Build A Better Internet (Part 2)
Two roads diverged...
It’s been a long journey to this point in Internet history: from the early days of dial-up modems and AOL (Web 1.0), through a maze of mobile apps and social media (Web 2.0), up and through the cloud, with an AI tailwind blowing at our backs, all leading us to here.
But where is here?
It’s not an easy question to answer: each of us is extremely online in their own way. The Internet in 2025 is a splintered “here,” a multiverse of habits and perspectives. My algorithm is different from yours.
With that said, let’s be real; we’re all basically doing the same thing:
Scrolling [Instagram, TikTok, Reddit, X, Bluesky]
Watching [YouTube, Netflix, Disney +, Apple TV]
Chatbotting [Chat GPT, Claude, Grok]
Gaming [Fortnite, Roblox, Minecraft, Steam games, mobile games]
Reading [E-mail, Substack, NY Times, Kindle App, whatever random website you land on while doomscrolling]
So what’s next?
For now, it appears that two paths lie ahead—each with its own set of advocates imploring us forward.
Path 1: Centralized Tech + AI Agents + Advertising
This is the way of the big platforms. More sophisticated algorithmic feeds and targeted advertising. New and improved iPhones. The seamless integration of AI-powered features that anticipate your needs and simply do whatever needs doing.
Open AI CEO Sam Altman recently described the “platonic ideal” of AI as “a very tiny model with superhuman reasoning, 1 trillion tokens of context, and access to every tool you can imagine” — an all-in-one sidekick that’s always there. Always.
Path 1 is the Internet as it exists now… on AI “agentic” steroids.
Path 2: Decentralized Tech + Blockchain + Open Protocols
This is the way of the smaller upstarts. The Rebels to the Empire. Payments that flow peer-to-peer, free from middlemen — (Bitcoin, Ethereum). Community-driven platforms where your identity isn’t harvested for ad data — (Bluesky, Mastodon). Messaging where all your chats are encrypted and no corporation is reading (or listening) to them to sell you ads (Signal).
Path 2 is the Internet as it could be… a little wilder, a little less convenient (for a time), but more democratic and competitive.
Before we go further, a quick show of hands: who wants to keep scrolling along the slick, tightly controlled highway that is Path 1? Conversely, who wants to escape corporate control, install a bunch of weird apps, and bump down the cobblestone road that is Path 2?
Alright, cards on the table: this is a classic false dilemma and an oversimplification of how the Internet’s future will likely unfold. These two “paths” will inevitably bleed into each other, and I don’t see many Path 2ers throwing away their iPhones. Meanwhile, the Path 1 cohort is like:
Still, the “fork-in-the-road” metaphor is a useful framework, so we’ll stick with it for now.
Let’s meet the folks pushing for Path 2—a diverse, tech-savvy group fond of coining new terms for the next internet (“next-ernet”? No, that doesn’t work).
Web 3. The Fediverse. The Social Web?
The term “Web3” gets a bad rap today. Like “the metaverse,” critics pounce on it, arguing that since it hasn’t fully materialized, it’s either a has-been or a never-was.
Coined by Ethereum co-founder Gavin Wood, Web3 is essentially the traditional internet + blockchain networks. As the story goes, Web 2.0 has meant massive platforms capturing users in increasingly enshittified walled gardens. Web3 would bring back the decentralized ethos of Web 1.0 by empowering individuals themselves to own and control their parcels of the Internet.
According to Chris Dixon, partner at a16z and author of Read, Write, Own: Building the Next Era of the Internet, blockchain networks “are a new construction material for building a better internet.”
The blockchain—decentralized systems that validate, record, and manage transactions across a distributed ledger—certainly holds promise for reshaping internet ownership. But, as Dixon helpfully explains, there are two camps within the blockchain—the “casino” and the “computer” —and only one has dominated the headlines to date, and not for the better.
The casino is the speculative, finance-driven side—NFT hype, volatile crypto investments, Ponzi-like schemes, scams, thefts. Sam Bankman-Fried in the Bahamas.
