Short-Circuiting of Long Circuits
A long circuit is the extended temporal process through which knowledge is built — years of reading, decades of practice, sustained engagement with difficulty through which savoir-faire accumulates. A short circuit is the direct production of the output the long circuit was supposed to produce, without going through the intervening process. Stiegler argued that the cultural industries of the twentieth century began short-circuiting long circuits of attention and desire — delivering pre-formed cultural products that replaced the longer processes of individual and collective meaning-making. AI extends short-circuiting into the long circuits of cognition itself: the production of philosophical prose, working code, competent analysis without the process through which such outputs used to be built.
In the AI Story
The distinction between long and short circuits gives pharmacological analysis its operational edge. It is not enough to observe that a tool produces output; the question is whether the output is produced through a circuit that simultaneously produces the practitioner. The long circuit of philosophical study produces both philosophical texts and philosophically formed minds. The short circuit of AI-generated philosophy produces philosophical texts while leaving the practitioner's philosophical formation untouched.
Segal's Deleuze fabrication is the canonical example. Claude produced an eloquent passage connecting Csikszentmihalyi's flow state to a concept attributed to Deleuze. The passage was beautiful, structurally elegant, rhetorically effective. The philosophical reference was wrong. The short circuit had generated the form of insight without the process of understanding that would have caught the error. Segal's verification was an act of long-circuit discipline performed against the short-circuit's seductive output.
The generational stakes of short-circuiting emerge from the relationship between the current transitional generation and those who follow. Today's senior practitioners possess long-circuit-built understanding that enables them to evaluate short-circuit output. Future practitioners, raised in environments where short circuits are the default path to output, will lack the evaluation capacity because they will not have undergone the circuits that build it.
The response is not to refuse short circuits but to protect and institutionalize selected long circuits. Which circuits must be preserved is itself a pharmacological judgment — one that cannot be codified but must be developed through the specific practice of attending to what is being built, and not being built, through the use of the tool.
Origin
Stiegler developed the concept across Symbolic Misery and later works, drawing on the psychoanalytic tradition's concern with the circuits of desire and identification.
The analysis has been extended to generative AI by Anne Alombert and other Stiegler heirs in work published from 2022 onward.
Key Ideas
Output without process. Short-circuiting delivers the product while eliminating the process — a gain in efficiency and a loss of formative development.
Long circuits produce practitioners, not just products. The value of the long circuit is not merely what it outputs but who it forms.
The Deleuze test. Short-circuit output may be indistinguishable from long-circuit output at the level of form while being hollow at the level of substance.
Generational transmission is at stake. Those who went through long circuits can evaluate short-circuit output; those who haven't, cannot.
Debates & Critiques
Defenders of AI-augmented practice argue that new long circuits emerge around evaluation, orchestration, and judgment. Stieglerians accept this in principle but ask whether the emerging circuits are actually being built or merely imagined, and whether they depend on prior long circuits that the tools themselves render economically unnecessary.
Appears in the Orange Pill Cycle
Further reading
- Bernard Stiegler, Symbolic Misery, Volume 2 (2006)
- Bernard Stiegler, Taking Care of Youth and the Generations (2008)
- Anne Alombert, "Reticulated Artificial Intelligence" (2024)