Not long ago, you could play the guitar while singing along to almost any song just by googling “song name + chords” on your phone. This was a really nice social gathering activity, especially among musicians. Your phone (or tablet) was invisible, just a tool. The focus was in expressing yourself through music and connecting with friends. Today, that is not possible any more.

Chord websites are plagued with ads between the lyrics, and even video! You need to scroll through the ads while playing and singing! The flow of making music with friends breaks apart. Not even using an ad blocker (which I do) solves the problem: some of the big sites (eg: Ultimate Guitar) won’t let you open the music sheets unless you install their mobile app. Or they force you to install it after you used their “3 free songs” or something. The apps suck. Believe me, I gave them a chance. They are full of ads and pop-ups trying to sell you their Premium Subscription™. From what I read in reviews, even when you do pay up they continue trying to upsell you teaching programs. Bummer. Besides, on principle: why on earth would I want to install a 108 MB app when all I want is to read plain text. Bottom line is: it’s no longer possible to just think of a song and play it.


Something similar is happening to the web in general. It’s fatiguing. Even before the advent of LLMs, content websites were becoming unusable. Daring internetors suggested adding “reddit” at the end of your google queries. More recently, the “before: 2023” trick appeared. Results since 2023 are of no use, it seems. Grifters are using LLMs to do SEO hacking at scale and bragging about it.

Over the past decade, 99% of our online experience has consolidated into 5 huge websites. The web has transformed into attention factory farming. As Cory Doctorow puts it, it’s suffering from a severe case of enshittification. I may be an old man yelling at a cloud, but most of the web feels awful these days. It used to feel awe-full. I don’t need to touch on the rot of social media yet again. It’s 2024 and everyone knows that. Web fatigue is real.

Some of you remember the 00s with nostalgia. The “good old days” of the Internet. Back then, most google searches took you to a real human’s blog. YouTube had no algorithm. Social media didn’t exist. People stayed weird online, instead of posing. The internet felt homey, not an unending ego contest where everyone’s trying to show how sophisticated and clever they are. Interest-themed online forums were actual communities. Some of us made life-long friends there!

The web today feels dry, overly polished, dramatic, and soulless—somehow all at the same time. What can we do about it?


There’s been a recent move towards recovering the lost internet mojo. A rebirth of newsletters, where you can read people that you like, at your own pace—as opposed to at the pace of the drama-maximizing algorithms. However, I still feel we can do better. Lately, I’ve been thinking about how we can reclaim part of the web back into our hands. The big problem is VC-funded, growth-maximizing, startup mentality. That model wins in the decade time-frame. But we need the web to work for centuries. Time to start thinking longer term.

What Signal is doing is interesting. A donation-fueled foundation building a core component of the internet: messaging. I used the app for years now and I greatly admire Moxie, the founder. However, their product is pretty bad. Instead of polishing the core features, they choose to add “stories”, which is more of the same enshittification process for-profit apps went through. After seeing this, I canceled my monthly donation, disappointed. Plus, as a latin-american I always felt weird supporting a US-based foundation: salaries are too high and they have a certain way of doing things which don’t exactly match my values. Still, a foundation building open-source seems like a model worth studying.

Another interesting model to think about is crypto protocols. What Optimism is doing with Retroactive Public Goods Funding is fantastic. They’ve rewarded open-source work with token drops to projects that had community-vetted impact. However, crypto still has a greed aura around it. Founders seem more focused on following the latest crypto-twitter narrative than in polishing their core product. And many users and developers jump from one project to the next merely trying to farm airdrops. Few crypto projects escape these traps. Still, retroactively funding public goods (with or without crypto) seems like a model worth studying.

Some people are doing experiments at the solo or micro-community level.

  • Derek Sivers shares his tech independence walkthrough, where he shows you how to set up and maintain your own private server as a foundational tool for sovereignty. It enables you to host websites, email, contacts, calendars, and storage on your own terms, free from the changing whims of any company.
  • Cory Dransfeldt speaks about how a possible solution is to push decentralization of content via making better open tools for personal blogging, forums and niche communities.
  • Herman Martinus and his “my product is my garden” approach to running an online business. His “Building software to last forever” is also short and sweet.
  • The permacomputing community is thinking about creating sustainable computing systems that minimize resource usage and enable adaptability—which industrial websites lack.
  • Ivan Vendrov, in his excellent “The Tyranny of the Marginal User”, shows how focus on user acquisition makes apps dumb down with time, with a brilliant characterization of the next user, which he calls “Marl”.

