Your brain does not think like a computer running serial commands. It thinks like ten billion neurons firing simultaneously, each one strengthening or weakening its connections based on what works. That's a neural network. And here's what's radical: your mind solves problems faster through this parallel messiness than any logical sequence could achieve. The traveling salesman problem—a mathematical puzzle that would take a serial computer millennia to solve—a neural net solves in minutes. Not through better logic. Through ten thousand pathways competing, converging, eliminating weak paths and amplifying successful ones.
This is not a metaphor for how the brain works. This is how the brain actually works.
A neural network is a system where:
Nodes (neurons, processing units) exist in layers. Sensory input arrives at the bottom layer. Decision output emerges from the top layer. Hidden layers in between do the actual computational work—and here's the key: they're doing work in parallel, not one step at a time.
Connections (synapses) link nodes. Each connection has a weight—a strength. When a neuron fires, that signal travels along weighted connections to the next layer. Strong connections transmit signal effectively. Weak connections barely transmit.
Learning happens through weight adjustment. When an output is right, the weights on successful pathways strengthen (reinforcement). When an output is wrong, unsuccessful pathways weaken (extinction). Over time, the network becomes good at the problem it's trained on.
This is Hebbian learning in its purest form: neurons that fire together wire together.
The neurobiological substrate: synaptic efficacy increases with repeated stimulation. AMPA and NMDA receptors proliferate at active synapses. Unused synapses prune. The brain literally rewires itself based on what it practices.
A serial computer tries one path at a time. A → B → C → solution. If the path is wrong, it backtracks and tries again. This is fast for simple problems. It is catastrophically slow for complex problems where the number of possible paths explodes exponentially.
A neural network tries ten thousand paths simultaneously. Most are weak. Some are stronger. A few are very strong. The network "settles" into the strongest configuration. This happens in milliseconds through parallel activation—far faster than any serial algorithm could achieve.
Example from Bloom: The traveling salesman problem. A serial approach: calculate distances for every possible route (factorial explosion: for 20 cities, 20! = 2.4 × 10^18 possibilities). A neural net: initialize all cities as nodes, all connection weights as random, then let parallel activation find the shortest route through gradient descent. Solution in minutes.
The brain does this constantly. Visual recognition: instead of checking "is this a face?" through serial feature-testing, the visual cortex activates thousands of filters in parallel—edges, curves, textures, spatial relationships—all competing and converging on "face" or "not face" in 100 milliseconds.
The brain is not hard-wired at birth. It's plastic—responsive to experience. Connections strengthen with use. Unused connections prune. This happens throughout life, though the window for rapid plasticity (critical periods) narrows with age.
What you practice, you get better at. Not through willpower or discipline. Through neurochemical rewiring. Practice increases BDNF (brain-derived neurotrophic factor), which supports neuron survival and growth. Practice increases dendritic spine density at active synapses. Practice literally grows the neural tissue devoted to that skill.
Inversely: disuse causes atrophy. Stroke victims losing use of one arm see neural tissue in the motor cortex devoted to that arm shrink. Rehabilitation reverses this—forcing use of the arm rebuilds motor representation.
This has radical implications: You are not fixed. Your neural network is optimized for what you've been doing. Change what you do, and your brain rewires.
Bloom emphasizes: male and female brains show structural differences in connectivity patterns. These differences are not absolute—there's vast overlap—but they're statistically significant.
Male brains show stronger within-hemisphere connectivity. Pathways connect left-to-right within one hemisphere efficiently. This architecture optimizes for focused, single-track processing. Advantage: targeting, throwing accuracy, sustained attention on one problem. The network consolidates around one solution path.
Female brains show stronger between-hemisphere connectivity. Pathways bridge left and right hemispheres frequently. This architecture optimizes for parallel processing across multiple domains. Advantage: integration, multi-tasking, perspective-shifting, seeing multiple frameworks simultaneously. The network keeps many pathways active.
This is not a moral or capability difference. It's a computational architecture difference. Males are optimized as specialists; females as generalists. In environments demanding single-track execution (combat, throwing-based hunting), male architecture wins. In environments demanding integration (child-rearing, coalition management), female architecture wins.
Bloom's point: evolutionary selection pressure shaped these architectures. Expendability made males risk-takers (focus narrowly on conquest). Investment made females alliance-builders (integrate across social domains).
How to recognize which network you're operating:
Identify the problem type. Are you solving something that requires focus (one best answer, precision required) or integration (multiple valid approaches, synthesis required)?
Notice your activation pattern. When you think about a problem, do you narrow to one solution path and pursue it relentlessly? Or do you hold multiple frameworks simultaneously and weave between them?
Check your history. Which tasks have you practiced extensively? Those pathways are strengthened. Which have you neglected? Those have pruned. Your network is optimized for your history.
Test under stress. When threatened or time-pressed, which mode dominates? Focused/narrow or integrative/broad? Stress reveals your default architecture.
How to deliberately rewire your network:
Practice the opposite mode intentionally. If you naturally focus narrowly, force yourself to hold multiple perspectives simultaneously. If you naturally integrate broadly, force yourself to single-track on one problem. Neuroplasticity requires deliberate practice against your default.
