# What remains human in a world of 'thinking' machines? #questions/what-remains-human · #ai · #philosophy/epistemology · #systems ^14590b This line of inquiry emerges from the accelerating capacity of artificial systems to not only do work, but to simulate or substitute aspects of cognition—pattern recognition, clustering, even “insight.” As automation expands from physical labor to mental modeling, what is left that is distinctly human? I'm particularly interested in the tension between the professional pressure (need?) to adapt to a new paradigm of [[semiotic capitalism]] and the very real existential threat of over-reliance on [[computational thinking]]. At the core, I'm wrestling with: - **Cognitive Sovereignty** — the ability to _own_, _shape_, and _understand_ your own thought processes, rather than outsource or conform them to machine logic. - **Epistemic Boundaries** — discerning _where human knowing ends and machine modeling begins_—and what gets lost when that line blurs. - **Design Ethics of Mind** — deciding what kinds of cognitive environments we want to build, inhabit, and defend—not just what we can automate, but what we _shouldn’t_. ## Working Language | Frame | Position | | ---------- | -------------------------------------------------------------------------------------------------------------------------------- | | Tension | Are machines helping us think better or making us think less? | | Hypothesis | That certain forms of thinking—discernment, care, narrative, ethics—are fundamentally embodied, relational, and non-automatable. | | Challenge | To design systems and practices that preserve human agency and understanding, even as AI grows more capable and opaque. | ^working-language ## Current Thinking ### Cognition and computational thinking The challenge is not just AI's capability, but the ***erosion of human habits of critical engagement.*** Technology that obscures, obfuscates, and ultimately controls. - On the progressive nature of [[computational thinking]]:![[New Dark Age (2018)#^6678f0]] - On the danger of outsourcing "[[thinking/concepts/reification]]": ![[reification and critique#^fa89e9]] ### Reification, risk, and efficiency There are *qualitative differences* in how humans think (meaning-making, ethics, embodiment), even if machines outperform in speed or scale. - "Human in the loop" must mean more than permission-granting; it may require re-centering human *discernment*, *care*, and *responsibility*: - Reverse centaurs a la [[Cory Doctorow]]![[What kind of bubble is AI (permalink) - Clips#^1ba9c8]] - On the tension between efficiency and essential "rightness" of discernment![[What kind of bubble is AI (permalink) - Clips#^108946]] - On the challenge of adapting, [[self-governing systems]]: ![[Rob Bagot’s Post on AI and Skill Premiums - Clips#^a31f1d]] ### Sensemaking and irrationality The value of thinking as a human act may be less about computation and more about attention, narrative, ethics, and caring. Machines may help us predict the world—but only humans can _make sense_ of it. And sensemaking includes doubt, error, feeling, and care. - **Sensemaking is not the same as information processing.** - It’s an _embodied_, _narrative_, and _ethical_ process. - It includes _ambiguity_, _uncertainty_, and even _contradiction_. - Machines excel at optimization; humans excel at meaning-making. - **Irrationality isn’t a flaw—it’s part of how humans navigate complexity.** - Doubt, intuition, and emotional salience aren’t bugs—they’re features in human cognition. - Systems that enforce “rational” coherence (like LLMs or computational models) risk flattening or pathologizing human ways of knowing. - Sometimes, the _right_ choice is not the _efficient_ one. - **In the pursuit of seamless systems, we may discard the very qualities that make cognition humane.** - If systems optimize away slowness, contradiction, or care, they risk alienating us from our own processes of understanding. - There’s danger in mistaking _legibility_ for _truth_, and _efficiency_ for _wisdom_. There’s value in *slowness*, *doubt*, and *perspective*—forms of cognition machines may not replicate, but we risk discarding in pursuit of efficiency.