For months, I’ve been subscribed to the Recomendo newsletter. With each weekly email, I wondered why on earth I was interested in 6 random recommendations (like backscratchers and other useless stuff) from Americans, whose lives have little in common with mine. Why the hell hadn’t I unsubscribed yet? I recently discovered the reason in their email footer: “Recomendo is an authentic, hand-crafted, human-written weekly newsletter”. The world is flooded with inauthentic, algorithm-pleasing, LLM-written, engagement-polished content. The sheer simplicity of a couple of folks writing about what they find cool has a refreshing value.


Human curation is coming back. As I said before, we’ll soon see social media as we see smoking: we all know it’s harmful, even if some people still choose to do it. But, what’s next?

What I see coming is a change of focus on how we build the web. Most websites today are built with the same growth- and metrics-driven mindset. To change the output, we need to change the process. We need to rethink the machine that builds the machine.

A noble goal, alluded by some of the thinkers cited above, is to think deeply about website robustness in time. A focus on optimizing stability and perfecting a simple thing rather than growing it. Instead of obsessing about how to get the next +10% growth in your metrics, think about how to make what you’re doing stable and to endure in time. Make the web last longer, not grow larger.

We already know how to grow online businesses. There’s more than 30 years of experience to learn from. But we don’t really know how to make digital tools that last centuries. Ask yourself: How would you build your product to last 25 years? To outlast you? To last 100 years? 200? What design decisions would you make? For sure they won’t be the same as if you’re looking for short-term quarterly growth. And they may bring a return to quality over quantity. Users are yearning for quality.

Some notions of software longevity can be found in crypto. Bitcoin seems to be running pretty fine after more than 15 years, thanks to good incentive design. The block size wars and forks like Bitcoin Cash, BitcoinSV and others show that the social pressure to grow at all costs will exist even when there’s no company behind the technology. I originally thought tech’s obsession with growth came from the governance structure of for-profit companies, and their fiduciary duty to investors. But it think it runs deeper than that: human nature seeks growth. Obsession with quality is rarer. The fact that Bitcoin remains so good after 15 years of development and many different people getting involved speaks of the fantastic incentive design built by Satoshi into the technology.

Not-for-profit foundations for software development like Signal and Nomic Foundation, have a well-written charter and governance that drives continuity in time.

Yet another path may lie in Neal Stephenson’s phyles. In his—usually prophetic, mind you—sci-fi voice, he shows how technology can bring new meaning to multigenerational organizations. Are we maybe close to seeing some of the first hereditary online estates, Zaibatsus of the digital age? What would it take to build an online phyles?

To achieve the goal of a multi-generationally stable website (or app, or tech), some requirements come to mind:

  • Clearly stated values and mission.
  • An almost obsessive focus on maintaining the project for a long period of time.
  • A minimalist technical architecture focused on simplicity, reliability and low maintenance.
  • Incentive design that promotes new players to join the project, to compensate natural turnover.
  • A legal or on-chain governance structure that discourages single points of failure or too much change.
  • An explicit rejection of growth hacking.
  • A sustainable funding model—not reliant on ads, selling data, or a constant need for growth.
  • A product design philosophy that values polishing over feature adding.

All of this seems very hard, compared to how software is commonly built. I think, however, that it’s worth trying new models. I don’t know what specific form the new web will take. But the age of attention factory-farming by 5 huge tech conglomerates will someday be over. We might as well start thinking about it.

Liked what you read? Consider subscribing below:

Additional Resources and Learning


Thanks to Camila Russo for providing feedback on earlier versions of this post.

Thanks to the Sur Computing Lab and deadpine for some inspiring conversations that led to writing this blog post.

Cover photo by MidJourney

Comments and Discussion

Email a private comment