“Normal People Don’t Care”

There is a minor, ongoing media kerfuffle with the internet-darling Larian Studios (makers of Baldur’s Gate 3, Original Sin 2, etc). It started with this Bloomberg article, wherein Jason Schreier writes:

Under Vincke, Larian has been pushing hard on generative AI, although the CEO says the technology hasn’t led to big gains in efficiency. He says there won’t be any AI-generated content in Divinity — “everything is human actors; we’re writing everything ourselves” — but the creators often use AI tools to explore ideas, flesh out PowerPoint presentations, develop concept art and write placeholder text.

The use of generative AI has led to some pushback at Larian, “but I think at this point everyone at the company is more or less OK with the way we’re using it,” Vincke said.

There are possible charitable and a not-so-chartable takes on those words, and suffice it to say, many people chose the latter. Vincke responded with a “Holy fuck guys [chill out]” Twitter response, with clarifications and emphasis that they only use AI for reference material and other boring things, and not with actual content. Jason Schreier also chimed in with an original transcript of the interview, as a response to others suggesting that what Schreier wrote was itself misleading.

As a side note, this portion of the transcript was extra interesting to me:

JS: It doesn’t seem like it’s causing more efficiency, so why use it?

SV: This is a tech driven industry, so you try stuff. You can’t afford not to try things because if somebody finds the golden egg and you’re not using it, you’re dead in this industry.

I suppose I should take Vincke’s word on the matter, considering how he released a critically-acclaimed game that sold 20 million copies, and I have… not. But, dead? Larian Studios has over 500 employees at this point, so things are likely different at these larger scales. I’m just saying the folks that made, you know, Silksong or Megabonk are probably going to be fine without pushing AI into their processes.

Anyway, all of that is actually a preamble to what sent me to this keyboard in the first place. In the Reddit comments of the second Schreier piece, this exchange took place:

TheBlightDoc: How could he NOT realize how controversial the genAI comments would be? Has he been living under a rock? Or does he himself believe AI is not a big deal? :laughing:

SexyJazzCat: The strong anti AI sentiment is a very chronically online thing. Normal people don’t actually care.

do not engage… do not engage… do not engage

Guys, it’s hard out here in 2025. And I’m kinda all done. Tapped out. Because SexyJazzCat is correct.

Normal people don’t actually care. We know this because “normal” people voted the current administration back into office. Normal people don’t understand that measles can reset your immune system, erasing all your hard-fought natural immunities. Normal people don’t understand that every AI data center that springs up in your area is subsidized by increases to your own electric bill. Normal people don’t understand that tariffs are taxes that they end up paying for. Normal people don’t understand that even if they didn’t use ACA subsidies, their health insurance is going to wildly increase anyway because hospitals won’t be reimbursed for emergency care from newly uninsured people. Nevermind the, you know, general human misery this creates.

Normal people don’t actually care about AI. But they should. Or perhaps should have, past tense, because we’re far past the end of a very slippery slope and fully airborne. Normal people are just going to be confused as to why computers, phones, and/or videogame consoles are wildly more expensive in 2026 (e.g. RAM crisis). Or if AI successfully demonstrates real efficiency gains, surprised when they are out of a job. Or if AI crashes and burns, why they also still lost their job and their 401k cratered (e.g. 40% of S&P 500 value is in AI companies).

The only thing that I still wish for these days, is this: people have the kind of day they voted for.

Posted on December 18, 2025, in Commentary and tagged , , , , , . Bookmark the permalink. 4 Comments.

  1. I’m not sure the ‘normal people don’t care about AI’ thing holds up for gaming, as gamers and ‘normal people’ aren’t the same. There have already been examples of companies caught using AI and having to address the issue and fix the ‘mistake’. I 100% believe a bad review score on Steam matters in terms of sales, and more than one game now has been bombed because of AI use.

    It’s not 100% “use AI and fail”, but it’s not 0% either. And maybe it swings more towards ‘people don’t care’, but right now it feels like things are swinging further towards “Do care and don’t like it”, for a mix of all the reasons you listed above.

    Liked by 1 person

  2. “Gamers” seem to be very tribal. What they will or won’t accept seems to have very little to do with logic. At the moment, even the rumor that it’s been used leaves a taint. How long that attitude will persist, though, is another matter. Gamers also have very short attention spans when it comes to things they object to and very little willingness to disadvantage themselves for moral or ethical rectitude, as we saw throughout the Blizzard scandal.

    At the moment, AI looks quite shoddy in most situations so it’s very easy to reject it. I wonder how that will hold up if some games come out in which the AI looks and plays very well, though. Or even if it looks and plays better than human-made equivalents? Always assuming anything like that ever happens. Doesn’t look likely at the moment but it’s early days still.

    Like

    • […] as we saw throughout the Blizzard scandal.

      Which one? ;)

      As for AI improving, there’s going to be a very interesting inflection point beyond which you have to ask the devs: “what is it that you do here?” Usage today seems to focus on promotional material (the worst possible place to get it wrong), placeholder assets, and I guess pre-concept art per Larian. That said, there doesn’t seem to be much distance between pre-concept art and regular-ass concept art, which then zips right into art art and game assets. Low-budget indie games already use Unity/UnReal assets all the time. It is easy to imagine AI improving/proliferating so many higher-quality assets in the future, such that the line between stock and bespoke becomes harder to determine. I mean, hell, we’re already there in terms of paint-on-canvas kind of art.

      Vibe coding is already a thing too. It’s hard to see someone using AI to completely craft a viable game engine, but augment an existing one, which is done all the time?

      So, sure, let’s imagine the near-future in which AI improves to the point that it becomes almost impossible to tell that it’s AI. What are the devs doing that you cannot (other than, presumably, having the $1000/month AI subscription)? Or the ten thousand wannabe-devs that can use the exact same tools to flood the market with very similar offerings?

      We live in interesting times.

      Like

  1. Pingback: Blurry Lines | In An Age

Leave a comment