NOT INTENDED FOR HUMAN CONSUMPTION

For most people, the internet is not the wires and protocols that tether us all together like an awkward group hug at a funeral, it is the web, – that twitching, blinding maze of pages and pop-ups, in a confetti of cat pictures, clickbait headlines, and the occasional weight loss life hack. And while it was born with sincerity, it soon learned that sincerity is poor for business. What mattered wasn’t what humans wanted to read, but what search engines wanted to index. Search engines – the great librarians of entropy, didn’t write anything themselves. They simply promised that if you wrote something mildly useful and tagged it like graffiti on a cubicle wall, someone might stumble upon it. But as soon as hits became currency, everything changed. Headlines learned to shout, introductions became endurance tests, and paragraphs – entire textual fields of wasteland – were grown solely to feed the monstrous, unblinking eye of keyword density. Meaning was mugged in a back alley by metadata as the pearls of a necklace bounced in slow motion across the floor. But instead of Batman, this arrangement birthed a mutual resentment that somehow functioned. Readers learned to scroll with the cold detachment of someone browsing through a will. Website owners mastered the dark art of pretending to answer a question while circling the drain. And the search engines? They didn’t care, not really, as long as the ads loaded, the apocalypse could wait. However, Large Language Models (LLMs) don’t need links or pages or even your carefully positioned “10 ways to improve X”. They ingest the raw meat of the web and regurgitate gourmet answers without so much as a breadcrumb in return. It’s like asking a butler for your coat and he hands you a tailored suit stitched from the fabric of everyone else’s. From the user’s perspective, this is delightful. From the content creators’, less so. When AI answers questions instantly, everything that once led people to those answers is bypassed. And what then? AI systems require a continuous influx of new human work to remain useful, but if there’s no incentive left in publishing, if the applause has left the auditorium, what madman keeps performing? At first, everything will retreat behind paywalls, newsletters, and private communities. So eventually, AI companies will have to start tipping a few coins into the begging bowls of those they want to pillage. And in much the way that SEO reshaped the web, an LLM tailored version of SEO will surely follow. And of course, this new optimisation layer, will not be intended for human consumption, so content will be engineered specifically for machine ingestion, perhaps resembling SQL, – raw, structured deposits of text and metadata, with no narrative or persuasion, not even a voice. Which, hopefully, will split the web by audience, and become a beautifully grim dystopia, where low-effort slop disappears, leaving behind smaller, quieter, better-lit corners of the web. I am a dreamer, but for now, we’ll just limp forward – disoriented, incoherent, still typing – because inertia is cheaper than innovation. But one thing is clear – there is no going back. Once the public has tasted instant certainty (or at least something that sounds like certainty), it will not gladly return to the scavenger hunt that is page 2 of search results. The next great flood of content won’t be meant to persuade or delight, it won’t even pretend to be readable. It will be useful – and nothing more. For now, I’m hopeful. The web began as a human thing, flawed and chaotic, full of blinking text and auto-playing music in a sea of creativity. Then came search engines, and creativity became commodified, but now, with machines pillaging the web, we might, hopefully, return to creating for the sake of creating, or maybe we’ll start going outside again. Either way, the web will continue – whether we’re invited or not.