One of Reddit’s co-founders has come out and said what most of us have been thinking for a while now: much of the internet is dead; the bots have taken over and now outnumber actual humans.
“I think we’ll see a next generation of social media emerge that’s verifiably human, because it’s all going down in the group chats now,” he said.
TBPN
Bots have always been an issue online, but AI has made the situation 10x worse. You can now deep fake anyone, including their voice, with nothing more than a few prompts and a basic understanding of video editing.
Michio Kaku Deep-Fake Video On YouTube
We now have deep-fake videos of Michio Kaku on YouTube spouting nonsense about the 3I Atlas comet (or whatever it actually is). I saw this posted on Reddit earlier this morning, but I’d actually seen the video in question myself inside my YouTube feed.
To the untrained eye, you would think it was Michio Kaku. The video is well done, looks legit, and the subject matter is well within Michio Kaku’s usual wheelhouse.
His voice is near perfect too. It sounds just like him and has the same cadence. If you didn’t know deep fakes were possible and easy to do, you’d just assume it was him.
Which brings you to the next logical question: why are tech companies developing this technology when they know for a fact it’ll mostly be used by bad actors to spam the internet?
Money, of course. But the idea that no regulatory body has stepped in to maybe pump the brakes is very telling about just how much power Wall Street and “The Market” have in most Western countries.
Drowning In AI Slop Content
AI content is everywhere now. Most of the articles you read online haven’t even been proofread by a human. Tools exist now that will crap out an unlimited number of articles per month for $40 to $50.
Google has done precisely zip to combat this in search. Yes, it’s issued a few algorithm updates, but these have done nothing to fix the problem.
It would appear that Google is now more interested in providing tools that expedite the enshitification of the open web than in preserving its quality.
The end goal of all this would appear to be to make the web so unusable, so chock-full of nonsense, that people simply exist in a perpetual state of confusion, fearful of being deep-faked by jilted-ex partners or enemies.
But when the web is ruined (and LLMs are trained on open web content, illegally in most cases, I might add), there’ll be no more sausage left for the meat grinder. Instead, it’ll be AI content feeding LLMs, and that’s when you get model collapse.
Just as a photocopy of a photocopy lacks the quality of the original, LLMs trained on LLM-generated content devolve and become all but useless and, ironically, even less reliable than they are today because they’re being fed their own derivative drivel.
The only shining light in this entire debacle is that human-crafted content and human-fronted brands that focus on authentic content and connections with their audience will be in higher demand.
Despite what anyone says about AI’s role in the future of the world, its ever-increasing presence online and the general disdain from the public mean there has never been a better time to be a content creator.
The future belongs to creators, not fakers.
Sharing is Caring!
