Web 2.0 - An Early Obituary
Can we let it go?
It has been predicted that by 2025, 90% of the content on the internet will be machine generated. That's text, images, videos - whatever you like.
We can debate the 90% figure or the date, but I'm not sure it's really worth investing a great deal of time in that. I'll happily state with a relatively high level of confidence that real (as opposed to apparent) AGI and Human Level Artificial Intelligence are much further away. Hyper-functionality will continue to grow, massively outpacing what people are capable of in many technique-driven fields, and people will continue to be bamboozled into believing that the machines that jump through those hoops are human-like too, but we’re a decade or more from flipping the few pages forward in the handbook of fundamental computing needed to get to the totally different architectures we’re going to need if we expect our machines to be like us in the way they reason, while also being hyper-functionally superior in every other possible way.
But anyway…while we're still on this particular page in the handbook, and the source of information to which we have become utterly addicted continues to be the web - where almost all of what’s there has been made by software that derives its inspiration from that very same internet - what sort of a self replicating garbage factory will we be left looking at in 2026?
Although It would be strange to describe the internet at that point as alive, it most certainly will be an organism of sorts. Our efforts will no longer be required to sustain it. Its input will come from its own output, and the tail-swallowing that follows must surely proceed at pace until the whole thing just descends into some sort of cacophonous electronic yell of unverifiability.
A couple of years ago, I was quick to dismiss the potential for what was being described as Web 3.0 to emerge from what we have today. It seemed like a spuriously sprouting teenage offshoot freakshow. A globulous grey version of the dark-web…
The descriptions I read back then were almost entirely focused on a vague anti-establishment doctrine of decentralization and the as-yet unproven promise of blockchain.
The decentralization premise that appeared to drive those waves of change towards the beach of improbability often ignored a few inherent facts of what the web (any version of it) actually is, which is a few trillion dollars worth of prior investment in infrastructure from telcos, big tech and a gazillion other vested interests, all of whom have established their life-sustaining recurring revenue streams on that investment, and will be highly unlikely to relinquished them any time soon.
Abandoning everything we’ve built so far seems unlikely and somewhat wasteful, but supplanting it with something entirely different could be an even greater challenge, requiring at least some pieces of the model to be rebuilt almost from scratch, dumping users and subscribers by the wayside as we go…which is not usually the way you go about making a viable business into an even more viable business (somebody might want to remind Elon of this).
However, there were things about the concept that did make sense.
Decentralization as a means of achieving resilience, segmented privacy and resource balancing could be a good thing. Decentralization so you can stick it to the man, not so much.
The non-repudiation and strong attribution capabilities of the distributed leger model that fell out of cryptocurrencies might be immensely valuable in countering the stormy waters of a web filled with machine generated content and Deep Fakes, in fact this could be the most crucial feature of any future web foundation - even if it is full of machine generated content with no means of controlling the rights to its consumption. As long as it doesn’t turn into a model where truth and trust belong only to the few with the power (both political and electrical) to be a part of the infrastructure.
If there's a way to ensure that encryption remains intact - possibly through this process of decentralization and the hope that quantum crypto crackers remain scarce enough that one won't appear over the horizon of your local domain any time soon - then this just might be okay for a while.
However, if today’s web continues on its current trajectory, the time is rapidly approaching when it will be impossible to get anything coherent done without engaging in some sort of algorithmic double-bluffed guessing game between the generative AIs and the tools you use to write your prompts, where they’re not all mangled and tangled into the same wilderness of digital untruths, conspiracies and the models that are built from them.
For the vast majority of people, the option (and the preference) to just let go and collapse into that vat of endorphin-laced mediocrity is probably going to seem just fine - especially if the real world is never going to be anywhere near as much fun as it once seemed to be. And let’s face it, I like endorphins too; but this can’t be the last stop on the road to internet nirvana, can it?
That figure of 25% machine content by 2025 needs to be unpacked a little bit. It’s not as though all of the current content on the internet is going to disappear and be replaced by machine generated stuff, it’s just that the pace at which machines will be able to generate new content is rising exponentially, and many of the new data centers popping up all over the planet will likely be both responsible for that as well as filled to the brim with it.
Cloud based applications and services that emerged in recent years will not disappear. That toaster you have that tells you when your sourdough is beyond optimal freshness is still going to work fine, it’s just that soon your refrigerator may try to persuade you to shop at a dozen places that don’t actually exist, even though each of them is being endorsed by celebrities and influencers who probably never said what their deep faked avatars might seem to have whispered in your ear.
Nobody will care about some IoT device pinging sewage level data to a cloud based logger for the water company, or a remote camera pushing average wave heights from the beach to your surferbuddy.com account (if there isn’t one of those then maybe there ought to be…for people in flowery shirts). These things could easily be manipulated and exploited too, but that’s maybe not very profitably.
When there’s more to win or lose from algorithmic skew, or when the information needs to be really relied upon - like for your bank account or to keep your lights on, for instance - being able to sort fact from truth is going to be something more of us are going to be concerned about.
This isn’t so much a doom and gloom post as a question.
When will the lack of trust in what we're consuming from the monolithic data-fondu that is today's World Wide Web push us to finally want to abandon it and move on to something completely new?
The whole idea of attempting to somehow watermark machine generated content for easy identification, or have some AI based system point out things that other AI based systems have made is just a non-starter. You remember how hard it was for Deckard to do? And by all accounts he was one of them…
Something more fundamental needs to change in how we attribute information and how we value real facts over fake ones. I don’t know what that thing is, or how we might build structures that aren’t hoodwinked by falsehoods all of the time, but it feels like something significant has to change in the underlying structures of information for that to happen.
Web 3.0 probably won’t bring all of that in one fell swoop. If it comes, it will probably want to still make a few bucks off what came before, but maybe a different thing could one day supplant it all. Maybe Web 4.0…?
Or maybe trust is just an outdated concept from a bygone generation.
Join the conversation and let others in the growing list of subscribers know where you think we’re going. If you’ve got something to say about Security, Cities or Technology, your input is always welcome here at Securiosity.