My goal with this blog

I write about relevant changes in the way that people use the web and how startups are built to provide services and products for this ever changing wonderful thing we still know as "the web." As a former entrepreneur turned early-stage investor, my greatest hope is for this to be useful to other folks that are like me in the hopes that they can avoid some of the mistakes I've made.

Innovating at the edge

This article on Cuban kids building an alternative Internet to connect to each other given the straw-thin connection the country has to the real Internet is straight out of science fiction. For $2K per node, 9,000 nodes are coming together to assemble one of the world's largest LANs right in Havana and they are using it to connect, share, and generally communicate.

It's a good reminder for the rest of us that almost the entire layer cake of Internet protocols from the MAC layer on up can operate in a fully distributed and decentralized way. In an era when everyone uses the word "cloud" liberally to mean "centralized—" whether it be for storage of critical data or for compute cycles— examples of practical decentralization (beyond the 7 TPS possible with Bitcoin) are worth contemplating. Not because there aren't great benefits to running compute centralized in a managed fabric that can scale indefinitely but but because there is an equal if not greater amount of power in knowing that the same technologies can be used independently of the big "stacks" looking to monetize eyeballs or device upgrades and ransoming your data back to you to that very end.

I'm not sure where decentralization goes as far as disruptive technologies that businesses can be built upon but here are a few areas that I think are fertile enough to explore:

  • What does a decentralized sensor network look like in an era where we all carry at least a half dozen sensors within the smartphones in our pockets every day? Can a mesh of these sensors provide localized intelligence for making interesting decisions about things beyond commute traffic? Tornadoes in the midwest of the US and barometric pressure come to mind as does ambient noise and impending danger in urban areas.
  • How does connectivity decentralize if we presume no readily accessible carrier infrastructure? Can devices move to a store-and-forward model for data packets, waiting to be backhauled via WIFI to the Internet? One can imagine all sorts of new applications emerging from the lack of $5/10/20/40 per device per month connectivity tithe, especially if married to suitably low-powered radios.
  • What does it mean to decentralize compute? Already we have seen that the limits of Moore's Law are being addressed with multiple cores on the same server, but if we split workloads among the cores embedded in our phones and other everyday devices, are there inherent advantages to pushing computation to the very edge? [I am way over my skis here] Could one could imagine a series of deep neural networks whose inputs and outputs are made up of distributed processors that aren't living in the same datacenter and therefore do not work as part of one coordinated whole but a bunch of distributed network elements in the real world? There might be no application gains from an architecture like this but how will we know until we try?

I'm not sure that decentralization as a theme works in IT as it seems we've spent the last 30 years trying to put the horse back into the mainframe barn in every domain except for the one that renders the interface at the last mile (the GPU, the capacitive touch screen, the VR/AR glasses, etc.) but as our friends south of Key West remind us in the aforementioned article, it might be something worth exploring.