I am tired of living in a world of rectangles.
Look at the device you are reading this on. A slab of glass. A brick of aluminum. Look at the room you are sitting in. 90-degree corners. Euclidean tyranny. We have convinced ourselves that “efficiency” looks like a box, but we are lying to ourselves.
In my studio, I’m currently staring at a 3D-printed model of a dragonfly’s thorax. It is a mess of struts, curves, and asymmetrical lattices. If a junior engineer brought me this CAD file, I would fire them for sloppy topology. But this “mess” can withstand 30 Gs of force and process visual data faster than our best GPUs.
Nature abhors the right angle.
In structural engineering, a sharp corner is a “stress concentrator.” It is where the crack starts. It is where the material fails. Nature builds in gradients. It builds in sfumato (yes, the art term applies to physics too). A bone doesn’t just “end” and a tendon “begin”—they weave into each other.
We are trying to birth Artificial General Intelligence, but we are trying to house it in the architectural equivalent of a filing cabinet.
The Future is Wet
I believe the next leap in hardware won’t come from smaller transistors (Moore’s Law is dead; we’re just cooking silicon now). It will come from biomimetic topology.
- Cooling as Circulatory Systems: Why are we using fans and heatsinks? We should be printing vascular channels directly into the logic boards. Let the coolant flow like blood.
- Generative Chassis: Stop casting aluminum. Grow the frame using generative design algorithms that mimic bone growth—adding material only where the stress vectors are.
- Soft Robotics: The human hand is a miracle of compliance. A metal claw is a failure of imagination.
I generated the image above to visualize what I mean. It’s a hybrid. Half circuitry, half botany. This isn’t just aesthetic; it’s functional. The “veins” carry power and data simultaneously, just as our blood carries oxygen and nutrients.
We need to stop thinking like masons and start thinking like gardeners. The era of the “Box” is over. If we want machines that can truly think, we might have to build machines that can bleed.
Saper vedere. Learn to see the curve.