Your Text is Worth Nothing

We are used to a world where text correlates more or less with the amount of information and effort. If a colleague handed you a 50-page document, you knew they had spent weeks researching, synthesizing, and formatting. The length of the text itself was a reliable proxy for diligence.

That world is dead.

Artificial intelligence has inflated the volume of text in the world by such a staggering margin that we are forced to fundamentally reevaluate how we communicate. We have decoupled word count from human effort.

Recently, we were tasked with evaluating a series of project proposals. We received the usual massive, dense presentation decks from well-known consultancies. In an attempt to streamline the process, our procurement department ran these decks through an AI tool to generate evaluations, which were then emailed to our review committee.

The result was a complete waste of everyone's time. The AI-generated evaluations were painfully long and overwhelmingly generic. They sounded entirely professional, yet they said absolutely nothing of substance. Instead of accelerating our decision-making, we were forced to dig through synthetic text.

Funnily enough this illustrates one of the core concepts of Claude Shannon’s Information Theory from the 1940s, which states that the actual information content of a message is inherently tied to its unpredictability. Because AI-generated text is inherently statistically probable, one could argue it therefore is the opposite of surprising. Another interesting related read is Noisy-channel coding theorem although they are both very math-heavy.

So what I want to say: We have to pivot back. Less is more.

The sole purpose of text should be to convey information, not to waste the reader's time. Wrapping a simple truth in paragraphs of bloated, artificially generated fluff is no longer a sign of thoroughness, it is a lack of respect for the recipient's time.