A central advantage of a cloud is that the bigger it gets, the smaller and thinner our devices can be. The cloud does all the work, while the device we hold is just the window into the cloud’s work. When I look into my phone screen and see a live video stream, I am looking into the cloud. When I flick through book pages on my tablet, I am surfing the cloud. When the face of my smartwatch lights up with a message, it is coming from the cloud. When I flip open my cloudbook laptop, everything that I work on is actually somewhere else, in a cloud.
The ambiguity of where my stuff is and whether it is in fact “mine” can be illustrated by the example of a doc on Google. I usually use the Google Drive app to write a marketing letter. “My” letter appears on my laptop or my phone, but its essence lives in Google’s cloud, dispersed across many far-flung machines. A key reason I use Google Drive is its ease of collaboration. A dozen or more collaborators can see that letter on their tablet and work on it—edit, add, delete, modify—as if it were “their” letter. Changes made on any of those copies will appear simultaneously—in real time—on all other copies anywhere in the world. It’s kind of miraculous, this distributed cloud existence. Each instance of the letter is much more than a mere copy, a term that suggests an inert reproduction. Rather, each person experiences the distributed copy as the original on their device! Each of the dozen copies is as authentic as the one on my laptop. Authenticity is distributed. This collective interaction and distributed being makes the letter feel less mine and more “ours.”
Because it lives on the cloud, Google could easily apply cloud-based AI to our letter in the future. Besides automatically correcting the spelling and critical grammar, Google might also fact-check the statements in the letter with its new truth-checker called Knowledge-Based Trust. It could add hyperlinks to appropriate terms, and add (with my assent) smart additions that improve it significantly so that it further erodes my sense of possession. More and more of our work and play will leave the isolated realm of individual ownership and migrate to the shared world of the cloud in order to take full advantage of AI and other cloud-based powers.
I already google the cloud for answers instead of trying to remember a URL, or even the spelling of a difficult word. If I re-google my own email (stored in a cloud) to find out what I said (which I do) or rely on the cloud for my memory, where does my “I” end and the cloud start? If all the images of my life, and all the snippets of my interests, and all my notes, and all my chitchat with friends, and all my choices, and all my recommendations, and all my thoughts, and all my wishes—if all this is sitting somewhere, but nowhere in particular, it changes how I think of myself. I am larger than before, but thinner too. I am faster, but at times shallower. I think more like a cloud with fewer boundaries, open to change and full of contradiction. I contain multitudes! This whole mix will be further enhanced with the intelligence of machines and AIs. I will be not just Me Plus, but We Plus.
But what happens if it were to go away? A very diffused me would go away. Friends of mine had to ground their teenager for a serious infraction. They confiscated her cell phone. They were horrified when she became physically ill, vomiting. It was almost as if she’d had an amputation. And in one sense she had. If a cloud company restricts or censors our actions, we’ll feel pain. Separation from the comfort and new identity afforded by the cloud will be horrendous and unbearable. If McLuhan is right that tools are extensions of our selves—a wheel an extended leg, a camera an extended eye—then the cloud is our extended soul. Or, if you prefer, our extended self. In one sense, it is not an extended self we own, but one we have access to.
Clouds are mostly commercial so far. There is the Oracle Cloud, IBM’s SmartCloud, and Amazon’s Elastic Compute Cloud. Google and Facebook run huge clouds internally. We keep coming back to clouds because they are more reliable than we are. They are certainly more reliable than other kinds of machines. My very stable Mac freezes or needs to be rebooted once a month. But Google’s cloud platform was down only 14 minutes in 2014, a near insignificant outage for the immense amount of traffic served. The cloud is the Backup. Our life’s backup.
All business and much of society today run on computers. Clouds offer computation with astounding reliability, fast speed, expandable depth, and no burdens of maintenance for users. Anyone who owns a computer recognizes those burdens: They take up space, need constant expert attention, and go obsolete instantly. Who would want to own their computer? The answer increasingly is no one. No more than you want to own an electric station, rather than buy electricity from the grid. Clouds enable organizations to access the benefits of computers without the hassle of possession. Expandable cloud computing at discount prices has made it a hundred times easier for a young technology company to start up. Instead of building their own complex computing infrastructure, they subscribe to a cloud’s infrastructure. In industry terms, this is infrastructure as service. Computers as service instead of computers as product: access instead of ownership. Gaining cheap access to the best infrastructure by operating on the cloud is a chief reason so many young companies have exploded out of Silicon Valley in the last decade. As they grow fast, they access more of what they don’t own. Scaling up with success is easy. The cloud companies welcome this growth and dependence, because the more that people use the cloud and share in accessing their services, the smarter and more powerful their service becomes.
There are practical limits to how gigantic one company’s cloud can get, so the next step in the rise of clouds over the coming decades will be toward merging the clouds into one intercloud. Just as the internet is the network of networks, the intercloud is the cloud of clouds. Slowly but surely Amazon’s cloud and Google’s cloud and Facebook’s cloud and all the other enterprise clouds are intertwining into one massive cloud that acts as a single cloud—The Cloud—to the average user or company. A counterforce resisting this merger is that an intercloud requires commercial clouds to share their data (a cloud is a network of linked data), and right now data tends to be hoarded like gold. Data hoards are seen as a competitive advantage, and sharing data freely is hampered by laws, so it will be many years (decades?) before companies learn how to share their data creatively, productively, and responsibly.
There is one final step in the inexorable march toward decentralized access. At the same time we are moving to an intercloud we will also move toward one that is fully decentralized and peer to peer. While the enormous clouds of Amazon, Facebook, and Google are distributed, they are not decentralized. The machines are run by enormous companies, not by a funky network of computers run by your funky peers. But there are ways to make clouds that run on decentralized hardware. We know a decentralized cloud can work, because one did during the student protests in Hong Kong in 2014. To escape the obsessive surveillance the Chinese government pours on its citizens’ communications, the Hong Kong students devised a way to communicate without sending their messages to a central cell phone tower or through the company servers of Weibo (the Chinese Twitter) or WeChat (their Facebook) or email. Instead they loaded a tiny app onto their phones called FireChat. Two FireChat-enabled phones could speak to each other directly, via wifi radio, without jumping up to a cell tower. More important, either of the two phones could forward a message to a third FireChat-enabled phone. Keep adding FireChat’d phones and you soon have a full network of phones without towers. Messages that are not meant for one phone are relayed to another phone until they reach their intended recipient. This intensely peer-to-peer variety of network (called a mesh) is not efficient, but it works. That cumbersome forwarding is exactly how the internet operates at one level, and why it is so robust. The result of the FireChat mesh was that the students created a radio cloud that no one owned (and was therefore hard to squelch). Relying entirely on a mesh of their own personal devices, they ran a communications system that held back the Chinese government for months. The same architecture could be scaled up to run any kind of cloud.