only question about money is how much money one makes, not how one makes money. In a capitalist society, that revolves around money, everying revolves around numbers. And guess what, computers are good at handling numbers, whether they express actual numbers, or things that have been encoded into numbers in computers, like text.

Corporations have transformed many jobs into number crunching processes. The numerous apps and software that are available force people to imprison any activity into computers, therefore digitizing everything. AI can do that better than humans, as they are built-in as number processing agents.

Back to imagination, creativity, connectivity (interdisciplinarity), deep thinking.

Not everything can be digitized. It is possible to digitize the form in which things are expressed, whether as text, video or audio recording, image. The means of communications are digitized, almost entirely. Even printed text is the result of a digitized process. But communications is not everything that exists. A thought, an understanding, a misunderstanding, an agreement, a disagreement, a conflict, satisfaction, dissatisfaction, frustration, contentment, happiness, sadness, success, failure result from communication and exchange. They are what constitute the core of our daily life, both professional and personal. Emojis are ways to encode emotions and make them prone to digitization. It enables some companies to produce sentiment analysis that is used to derive trends of satisfaction level, etc. But they are unable to capture the nuances and the reasons why people feel a certain way.