Dear friends,

In my last piece, I wrote about Apollo Go’s robotaxis in Wuhan and framed their technical failures as a strategic opening for workers. I want to stay with this a little longer, because that technological limit is telling us something important about what “artificial intelligence” actually is, and what workers actually do.

A book I’ve been reading, Ways of Being by James Bridle, opens with a provocation: intelligence is not what we think it is. We’ve built AI on a narrow idea of what minds do: calculating, optimizing, processing symbols. But Bridle argues this image was always incomplete. Intelligence doesn’t happen inside a brain; it happens between a body and a world.

I want to use that as a jumping point for rethinking solidarity, not as feelings or a posture, but as a form of social infrastructure.

Asian Labor Futures is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

The Collective Knowledge that Machines (Fail to) Replicate

Driving is one of the most complex tasks humans perform, yet it feels effortless to most of us. It’s nearly impossible to replicate in a machine. There’s a technical name for this: the Moravec paradox. What is easiest for us is hardest for machines, and vice versa.

Driving is a relational activity: reading the slight hesitation of a motorbike about to cut across your lane, sensing the rhythm of a street that slows at certain hours, knowing from experience that the puddle on this corner is deeper than it looks in the rainy season. This knowledge lives in the relationship between a body and a place, built over years of moving through it. It cannot be cleanly extracted and fed into a training dataset.

Yet workers are mislabeled “unskilled,” a classification with a long history, from the factory to the platform economy, of management taking exactly this kind of knowledge and placing it up the chain of command. What remains with the workers gets reclassified as simple execution requiring no particular intelligence. The robotaxi is the same story, albeit unfinished, played out on the street.

And where the machine still falls short, tech companies now recruit people who desperately need income to close the gap, training algorithms in real-life conditions, on poverty wages, so the machine can eventually replace them too.

What Workers Build Within the Entanglement

I’ve often felt a disconnect between the reality I see on the ground and calls for “full automation” or “post-work” futures coming from some corners of the Western left. These are important political horizons, but they can obscure the immediate question facing workers who are already inside the system and cannot simply opt out.

But the delivery riders I work with don’t have that option. By economic necessity, they are inside the relationship with the machine—the app, the algorithm, the gamification system. The question for them is not whether to engage with technology, but what to build within that entanglement.

Workers are not jut redefining themselves in the process. They are building something inside these platforms that the algorithm never intended and cannot control. This isn’t just about collective bargaining; it’s about workers coming to know one another, recognizing that the social world the algorithm depends on is actually ours.

The routing algorithms that guide platform logistics are built on the accumulated navigational intelligence of thousands of workers. This is collective labor in its purest form. And it raises a clear political demand: what is collectively produced must be collectively owned. The moment we begin to claim that knowledge together, we reclaim our “world-making rights.”

What This Means for Solidarity Building

Building collective power from within is difficult. Not just strategically, but subjectively. Platforms are designed to deplete the very capacities that organizing depends on: the desire to connect, the capacity for trust.

What we can do as movement builders is support workers' initiatives that start small and molecular: a mutual aid project or a savings circle. These are the first bricks of a solidarity infrastructure. This is similar to what scholars Roseann Liu and Savannah Shange call thick solidarity: building from workers' actual, embodied positions, across real differences in their conditions, led by those most directly inside the machine.

Unlike thin digital connections designed to keep us isolated, thick solidarity creates recurring obligations, shared risk, and belonging that doesn't disappear when accounts are deactivated.

One framework I’ve found generative comes from the US-based Building Movement Project, which uses scaffolding as a central metaphor. Scaffolding is the support system that makes it possible to build something larger than ourselves. Not the final structure, but what protects our collective energy while we’re building it. It’s what moves us from constant reaction toward shared governance and lasting power.

Platform worker movements desperately need this kind of bridge between current capacity and actual need. Between what capital permits and what justice requires. And we should be honest with ourselves: what we’re building in the meantime may not look pretty, but that’s ok.

Until next time,

Kriangsak (Kiang)

Get full access to Asian Labor Futures at asianlaborfutures.substack.com/subscribe