Tag: dead internet

#oneaday Day 838: A shitty new trend

One of the most irritating new developments in the world of generative AI is the use of "agents" to spam you. This can take numerous forms; earlier today, Luke Plunkett of Aftermath shared an AI "pitch" he had received via email, Frank Cifaldi of the Video Game History Foundation shared something similar, and of late I've seen a marked uptick in blog comments that are obviously (but, of course, undisclosed) AI-generated spam.

The reason I can tell they are AI-generated spam is because they don't come from a person. Inevitably, they come from a "company", with the email address and website you can optionally include when leaving a comment linking to some sketchy website or other that wants you to download their new AI-powered tool to, I don't know, delete all the shit on your hard drive or some such shit.

The reason why these are infinitely more frustrating than good old-fashioned spam (and man, at this point I really miss the surreal narrative spam from 2009) is because, using the magic of AI, they might initially appear to be a genuine comment. In other words, the "agent" (spit) has taken a blog post as input, been tasked with composing a comment that might initially appear to be engaging directly with the subject matter, but which is actually just stating what you wrote back to you with slightly different wording, then attached a sketchy link and email address to it. The hope, presumably, is twofold: firstly, that the blog's owner will fall for the spam by clicking on the link or sending a message to the email address; and secondly, that the blog's owner will publish the comment, making the link public and thereby, potentially, tricking others into inadvertently clicking on it.

This is fucking scummy behaviour! Moreover, this is not convincing me that "agents" are in any way a net good for humanity! Not to the absurd degree that I think they're going to "destroy humanity" or any of that utter nonsense, but slowly but surely, generative AI in general is making the Internet harder and harder to use effectively — and, moreover, it's destroying our trust in one another.

Earlier today on Bluesky, I saw someone share a post from an "Atari news" account that looked interesting. Supposedly there's a project out there that is attempting (and starting to succeed) to get Sierra's Adventure Game Interpreter (AGI, but not that one) games working on Atari 8-bit. They've supposedly already got King's Quest up and running, and the parallel C64 project is, it seems, a little further along.

However, I was immediately inclined to distrust everything about this news, firstly because of the amount of vibecoded bullshit that is out there at the moment — and apologies to the creators of the project in question if it has all been done by hand; you might want to put that very fucking prominently on your website — and secondly, because the Atari news Bluesky account that shared it in the first place is absolutely fucking infested with garbage AI-generated artwork, leading me to believe that the account is automated. An automated news-aggregating account isn't necessarily a bad thing in its own right, but when combined with the amount of AI imagery that account shares, every single one of my bullshit sensors starts going off.

Want another example? Time Extension, the retro gaming website, has recently been dealing with AI-powered plagiarism and harassment from an account calling itself "Generation Amiga", ostensibly a news site about… well, the Amiga. And yet if you ever accidentally have the misfortune to end up on Generation Amiga's website, you will discover that not one of the articles there (very few of which are actually Amiga-related) is written by a person. Their "exclusive reviews" consist of LLM-generated slop derived from marketing material. And they even had the gall to accuse Time Extension of being "Big Media" harassing a poor, innocent little independent creator who was just using an ickle-wickle bit of planet-destroying AI to construct a vibecoded website that has a net negative value to humanity — all because Time Extension's editor, Damien McFerran, quite rightly called this absolute bullshit out, for the benefit of anyone in retro gaming who had never had the misfortune of stumbling across these fucknuts.

You know, contrary to how it might appear, I don't like ranting and raving and complaining about this all the time! (All right, maybe a little bit.) I would much rather live in a world where it wasn't happening at all! But when I'm in a situation where I can no longer trust that comments on my fucking personal blog are written by actual people, I think I am more than justified in losing my patience. When I am in a situation where my first response to a potentially cool piece of retro gaming-related news is not "wow, that's cool!" but instead "is this AI bullshit…?" I think I am very fucking justified in losing my patience.

I'm so, so tired of this all. None of this is making the Internet better. None of this is making the world better. All it's doing is, at best, sowing division and mistrust, and at worst, completely obliterating what was once the world's most valuable repository of information, knowledge and human creativity. That's the "AIpocalypse" we should be concerned about.

Want to read my thoughts on various video games, visual novels and other popular culture things? Stop by MoeGamer.net, my site for all things fun where I am generally a lot more cheerful. And if you fancy watching some vids on classic games, drop by my YouTube channel.

