The mechanism is attentional. Attention is a finite biological resource. When semiotic input exceeds attentional capacity, the organism has three options: process some signals at the cost of ignoring others (selective attention), process all signals shallowly (diffused attention), or withdraw from semiotic engagement entirely (defensive disengagement). The first produces comprehension but incomplete coverage. The second produces coverage but without comprehension. The third produces relief but at the cost of informational impoverishment. All three are forms of adaptation to overload, and none of them constitutes genuine processing at the depth that meaningful engagement requires.
The AI-augmented builder operates at or beyond the overload threshold much of the time. The tool generates output — code, text, design options — at a rate that demands continuous evaluation. Each piece of output must be assessed for quality, relevance, and alignment with intention. Each assessment generates new direction, which generates new output, which requires new assessment. The cycle is rapid, continuous, and cognitively demanding. The tool consolidates into the attentional field of one person the semiotic output that would previously have been distributed across a team operating over weeks.
The consequences extend beyond the individual builder. Every AI-augmented builder adds to the aggregate semiotic output flowing into the shared semiosphere. More code, more text, more content is produced per unit of time than at any point in human history. The attention that might process this output, however, remains biologically constrained. The result is semiotic overload at civilizational scale: a world in which more signs compete for less available attention, producing the specific pathology Berardi calls semiotic inflation — more content carrying less meaning per unit of content.
Berardi's diagnosis implies that individual solutions to semiotic overload are insufficient. The builder who protects her own attention is still swimming in an ocean that grows deeper and more turbulent. Individual discipline can reduce exposure but cannot change the overall level of semiosis. The problem, like climate change or antibiotic resistance, is collective — produced by the aggregate of individually rational choices and requiring collective institutional response to address.
The concept appears throughout Berardi's work from the 2000s onward, with particular development in The Uprising (2012), And: Phenomenology of the End (2015), and Breathing (2018). Related concepts appear in the work of Bernard Stiegler, Byung-Chul Han, and contemporary attention researchers.
The framework has acquired specific new applications in the AI moment, where machine generation represents a categorical increase in semiotic production capacity.
Threshold phenomenon. Below capacity, more input means more understanding; above capacity, more means less.
Three adaptive failures. Selective attention, diffused attention, defensive disengagement — all suboptimal responses to overload.
AI as overload accelerant. Machine generation categorically increases production volume without increasing processing capacity.
Semiotic inflation. More signs carrying less meaning per sign — the qualitative consequence of quantitative increase.
Collective, not individual, problem. Individual discipline cannot address an ambient condition produced by aggregate choices.
Whether semiotic overload is a genuinely new phenomenon or a recurrent condition (every information technology has produced similar complaints — Gutenberg, the newspaper, television) is empirically contested. Berardi acknowledges the historical pattern but insists that the current moment represents a qualitative break because machine generation removes the human-capacity constraint that limited all previous media.