It was a quiet Tuesday morning when the scales finally fell from my eyes, prompted by a beautifully provocative piece in Quanta Magazine asking a question we should have asked fifty years ago: does computer science actually need computers? The answer, I realized with the force of a religious awakening, is a resounding no. In fact, physical machines are the worst thing to ever happen to the discipline.
We have lost our way. The original sin of modern technology was not data harvesting or the invention of the reply-all button, but the tragic, pedestrian decision to take the pristine mathematics of the Turing machine and attempt to build one out of copper, silicon, and plastic. What was meant to be a beautiful, ethereal thought experiment was dragged into the mud of physical reality.
I brought this up on Thursday during a two-hour matcha ceremony in Aspen with a prominent venture capitalist and a former director of the National Science Foundation. They were initially skeptical, shackled as they are to the idea that a spreadsheet needs to turn on or that an MRI machine requires electricity. But as I patiently explained to them, the theoretical framework for processing data is flawless. It is only the vulgar execution of that data by actual microchips that introduces bugs, lag, and overheating.
Think of the majesty of an algorithm in its natural state. On a chalkboard, a sorting algorithm is a testament to human ingenuity, a flawless ballet of logic and symbolic reasoning. But the moment you force it to run on a Dell Inspiron to organize a mid-sized logistics company’s payroll, you degrade it. You force a Pegasus to pull a garbage truck.
Just yesterday, I was reviewing a brilliant preprint from a lab at MIT that proved, mathematically, that a new machine-learning model could cure a specific strain of mouse leukemia. The researchers ruined the paper in the final section by demanding a six-million-dollar supercomputer cluster to actually run the simulation. I sent it back with a simple note: the proof is elegant; don't cheapen it by turning it on. If the math says the mice live, we should simply rejoice in the theoretical survival of the mice.
Naturally, the entrenched interests in the hardware sector are terrified of this paradigm shift. They cling to their physical infrastructure like a safety blanket, terrified of a world where we simply think about the code rather than executing it.
If you advocate for the total cessation of physical computing, you are advocating for the immediate collapse of the global supply chain, aviation networks, and modern agriculture, none of which can be sustained by a guy staring thoughtfully at a whiteboard.
It is exactly this kind of defensive, small-minded pedantry that keeps humanity tethered to the dirt. Hewitt and his ilk are so obsessed with the supply chain that they completely miss the breathtaking elegance of a zero-hardware future. They cannot see that by abandoning the physical computer, we elevate the science.
As Carl Sagan would say, the cosmos is within us. We do not need a liquid-cooled server rack to understand the universe; we only need the limitless capacity of the human mind to conceptualize finite automata. By stripping away the keyboards and the monitors, we return computer science to the philosophers, the dreamers, and the columnists who have the vision to see beyond the screen.
I have already begun implementing this transition in my own life. Last week, I unplugged my home router, carried it to the edge of the Santa Monica pier, and let it sink into the Pacific. I now handle all of my digital correspondence theoretically. When an editor asks for a draft, I simply construct the optimal routing protocol in my mind, prove mathematically that the essay could be transmitted, and consider the job done. The fact that they have not published my recent work only proves they are trapped in the old, hardware-dependent mindset.
The economic implications are, frankly, thrilling. Silicon Valley could pivot entirely to a conceptual model. Imagine the venture capital landscape if Andreessen Horowitz could raise billions for a startup that promises absolutely no physical deliverables. Instead of a software-as-a-service model, we transition to a software-as-a-silhouette model. You don't subscribe to a cloud storage provider; you pay a monthly fee to conceptualize an infinitely expanding digital warehouse. The profit margins would be staggering, unburdened by the vulgar realities of cooling costs and real estate.
Apple wouldn't need to mine rare earth metals or run massive assembly plants; Tim Cook could simply stand on a stage in Cupertino and describe the idea of the iPhone 17. We would gladly pay a premium for the privilege of knowing that a faster, thinner device exists in the Platonic ideal, completely free from battery degradation.
It is time to let go of the crutch. We must unplug the servers, smash the silicon, and allow computer science to finally soar as a pure exercise of the intellect. I, for one, have never felt more connected to the digital world than I do right now, dictating this column to an unpaid intern who is desperately typing it into one of the last remaining laptops I haven't managed to destroy.