Remember NFTs and the Bored Ape?
The computer, by contrast, are the people interested in genuine innovation in decentralization and ownership. They’re a quieter group whose primary concern is infrastructure and architecture: secure, open, decentralized networks that users control directly, free from Big Tech intermediaries.
For Dixon, a Web3 “blockchain network” combines the decentralization-benefits of a Web 1.0 “protocol network” with the incentive structure of a Web 2.0 “corporate network.” Specifically, blockchain networks offer built-in economic incentives through tokens and smart contracts, rewarding users and developers directly for their contributions. This incentive alignment theoretically ensures network sustainability, active participation, and innovation without sacrificing the decentralized ideals of the early internet.
But, even with Dixon’s casino vs computer distinction, some aren’t buying it.
Molly White writes the blog “Web 3 is Going Just Great (and is not definitely not an enormous grift that’s pouring lighter fluid on our already smoldering planet).” Her site meticulously tracks failures, scams, and absurdities in the crypto space with a snarky edge. Reviewing Dixon’s book, she calls him to task for glossing over the blockchain’s many flaws and failing to properly acknowledge that a16z, his own venture firm, not only contributed to the formation of the Big Tech monopolies in the first place but now stands to profit handsomely if Web3 ever takes off.
She agrees with Dixon that the current Internet is fundamentally broken, calling it “a commercialized, capitalist hellscape controlled by relatively few massive companies.”
(“Hellscape” seems a tad harsh to me, but okay.)
In contrast to Web3, White advocates for "Protocols, not Platforms," a term popularized by tech writer Mike Masnick in an article of the same name. The idea behind it is straightforward: emphasizing open standards and interoperable systems that anyone can build upon, reducing the risk of centralized control and monopolization by any single entity.
Dixon counters that merely promoting regulations and open protocols hasn’t significantly undermined the entrenched dominance of Big Tech; he writes, “There are only two known network architectures that preserve the democratic and egalitarian spirit of the early internet: protocol networks and blockchain networks. If new protocol networks could succeed, I would be the first to support them. But after decades of disappointment, I'm skeptical.”
This debate brings us to the so-called Fediverse, a decentralized network of social platforms. It’s a little hard to explain, but Reddit user Beneficial-Bid-8850 does a nice job:
“The Fediverse is an interconnected social platform ecosystem based on an open protocol called ActivityPub (AP), which allows you to port your content, data, and follower graph between networks. Just imagine if you took X, TikTok, Snapchat, Instagram, and Facebook and made them all interoperable so you could post anything from anywhere, and all your followers would be guaranteed to see it. And if you wanted to leave one platform for another, you could bring all your content, all your followers, everything with you. Best of all: most of the Fediverse is non-commercial and not for profit (with the exception of Threads, which is part of the Fediverse) and nobody can "buy" the AP protocol.”
Representative apps include Mastodon—a Twitter-like platform without corporate oversight—and Bluesky, a decentralized alternative backed by Jack Dorsey. Mike McCue, the CEO of Flipboard and an internet entrepreneur since the Netscape days, is a big believer in the Fediverse. He doesn’t love the word itself; he prefers calling it “the social web.”
Whatever you call this open, decentralized, interconnected social network, it’s clear there’s momentum behind it—just not from the big platforms.
The problem is that most people don’t care if content creators are shortchanged, or if their data is sold to the highest bidder. They like their feeds. So do I. So do you.
There’s little market incentive to disrupt what’s working for big platforms and “mostly working” for users, who refresh their social feeds like kids hooked on sugar. Just building an open alternative and banking on people’s sense of virtue isn’t enough.
This brings us back to Path 1: instead of abandoning entrenched internet giants—deeply embedded through software, data centers, and costly compute—the solution may be to reform from within. Better regulation, greater transparency, and selective integration of Path 2 features, as companies like Meta and Microsoft have signaled openness to, could bridge the gap.
So let’s say we just keep cruising right along Path 1, as I would argue 99% of people will.
