The Computational Web
This note is a placeholder for a potentially longer, more in-depth essay I hope to publish in the future.
Decentralized social media, like blockchain and crypto, promises to topple old power structures and hand the Internet back to the people. But the Internet runs on a stack of technologies, and if you venture just a few layers down, you'll see some familiar faces. The techno-oligarchy is not only alive and well, but it's getting stronger. They moved underground to the Internet's infrastructure layer.
While companies like Bluesky are open-sourcing the protocol layer, Zuckerberg, Bezos, Pichai, and Nadella are building a new Internet on which those open-source technologies will run. It's an Internet that relies on massive kingdoms of data centers, thousands of miles of deepsea cable, and billions of dollars in specialized GPUs. This is Web 3.0. This is The Computational Web.
I'm defining The Computational Web by the increasingly massive amounts of computing required to run the modern Internet, thanks to AI and decentralized technologies and the elite group of tech firms that can meet those demands—Google, Amazon, Microsoft, and Meta (GAMM). The more we expect these eyewatering levels of computation from our apps and websites, the more The Computational Web takes hold as the successor to Web 2.0.
Take AI, for example. We may not like it, but these large language models and recommendation engines have been able to intertwine into our daily lives in a way blockchain and crypto haven't. AI is here to stay for the foreseeable future.
Compute is expensive and increasingly difficult to scale. These hurdles make compute accessible only to the largest tech firms in the world. Shoehorning AI features into our apps isn't just tech bros following their tail. It's a strategy to set the expectation that all consumer technology requires resource-hungry AI. If all technology requires AI, and only a handful of companies are equipped to handle the computational power that Al requires, then computation becomes a moat too deep for competition to cross. The Computational Web grows stronger.
"Computational power, or compute, is a core dependency in building large-scale Al. […] It is profoundly monopolized at key points in the supply chain by one or a small handful of firms." –AI Now, Computational power and AI
We can debate what Web 3.0 will be, but it's been here the whole time, and it's not Web3.
Decentralized social media, blockchain, cryptocurrency, and artificial intelligence are part of the new web, just as the automobile is part of the electric battery market. Tesla isn't an automobile manufacturer. It's a battery company that sells the most battery-hungry product possible. Similarly, Google, Amazon, Microsoft, and Meta aren't platform companies. Not anymore, at least. They're an Internet infrastructure oligopoly that sells retail computing power (Meta is not quite there yet, but I believe they're pivoting to offering cloud computing services).
What we're seeing with all these computing-hungry products and decentralized networks is just the top layers of the stack falling in line for the new iteration of cyberspace—The Computational Web.
So, I wonder, how can we have this new decentralized Internet if all the infrastructure required to run it is owned by just four companies?
Related links #
Big tech’s great AI power grab
The Great Computing Power Chase: Why It Matters and How We’ll Win
How cloud computing became a global monopoly
Meta plans to build a $10B subsea cable spanning the world, sources say