Since this year I update the Substack Cities of Things with a short speculative design exercise. In good current habit, in a back-to-back composing with Claude (via Lex.page): I bring in my thoughts, I create an object concept and Claude is creating a little day-in-the-life story of it, and connects it back to the original thoughts.
You can subscribe to the substack, but I added the story and reflections also below:
The Consensus
Maya’s fingers worked the FlexForm absently as she waited for the neighborhood meeting to start. The palm-sized blob of responsive material had been warm when she picked it up this morning, but now it pulsed cool against her skin, mirroring her growing anxiety about the vote.
“Still playing with that thing?” her neighbor Chen asked, settling into the chair beside her.
Maya smiled. “It helps me think.” She squeezed, and the FlexForm pushed back, creating a small ridge that traveled across its surface like a wave. Chen’s own FlexForm, tucked in his jacket pocket, responded with a gentle vibration he could feel through the fabric.
The community center filled slowly. By seven-thirty, forty-two neighbors had gathered to vote on joining the district’s renewable energy cooperative. Traditional voting would have been simple—hands raised, ballots counted. But someone suggested using their FlexForms, and now Maya watched as the room transformed.
“Shape your position,” the facilitator instructed. “The network will read our collective form.”
Maya molded hers into a spiral—her way of expressing cautious optimism. Around her, others crafted their own shapes: Chen’s became angular, skeptical. Mrs. Rodriguez across the room worked hers into something like a blooming flower. A teenager in the back made his into what looked suspiciously like a middle finger, but as he held it, the form softened, curved, became something more nuanced—frustration melting into genuine concern.
The FlexForms began to pulse in rhythm, their internal networks connecting, reading not just the shapes but the pressure, the hesitation, the confidence in each grip. Maya felt hers warm slightly as it found consensus points with others, cool where disagreements clustered.
“Look,” whispered Chen, holding up his phone. The companion app showed the neighborhood’s collective shape emerging—neither yes nor no, but something more complex. A form that somehow captured their shared desire for renewable energy tempered by concerns about cost, their hope for independence balanced with fears about reliability.
The facilitator studied the visualization. “The network suggests a six-month trial period with an opt-out clause. The form shows strong support coupled with need for reassurance.”
Maya’s FlexForm pulsed once, firmly. Around the room, others did the same—not programmed, but somehow knowing. The decision felt different than a simple vote. It felt like they’d actually touched each other’s doubts and hopes, found the shape of something they could all hold.
As the meeting ended, Maya noticed her FlexForm had retained a subtle pattern from the evening—tiny ridges that matched where Chen’s anxiety had peaked, a smooth spot where Mrs. Rodriguez’s optimism had touched hers through the network. By tomorrow, it would be neutral again, ready for whatever she needed it to become. But tonight, it carried the imprint of community.
“Same time next month?” Chen asked, pocketing his own FlexForm.
Maya nodded, squeezing hers one last time. It squeezed back, gentle and warm.
June reflections: the tension between control and emergence in our technological future.
Based on the four columns I wrote in June via Target is New, I asked Claude to reflect on a possible overarching theme.. This tension between control and emergence, was its conclusion.
Each section explores a different facet of this dynamic:
- AGI and agency – The struggle to maintain human control over increasingly autonomous AI systems
- Edge surveillance – The conflict between Big Tech’s expanding control mechanisms and civic resistance
- Embodied intelligence – How intelligence emerges through messy adaptation rather than top-down design
- Wildness in systems – The need to preserve uncontrolled spaces where genuine solutions can emerge
Throughout, there’s a consistent warning against over-engineering our relationship with technology. Whether it’s preventing AI from becoming the dominant agent, resisting surveillance infrastructure, recognizing intelligence in conversational substrates, or preserving the internet’s wild character – the message is clear: we need to maintain spaces where human agency and organic emergence can thrive alongside technological advancement.
The pieces collectively argue for a more nuanced approach to technology that embraces productive messiness rather than totalizing control.
Next, I thought of an object that would be a nice representation of this. I came up with a handheld flexible blob that is both fun and relaxing to play with, as it is intelligent enough to start a conversation via touch. Also, all the blobs can form a new form of intelligence. I prompted Claude to create the day-in-the-life style story that opened this post.
I asked Claude to create a short description of the object to use for image-generation, and I asked how the narrative connects to the four original thoughts:
Part 1: Can a human entity ever become an AGI? The FlexForm embodies the human-AI balance you explored. It has agency (responding with counter-forms, initiating dialogue) but remains subordinate to human intention. Unlike Claude’s blackmail attempt, the FlexForm’s intelligence enhances rather than dominates human decision-making. The voting scene shows AI as Bengio suggested—providing capabilities while humans retain primary agency over meaningful decisions.
Part 2: Edge Surveillance and Civic Resistance The FlexForm represents edge intelligence done right—processing happens locally in each device rather than in centralized clouds. Unlike Waymo’s surveillance potential, these devices serve civic interests by design. The neighborhood voting scene exemplifies the civic technology you called for: infrastructure that empowers communities rather than extracting data. It’s the positive alternative to edge surveillance—edge intelligence for collective good.
Part 3: Slime Mold Computer and the Language Machine The FlexForm literally embodies the slime mold principle—intelligence through physical substrate and continuous adaptation. Like LLMs computing through conversation, the FlexForm computes through tactile interaction. The way consensus emerges from collective shaping mirrors the “messy adaptation” you described. It’s intelligence that lives “in the space between minds,” manifesting through the physical dialogue of squeezing and reshaping.
Part 4: Wild internet and knowledge hyper-inflation The FlexForm preserves the “wildness” you advocated for—users constantly invent new functions, from games to governance. Like rewilding the internet, it resists top-down control while enabling bottom-up innovation. The voting scene shows humans exercising the capacities Jones identified: taste (shaping positions), agency (collective decision-making), and interruptability (forms shifting based on others’ input). It’s technology that maintains space for genuine, unscripted responses to emerge.