I’ve said multiple times that I believe the web is the future. I still think that is true—it’s by far the best platform to publish most GUI apps which don’t depend on deep system or hardware integration. And I feel validated in that claim by the myriad of “AI” applications which have been sprouting up in the past few years, and the UIs of which have largely been web apps.
And while I do have very strong (and lately, quite mixed) feelings about this bubble, it has prompted a shift in the web landscape. So if the web is the future, what is the future of the web?
The Great Centralisation
I think, in a way, the web has always been changing. It’s a continuously evolving platform, and even before large language models there was a shift from many independent sites and publications towards more centralised content silos. Over the years, we were browsing the web less and less, instead spending more and more of our time on platforms built on the web.
Sure, there’s the “Indieweb”, which aims to keep the spirit of many smaller interconnected websites alive, but I feel like the size of that bubble is negligible compared to the amount of people to which “the web” means Google, Facebook, Instagram, Reddit, Wikipedia, and YouTube.
The AI-pocalypse
So is it really surprising that the centralisation keeps progressing? Content went from small, independent websites to centralised silos and now into the training data of LLMs. Already, there have been articles published about how traffic to large websites has dropped significantly since Google launched its “AI” summaries. Small, independent websites didn’t have much of a chance to be found and get traffic even before that introduction.
You either stumbled on a link to a page on a social network or in another piece of content, or you simply wouldn’t know it existed. There’s a reason I search for my own blog posts on my own website instead of using a search engine when I need to reference something.
I think the introduction of ChatGPT and other such tools created an expectation in users that they would get all the information they need based on a single query without further actions, so it’s only natural that search engines followed that expectation and cut the other interactions like clicking a link and probably dismissing at least one banner from the process.
Fluff Exhaustion
In part, I feel like people would rather reach for a quick “AI” summary than explore a page because the retrieval of information has become such a pain in recent years. It seems like every article has to be fluffed up with hundreds of tangential keywords just to land anywhere near the top of classic search results.
Looking something up has become a chore. In a forest of captchas, cookie banners, special promotions, paywalls, and other ads, who wouldn’t prefer a simple, short and sweet article without distractions?
Yes, there’s a reader mode, but it often doesn’t work on these websites with messy source code and I doubt many average users even know of its existence.
And yes, we all know that it won’t be long before the neat “AI” summaries will be riddled with ads and banners as well. It’s a trap we keep walking into every single time.
It doesn’t matter that LLM results can’t be trusted when the alternative is so much worse, right?
Well, of course it matters to me, but I’m part of a minority which enjoys personal websites, indie blogs and understands some of what’s happening behind the scenes of these models.
The Dia Controversy
At the end of last month, the Browser Company released a post on why their beloved product Arc is dead, and they’re going full steam ahead on their newest LLM-first product, Dia. In this article, it’s stated that people don’t need websites any more. To quote directly:
Webpages won’t be the primary interface anymore. Traditional browsers were built to load webpages. But increasingly, webpages — apps, articles, and files — will become tool calls with AI chat interfaces.
A statement I couldn’t disagree with more—but I have to admit that I think it will be true for some users. It’s a realisation I came to while discussing the aforementioned post on Mastodon.
I feel like there’s two principal reasons why people go to the web:
- To find and aggregate information to further a goal
- To have an experience
That first use-case is where I can see a place for LLMs and “AI” summaries, once they’re truly good and reliable enough. Perhaps even ethical, although I hold my horses on that ever happening.
But to have an experience, to get lost in a rabbit hole of information you didn’t even think you were looking for, to get highly individual perspectives or experience content in the context it was published in, we very much still need websites.
Well-built websites which don’t just try to rank high to earn money from ads, which aren’t built to sell something or to grab traffic. Websites which provide an experience. Websites, which have an identity, are individual and independent. Websites you want to visit.
I know such sites and I treasure them. I hope, you do, too–because it is sites like these that are the future of the web.