Use spacing and variation. Don't practice the same thing the same way repeatedly—that creates brittle expertise. Vary the context, timing, and framing. This forces the network to generalize rather than memorize.
Increase difficulty gradually. The network strengthens at the edge of current capability. Too easy = no rewiring. Too hard = failure and frustration. Slightly beyond comfortable = maximal neuroplasticity.
Leverage the critical period windows. Childhood and early adulthood show the fastest neural rewiring. After 25, change requires more deliberate effort. But it's always possible—the brain remains plastic throughout life.
Use reward and punishment intentionally. Dopamine strengthens pathways associated with reward. Punishment (negative feedback) weakens failed pathways. Direct your feedback system toward the changes you want to make.
Evidence:
Tensions:
Open questions:
Bloom's neural network framework parallels connectionist cognitive science (Rumelhart, McClelland, Parallel Distributed Processing), but extends it from explaining behavior to explaining civilizational dynamics. Connectionist theory in psychology focuses on individual-level learning and representation. Bloom applies the same logic to group-level decision-making and cultural evolution—societies as macro neural networks solving problems through parallel activation and pathway strengthening.
The tension reveals: Neural networks operate at every scale of organization. Individual neurons forming networks form brains forming minds forming societies. The same computational principles apply across levels. Understanding how a single synapse strengthens explains how a civilization mobilizes.
Neural Networks as Connectionist Processing explains how this architecture functions as a tactical system. Where psychology explains how the brain physically implements connectionist processing, behavioral-mechanics explains how to operate that processing to achieve strategic outcomes.
Psychology shows: parallel activation is faster than serial logic. Behavioral-mechanics shows: this principle scales to group-level operations. A business organized as a connectionist network (distributed decision-making, parallel experimentation, rapid feedback) outpaces a serial hierarchy (centralized decisions, sequential execution) in complex environments. The principle is identical—leverage parallelism and let gradient descent find the solution.
The tension: individual brains can't always exploit their parallel capacity (stress narrows attention, amygdala hijacking forces serial threat-response). Groups can—if they have the right institutional architecture. A single person cannot hold 10,000 solution paths simultaneously in consciousness. A distributed organization can. Network effects enable what individual wetware cannot.
Distributed Decision-Making and Civilizational Resilience documents that civilizations with distributed decision-making (parallel authority centers, multiple information flows, diverse problem-solving approaches) adapt better to novel crises than centralized hierarchies.
Where psychology explains the mechanism (parallel processing solves complex problems faster), history explains the evidence (distributed societies survived crises that centralized societies collapsed under). Venice's merchant republic network outadapted Ottoman centralized hierarchy when trade routes shifted. Swiss cantons outadapted kingdoms with unified command.
The handshake: A connectionist society is one organized to leverage parallel processing at scale. History shows which societies unconsciously stumbled into connectionist organization (and survived). Psychology shows why it works. Behavioral-mechanics shows how to deliberately design for it.
Worldviews as Problem-Solving Nets applies the connectionist model to belief systems. A worldview is a neural network trained on specific data—a person's experience, culture, media environment. That network has certain pathways strengthened (beliefs that paid off) and others weakened (beliefs that failed).
Where psychology explains how individual beliefs form (through network training), behavioral-mechanics explains how belief systems operate tactically (as lenses that mobilize action), and cross-domain synthesis explains why certain worldviews persist despite contradicting reality (they're optimized for the problem-space they were trained on, not for truth itself).
A person trained in scarcity experiences prosperity as dangerous and clings to conservation. A person trained in abundance experiences scarcity with curiosity. Neither is seeing reality clearly—both are running their trained networks. Understanding belief as network activation, not as conscious choice, explains why persuasion is so difficult and why people don't change beliefs when presented with contradictory evidence.
Your neural network is optimized for your history, not for your future. Every belief you hold, every skill you possess, every habitual thought pattern—these are pathways strengthened by repetition. They worked in your past. But the world changed. Your network is still running yesterday's solution.
This explains why people fail in novel situations despite being intelligent. Raw IQ is not the constraint. Network mismatch is. You have a brain optimized for the environment you came from, not the environment you're in now.
The radical implication: you must deliberately rewire. This is not about trying harder or believing differently. It's about practicing differently, repeatedly, until new pathways strengthen and old ones prune. This takes months to years, depending on the magnitude of change. There is no shortcut.
What problem was your neural network trained to solve? Look at your life history—the problems you had to solve, the environments you grew up in. Your brain is expert at those problems. What new problems are you facing that your history didn't prepare you for?
Which of your "intuitions" are actually trained responses, not genuine wisdom? The things you know without thinking—these are strongly weighted pathways. Are they correct for your current situation, or are they optimized for a past context that no longer applies?
What would you need to practice repeatedly to rewire your default responses? If you wanted to shift from narrow focus to broad integration (or vice versa), what specific practices would strengthen the new pathways? Not willpower—practices. Deliberate, varied, progressively harder practice.