After reading the first part of the article, maybe you already see that books and science are plateauing, but you may still find some hope in the internet and the digital revolution. It’s certainly true that we now have a planetary network that could theoretically connect any two people on the planet, if they knew the other person was worth connecting with. The speed and sheer volume of data flowing through this network are massive, and we even have a satellite system capable of connecting you to the internet almost anywhere on the globe. There are multiple technical layers to networking, and as a software engineer, I can attest that the hardware layer works in a genuinely clever way, and that even the low-level protocols like Ethernet and TCP/IP are well designed, almost in an academic way. Such technologies could well stay with us for a very long time, and where they have flaws, we mostly have a successor already designed that could easily replace them if the need arose. Unfortunately, this is where it ends: starting with the HTML layer — a human-readable language used to create web pages — the view becomes significantly more bleak.
HTML was originally designed to describe the structure and meaning of documents shared on the internet. The idea was that text would carry its own semantics, not merely its appearance. The hope was that this structure could later be harnessed — by internet browsers, yes, but also by the machines and algorithms that would one day have to organize and search the immense amount of content that was coming. Instead, we mostly used it to control how things looked. We reached for emphasis tags because they happened to render as bold or italic, not because anything was actually being emphasized; structure became a side-effect of chasing appearance. Over time this proved a real hassle, as the way we used HTML drifted too far from what it was meant to express — so we invented elements that carry no meaning at all, useful only as something to hang layout and styling on. Since then we’ve built one intricate layer of abstraction after another to smooth out a fabric that was practically designed to stay wrinkled. Technology after technology comes and goes, each offering some ingenious way to bend the ideal curve into a straight line — there’s even a dedicated programming language sitting on top, one that arrived from elsewhere to the rescue, flexible enough to accommodate any improvisation or ad hoc fix. It’s the most popular language in the world. Among programmers the tech stack is considered rather fancy and sophisticated, and AI is now expected to solve the whole thing once and for all, implementing custom hacks of any complexity fully automatically. This is the foundation on which the entire World Wide Web rests.
In information technology, AI is unbelievably effective wherever we’ve gotten stuck in a complex mess caused by poorly designed fundamental building blocks. AI “thinks” differently than any human, and it can take building blocks of any quality and assemble something complex out of them that simply works. Most of the time, at least. And yes — if you had a hunch, it was right: what you’re looking at is another local optimum. Once again. It seems almost obvious that at some point we’ll have to redesign the fundamental building blocks, drawing on several decades of experience. But for now it seems a couple dozen nuclear power plants fueling AI data centers can buy us enough time to stay stuck for another decade — maybe longer.
If you’re not a technical person, you may think I’m just echoing a disappointment common among software engineers about the way things are. In reality, the very opposite is true. People are genuinely enthusiastic about the way things are; I have never heard anyone complain. People like what brings in more business, and they’re willing to do things the right way only as long as it soon brings more revenue — or lets you control a larger share of the market. Maybe in the early days the idealistic attitude was acceptable, but the story of the big tech industry is one of recognizing that controlling the market and building the best technologies and products are two different objectives.
Initially people tried to invent technologies capable of searching the massive, ever-growing pile of internet documents. But they soon hit the barrier of digital literacy. It’s certainly possible to build a search engine that can find what you’re looking for — but it’s hard to command one when natural language is the only language you’re able to use. The current wave of AI has succeeded precisely because it promises that everyone can stay illiterate in technology: it provides a layer between human language and the various technical formalisms. I certainly can’t see this as a big win for humanity, but before AI there was only one way to rule the market — serve the average mediocre Joe and disregard everyone else. This is how Google became the planetary search engine: not by letting you find whatever you searched for, but by retrieving the stuff the average mediocre Joe searches for in the first place, with some respect for what he actually typed into the box. This is how most people will always find what they want — or at least during the part of the day each of us has to be the average mediocre Joe. My own experience during the rest of the time was quite different. I couldn’t find anything, and Google search kept getting worse, until the large language models came and helped me search again. By then Google was already an advertising behemoth controlling the market in key areas.
Absurdly, a lot of research exists into how to make the internet searchable in a better way, even for people like me. But this research can barely happen in universities. The only entities capable of experimenting on top of a giant base of data are the tech giants. So the solution may already exist — it may even be getting deployed during the training of AI models, in order to win market share there — but we may never benefit from it directly, since building the best possible product simply isn’t a priority.
It’s very similar with social networks. It’s certainly possible to build an infrastructure that connects you with the people on the planet you genuinely ought to know, given your vocation — but have you ever been able to do anything like that? No. On Facebook, for example, we don’t search for people; we add the ones we already know, the ones who help Facebook keep us scrolling so it can sell us more advertising. Here the distance from the ideal product couldn’t be more apparent. We don’t meet anyone new, and we pay for it by wasting time scrolling — and wasting even more time producing things to photo-document, so that others keep scrolling too, watching the perfectly wasted minutes of our perfectly useless active living.
We can’t find what we want, we can’t meet whom we want, we can’t fix the design flaws in the tech stack — and we can only soothe ourselves that AI will build a layer of natural-language interface to make it all more bearable and, simultaneously, more cemented, at the price of extreme energy consumption and significant environmental damage. What is this if not being stuck in an even worse local optimum — one that encompasses the whole of society, disposes of the time of our lives, and makes the earth barely livable?
As a practicing Catholic, I have to say I would rather die on the spot than accept the premises that force us to live on this kind of boring, damaging, exhausting plateau. To me it seems like hell. In many previous articles I’ve explained what it means to love God above everything else. It means trying to be here for others as much as possible, as He exists solely and perfectly to be here for others. In that view, it is not acceptable to seek profit without seeking the product that serves others best. And any advantage you gain by going against this rule is temporary almost by economic necessity — at least in a customer-based economy. I refuse to do it, and so can you and anyone else.
The same holds for fixing fundamental design flaws in technology and promoting basic digital literacy. You can refuse to give up; it isn’t some optional flavor of your life, it’s the essence of it — something without which life is not worth living.
By the same token, if a line of scientific research prevents you from looking into crucial aspects of creation, you should not engage in it, or at least you should not give up.
And you can never make peace with the idea that books and other cultural works can’t be objectively judged. When you read something and suddenly feel the desire that everybody would read it, it’s a good book — that’s the metric. It isn’t about perfect language, originality, enjoyment, or an unusual plot.
If everyone adopted the same attitude, our society would immediately stop plateauing. This is true by definition. So if you think God is not missed in this world anymore, that secularization doesn’t cripple us — think again. It does!