← Back to context

Comment by SoftTalker

11 hours ago

The web as we know it is over.

Websites will die on the vine if LLMs intermediate all the content.

The "website" of the future will be an API optimized for LLM crawlers, serving plain-text content that no end-user will ever view directly. The SEO game will change to LLMAO.

17 comments

SoftTalker

Reply
BrunoBernardino  10 hours ago

Alternatively, we can collectively "fight back" by not using Google and teaching others around us to do so as well. There are plenty of decent [1] and great (better) alternatives, where you're not the product [2]!

[1]: https://alternativeto.net/software/google-search/?license=fr...

[2]: https://alternativeto.net/software/google-search/?license=co...

ben_w  10 hours ago

> Websites will die on the vine if LLMs intermediate all the content.

The current zeitgeist of them will, but I think not all.

My first website (GeoCities) was either before Google existed or very close to it. Connected to people via WebRings and directory listings. More recently, RSS feeds.

  • SoftTalker  10 hours ago

    Yeah there will likely continue be a small underground of old-style websites I guess. But you'll have to be in the loop on how to find them, and very few people will pay to advertise on them.

    • mrec  10 hours ago

      > very few people will pay to advertize on them

      That sounds like an unalloyed plus. The perverse incentives caused by advertising have been the biggest driver of the web's decline, IMO.

abirch  10 hours ago

Here is what I think the future web may look like:

   1) Sites will have mcp / APIs for LLMs. So that when I ask my AI Agent du jour. It can call any of the sites where I have subscriptions for information. 
   2) Sites that are passion projects will be harvested by our LLM overlords.
   3) Sites that people don't type into their web browser and need ad revenue will die.
   4) SEO will finally die.

thisisit  10 hours ago

Or more likely move towards substack or newsletters where the pitch is - Don’t let the LLM chose the output for you, go directly to our Substack/newsletter instead.

This will happen especially with things like conspiracy theories because the choice might be to pollute the output or share the general consensus. Like searches for Apollo landing conspiracy theories can either chose to present “alternate facts” so that people can “do their own research” and conclude it is fake or LLM auto corrects to “Apollo landing happened”.

  • rchaud  6 hours ago

    Newsletters have a webview fallback with a public URL that makes them just as susceptible to scraping. If that ever gets fixed, Google will just scrape the full-text content in Gmail instead.

    Newsletters have been around forever and never taken off like the open web and free blogging have. Slapping a Stripe integration on the backend hasn't led to Substack becoming a sustainable business not propped up by VC cash.