Fascism exists always upstream, with the delusional control of truth values. Only until inevitability reality comes crashing down.

Taking cue from Trump fascist administration, having seen how they can easily manipulate and gaslight a very obvious murder into “self-defense”, Larian decided they could try gaslighting their fans into believing they “listened” to protests and corrected their course, while doubling down on their AI use.

And it’s done through the well trodden path of hiding their tracks.

Well, that’s great, right? There’s absolutely nothing to criticize here.

Problem is the message doesn’t stop there, despite many eyes will glaze over what comes next. And the message proceeds to radically UNDO and contradict everything it claimed in that first part:

Before you can even get to the stage of figuring out if AI can help or not, because, FOR SURE THEFT ALWAYS HELPS, maybe you should figure out what the fuck you’re actually doing and using, right? Just because you “try things” in a dumb, clueless way doesn’t absolve you from responsibilities.

Illegally appropriating the job done by someone else is surely a safe and proven way of “doing things faster.” It just works! I don’t even know why you want a debate on the usefulness of theft. All tech companies have figured out that theft is a successful model, the fastest path to getting rich, especially when governments are now completely corrupt.

So now you are using AI widely “across departments”, having decided that people’s criticism was ONLY ABOUT THE ART. So you figured out that EVERYTHING CAN BE STOLEN, just as long you don’t hurt some artistic sensibilities. Everything else must be up for grabs, right? Steal as fast as possible, loot this miserable world until it’s all gone.

Because that’s the point: everyone is doing the same. Everyone is participating in this looting, so who’s innocent? And if no one’s innocent, how can anyone be guilty?

We have successfully isolated this problem of “artistic origin”, then, having cleansed every other form of theft. So now lets cleanse the artistic side as well, so that WE CAN STEAL WHATEVER WE WANT?

“Without being 100% sure about the origins of the training data.” 100%! Do you know math? 100% is all there is. This is not a tactical game with 110% chance to hit, the d100 only goes 0-99, or 1-100. Statistics! Larian makes RPGs, after all.

“Then it’ll be trained on data we own.”

Oh, nice. So, if this wasn’t just a factually false statement, the only way to make this part “true” would indicate that Larian is now in the AI business and is developing, in-house, their own proprietary LLM model. They would be setting up their own server farm (good luck buying all that RAM required, considering you are trying to set your foot in what’s actually an oligarchy) and hire engineers to build an AI model from scratch. Because, technically, no other way exists.

Even assuming this obviously absurd scenario, there’s no way Larian could do that, because when we speak of LLMs (we’re really barely past kindergarten level here), we speak of something that SOLELY operates through “large amounts of data.” Where large is euphemisms for something HUGE. LLMs are heuristics (much like the human brain), they are statistical models. No matter how perfected is the algorithm upstream, the output can only be remotely useful when INCLUDING THE TRAINING DATA.

Unless Larian is developing LLM from scratch, they DON’T OWN THE DATA. Because even if they fed to it all they produced through their own existence, for another HUNDRED YEARS, that would probably still be less than 2% of the amount of data necessary for anything remotely useful. Without even counting how much work and time and resources would require performing and fine-tuning that training. We are speaking of highly specialized work that ALSO has a cost.

It’s not a case that all those proprietary LLMs have been built coupled with “large amount of data” that was factually stolen. Or “legally” stolen because, again, when corruption runs rampant then you can just bend the rules. No company has enough data to feed any LLM. Even if completely stripping their own employees of any right, that still wouldn’t be enough to train an LLM. You need MORE. You need access to everything.

But here we have Larian, the ingenue. They only use a piece of technology they do not understand. They merely want to make a product faster to sell, as it is their capitalistic purpose. Chop, chop. Move on. If they simply, innocently use a tool, they DO NOT SEE anything the LLM has eaten and digested. They do not see the stolen data. They do not see the stolen art. They just FEED THEIR OWN, to this cancer spreading and devouring more. It’s just “their own end” of the bargain. They just use “the prompt.” Once a body has been masticated, it ceases to be a body. It’s blood and gore now.

So, let’s cover ourselves in blood and gore, to participate in this ritualistic sacrifice that is humanity.

Larian saying that they’ll use LLM on “data they own” is an euphemism saying they intend to continue stealing precisely has they’ve been doing. Nothing changes because as long they can persuade their fans that LLM “start where you see them”, with their prompt, and their actual process is occluded and hidden away from awareness, then… everything can be legally stolen because it can’t be tracked. And they intend to continue to participate in that vicious looting.

A note on the previous.

Twin Peaks was in some way a mythology of meaning. In the third season it blended the origin of evil with nuclear power, then coming after that a godlike presence through the electrical network.

Generative AI is going down a similar path, it’s becoming god-like in its symbolic meaning: a false idol, for a false people.

In the previous, again, my point was to emphasize how “all the AI does” is the magical trick of hiding some process. It’s hiding the money laundering, it’s hiding its costs, and it’s hiding the causality chain, the link between input and output. Which is ironic because that’s exactly what human consciousness is: a trick. A perceptive illusion.

Everyone is looking to that one magical and mysterious ingredient that will make AI truly “alive”, human-like. But it all becomes very clear, dismissing this mysterious fog, after you realize that it’s all about something you need to remove from the recipe. Because if you want to reproduce the ILLUSION of consciousness, then you need to understand the TRICK behind that illusion. Only then you can make it appear the way you want. The way everyone of us sees it.

And it’s all back to the usual. Bouncing between, the description of what we see and feel, and how to build artificial consciousness.

Through that clarity you can then observe.

Of course AIs cannot apologize, because as the scam is all about money laundering and corruption and the erasure of cause and effect, the same happens with the concept of “agency” and responsibility. When it comes to the theft of intellectual property AI companies are free of blame because “the machine did it”, whereas the process can’t be tracked. No one takes responsibility, you can use what the AI does because it’s “felt” as the machine produced it. So you can now own it and sell it again, having “laundered” the process. You stole, cleaned it up, now it’s yours.

The scam is all about hiding those tracks, which is why no one is responsible if AIs cause damage. Responsibility is simply redirected where it is the most convenient. “The machine did it.”

Everything about AIs is about erasure of some kind. Manipulation.

But the irony of all this is that we are looking into a mirror. The AI is false because we are all false. We are finding a perfect match, for our own final and perfect annihilation.

It just takes the easiest final step, to isolate the AI into a fully sealed black box to reach that mythical “superintelligence.” Because that’s the magical difference: we are blind to our own processes. “Not being privy to information about our own training” is what makes human consciousness possible.