If you want this nonsense in your inbox every day, please feel free to subscribe via email. Your email address won't be used for anything else.

#oneaday Day 793: The Web is so broken even offline copies of pages are fucked up

Hey! Did you know the Internet is so broken that even a locally saved copy of something can be unreadable without a bunch of annoying "tinkering"? 'Cause it fuckin' is.

Prior to USgamer's closure in 2020, I backed up most of my articles that I could find and/or remember. I saved them as local copies on my computer, with the intention of gradually republishing at least some of them on MoeGamer — a process I started but still need to get back to. (Some of these articles found their way over to VG247, but they're buried in an archive and not exactly what I'd call "easily accessible" today, so I have absolutely no qualms in taking these pieces and republishing them myself.)

Anyway, after my enforced reinstall of Windows a little while ago, it suddenly occurred to me that I might have lost these locally saved archives. Thankfully, I had apparently had the foresight to back them up to my Google Drive (I know I should get off Google, I know) and thus I was able to recover them. After extracting them to my Documents folder, I thought I'd have a browse through a few to remind myself of what I'd written back then.

The first one I attempted to open — a review of Imageepoch's Time and Eternity, a game which I liked but none of the rest of the Internet did — opened for a split-second before automatically redirecting to some website I'd never heard of: "spl.zeotap.com", which is apparently an "AI-powered Customer Data Platform that helps enterprises unlock the true value of their customer data". Of course it's a fucking AI thing now — though I don't know what it was back in 2020 when I archived these pages. Whatever it was, it is broken, because every time I open certain articles from my locally saved archive, they immediately redirect to part of "ZeoTap"'s website that no longer exists, with no way of telling the browser to go "Back" and just look at the fucking page I was trying to look at in the first place.

Thankfully, this isn't the case for everything in the archive — and I've determined that you can tell which ones will fail by the file size of the main article HTML file — and I also found a workaround using Firefox's Developer Tools. Unfortunately it's not a workaround I can just apply globally; I have to do it every time I open one of these bastard articles. So for the sake of my future self, I'm going to recount the process here, just in case:

  • Open a new tab
  • Open the Developer Tools console
  • Go to the Network tab of Developer Tools
  • Drag the problematic HTML file into the browser window
  • Scroll through the stuff in the Network tab that appears until you find the problematic ZeoTap URL request
  • Right click on it and choose "Block URL"
  • Drag the HTML file into the browser window again
  • The article should now appear, and you can close Developer Tools to read it
  • You have to repeat this process each time you open a new tab or browser window

This is fucking annoying though, isn't it? I would have thought that the entire point of saving a web page locally would be so that you could, you know, look at it locally without having to be online — or, indeed, dependent on URLs that no longer resolve correctly. But apparently at some point in USgamer's history, certain pages on its site became so dependent on whatever the fuck this ZeoTap thing is that it became impossible to just look at the damned pages. And that dependency is somehow maintained when you save the page for offline viewing.

Well, you might think, surely it's just a matter of looking at the page source, searching for that troublesome URL and removing the bit of the HTML file that references it? You'd think, wouldn't you?! But there are so many fucking Javascript files, so many CSS files, so many external HTML files that are loaded when just looking at one sodding page, which, I'll remind you, is saved locally, that it's impossible to track down which one of these myriad dependencies is causing the problem!

I remember a while back, someone posted on Bluesky that they looked at a page on somewhere like PC Gamer, and before the article had appeared the browser had downloaded something like 20MB of… stuff to their computer. I think back to the good old, bad old days of dial-up, when downloading a single megabyte of anything took upwards of an hour at the best of times, and wonder exactly how we let the Web become such an unwieldy mess.

(Ads and tracking is the answer, of course. But you knew that.)

Anyway, thankfully I have at least found a workaround that appears to work for the articles that are affected, so I guess I will prioritise getting those pieces onto MoeGamer sooner rather than later!

Want to read my thoughts on various video games, visual novels and other popular culture things? Stop by MoeGamer.net, my site for all things fun where I am generally a lot more cheerful. And if you fancy watching some vids on classic games, drop by my YouTube channel.

If you want this nonsense in your inbox every day, please feel free to subscribe via email. Your email address won't be used for anything else.