The dumb trend toward crippled computing
Tech companies refuse to unlock the exponentially increasing compute power that humanity is manufacturing at an impressive scale. It's stunting our ability to solve novel and planet-scale problems.
Central Processor output is exploding…
One true thing about our world is that we’re seeing an absolute explosion in the availability of ‘compute’. Which is a sophisticated tech way of describing anything with a silicon chip that can do significant math. Compute-capable processors are increasingly in most things with electronics in them (thanks to Cortex, the god of ARM, may her name be ever blessed).
But we’re not connecting and aggregating
‘Central processing’ chips - that process math and are programmable - are in everything. But the tragedy of our current trajectory is that aggregate compute power is not easily accessible or accessed. It’s all fragmented within small, mostly single-purpose devices. From the smallest IoT device in your home to your Ryzen 7000 desktop, you cannot easily connect available computing power together to solve a ‘bigger’ problem. Talk less of connecting to the compute power owned by your neighbor or your friend across the world in say, Singapore. Even though you are connected by capable and fast internet connections. And that most of the time these machines are idle.
Aggregating computing power is more critical than ever
Worse, no current open protocols or standards allow us to do this easily. The only real attempt I could find (If I missed sth major, tell me in the comments) are dated and really old. In fact, major OS vendors make it even hard to properly harness the processing power you have purchased in the devices they manufacture. The compute inside an iPhone is not easily unlocked. Writing software for it has an unreasonably high bar set by and guarded by Apple via their App Store and developer program. Of which they exact complete control. Even a simple iPhone console app that processes text cannot be deployed without Apple as the mediator, talk less of one that uses some of the growing AI processing capabilities in the phone’s CPU. Even if you have a jailbroken phone, it’s nigh impossible for most people. Even if you manage to get an app published (to do it on your own phone you have to get a developer account from Apple), you only get to use slivers of the processing power since it’s mostly mediated via APIs that barely unlock the full power of the hardware.
This vision of crippled computing has gained currency at the exact same time that the knowledge about how to code and the means to do so is becoming more commonplace in the world’s population. In the previous WinTel-dominated world, it was much easier to unlock all your computer’s power. However, in those early days of computing, it was a specialized skill, not well distributed in the human population.
Further, compute is being trapped inside smaller and smaller devices, which are becoming more and more plentiful. Around me in my office are various devices with decent computing power. Some x86, a lot are ARM Cortex processors; single, dual, quad and octa-core. Each idles and labors at about 1 - 40% capability by design. All rarely busting a gut at all the pre-approved app store programs I throw at them. All are ‘living’ under their potential, but also uniquely unable to combine to solve a big problem:
Such as the fact that I recently traveled for the Thanksgiving holiday and wanted to write an AI app that could recognize a person picking up a package previously dropped off by a delivery person (I would assume theft in that case). I was completely unable to do so across any of these devices including my doorbell camera and its quad-core processor, which essentially runs only one pre-approved application.
Apple made ‘Crippled Computing’ popular 😖
The advent of small processor-laden devices that live inside walled computing gardens is slavishly following the example of Apple Computing, which seems to be the biggest proponent of a future of crippled computing1. By controlling access to the platform, they ensure that their business model will maximize profits. I still remember when Apple revoked its Mac OS license in 1997 to reverse its trend towards a more open platform. Even the future of cars is generally leaning towards crippled computing at a time when they are becoming just specialized computers (both ICE and EVs) on wheels.
This view of the world is obviously wrong. Even as we trend towards a planet saturated by compute, we (tech companies and tech people) are determined to isolate this exponentially growing computing power and neuter it in different walled gardens; inside unopenable enclosures and inaccessible OSes. We are not building open tools and operating systems that enable citizen developers to autonomously solve novel problems using interconnected compute. Even as the planet’s population becomes smarter, more creative, and has more free time to contemplate and attempt to solve the common problems of humanity. We have not considered how to do with processing power, what idealistic researchers did with TCP/IP in the 70s and 80s.
We need to do better
We need to more seriously contemplate and build a world where all compute, in every device, can be freed for developers. And where that compute can be easily connected together, to solve potentially planet-scale problems that companies and their employees, in the tech enclaves around the world, cannot contemplate.
Aside from MacOS. For now.