What will the Big Internet look like between now and 2030? How much wiggle room do we have on this noisy and crowded interstate to make it a little less soul-sapping?
The Agentic Web and Its Tradeoffs
As we cruise down Path 1, the centralized tech highway, the future looks slicker than ever. Microsoft CTO Kevin Scott recently pitched the idea of an “Open Agentic Web,” where AI agents act as hyper-capable personal assistants, seamlessly connecting you to whatever you need—information, services, or products—without the friction of traditional browsing.
As Scott’s title suggests, he envisions these agents operating according to open protocols: “So you need protocols, things like MCP and A2A and things that likely are going to be emerging over the coming year that will help connect in an open, reliable, interoperable way the agents that you are writing and agents that are being used so actively now by hundreds of millions of people to be able to go access content, to access services, to take action on behalf of users in fulfilling the tasks that have been delegated to them.”
These agents, powered by increasingly sophisticated models, are already transforming how we interact with the internet. Consider search: AI models are, to use a memorable term from Chris Dixon, “one-boxing” results, meaning when you Google something, you often get a direct answer from the AI, no need to click through to a website. The rise of AI Overviews has led to a notable decline in click-through rates (CTR) to third-party websites, with some studies showing a drop of nearly 30% in CTR since May 2024.
This shift is shrinking the open internet. Fewer clicks to websites mean less traffic for independent creators, bloggers, and small businesses, pushing more users into the walled gardens of big platforms—e.g., Google, Amazon, or Meta—where the walls are higher than ever. These platforms control the data, the algorithms, and the compute, locking users into curated ecosystems. The dystopic vision is that, as users are placated by the platform’s AI-powered, in-house solutions, the vast majority of the internet simply goes away.
Meanwhile, AI agents are becoming ridiculously powerful. By 2030, agents will book your vacations, negotiate deals, and even craft personalized content in seconds, connecting you to your desires with unprecedented ease. You’ll tell an agent, “Make a 20 minute children’s show about my kid attending Hogwarts, being sorted into Gryffindor, and winning the Quidditch cup,” and then there it’ll be. I suspect legacy IP (the Mickey Mouses and Harry Potters of the world) will become extremely valuable as AI-created content floods the internet and people long for the old and familiar, the tried and true.
Sam Altman wasn’t kidding about that “platonic ideal” of AI—a tiny model with superhuman reasoning and infinite context, acting as your always-on sidekick. This is where the internet is headed.
Here’s the rub though: as convenient as this Path 1 future is, it feels unsatisfying. The economic picture is skewed, with big platforms hoarding the profits while creators and smaller players scramble for scraps. The internet risks becoming a vertical scrolling trap—endless feeds of algorithmically served content, optimized for engagement but myopic in scope.
It’s a far cry from the vibrant, horizontal network we might imagine.
Since he published the idea as a feature essay at Wired in 2019, I’ve been fascinated by Kevin Kelly’s vision of the “Mirrorworld,” a digital twin of reality, a “1-to-1 map of almost unimaginable scope. When it's complete, our physical reality will merge with the digital universe.”
Shouldn’t the internet let you virtually attend your local church service, join a town hall meeting, check how busy the beach is, or buy from a nearby store with the same ease as ordering from Amazon? Shouldn’t you have a provable digital identity—see WorldID, and yep it’s also backed by Sam Altman—that lets you be your authentic self online, seamlessly blending your real and digital lives?
Path 1’s AI-driven convenience is impressive, but it falls short. The internet could be a seamless blend of physical and digital, where people create, connect, and transact freely.
Instead of walled gardens, picture an internet so integrated it feels invisible, like water to a fish, enabling authentic interaction and commerce. That’s the goal, and Path 1, despite its strengths, doesn’t get us there.
So what, my fellow adventurers, should we do? Is there a third path? What else could the internet be?
We’ll explore these questions (and more!) in the next and final leg of the Build a Better Internet questline. I hope you’ll join, and, as always, thanks for reading!