Yeah the AI slop hell that gives us terrible slop like expedition 33. Shudder at the thought


"I will be investigating this shortly." That way, you don't have to commit to AI and can distance a bit from the micromanagement. If he persists. "I have a number of avenues I'd like to go down and will update on progress tomorrow". Though I'd be tempted to flippant, "if you're feeling confident to pick it up, I'm happy to review it". If they hesitate, " that's OK, I'll go through the process. Standups should be quick. Any progress, any issues, what you're focussing on. Otherwise you waste everyone's time. Any messages I'll ignore until I have 5 mins. Micromanagement environments are not worth it.

Code analysis and suggestion tools in many professional IDEs are not powered by LLMs, in the IDEs I use, there's an available LLM that I've disabled the plugin for (and never paid for so it did nothing anyways). LLMs are simply too slow for the kind of code completion and recommendation algorithms used by IDEs and so using them is *not* "using genAI"

I'm not going to fault someone for driving to work in a car, but I certainly wouldn't call them the winner of a marathon even if they only drove for a few minutes of that marathon. There's a difference between something that runs the race for you (LLM AI) and something that simply helps you do what you are already doing (I suppose photoshop is the equivalent of drinking gatorade).

If there was an AI that licensed every bit of art/code/etc that it trained on, then I think I would be fine if they used it. BUT, I’d never think their final product was ever a more than just a madlibs of other people’s work, cobbled together for cheap commercial consumption. My time is worth something, and I’m not spending a minute of it on AI generated crap when I could be spending it on the product of a true author, artist, coder, craftsman. They deserve my dollar, not the AI company and their ai-using middle man who produced shit with it.

I don't think that's a relevant comparison - marathon is a race meant specifically to test what the human body is capable of. Using a car there is obviously against the goal of the competition. When I'm writing code, I'll happily offload the boring parts to AI. There's only so many times you can solve the same problem without it being boring. And I've been doing this long enough that actually new problems I haven't solved yet are pretty rare.

> I’m all for laws to regulate data centers and manufacturing, but again, that’s not what is being pushed for. Most anti-AI peeps seem the be helping the enemy a lot more then they realize. I'm guessing there's a lot of controlled opposition which is incredibly cheap to produce, doesn't leave much of a paper trail, and is reasonably effective.

I'm not going to fault you for that - but do you think you should receive an award for the work you didn't do? In the case of this particular game, perhaps the bulk of the creative work was done by humans. But if the GOTY committee isn't confident with drawing a line between what work is OK to offload onto an AI and what work isn't, then I think it's fair for then to say this year that any generative AI use is a disqualifier. You and I can say with ease that an implementation of a basic swap function (c=a, a=b, b=c) doesn't require any creative work and has been done to death, so there's no shame in copypasteing something from stackoverflow or chatgpt into your own code to save time. But it's harder to gauge that for more complex things. Especially with art - where would you draw the line? Reference material? Concept art? Background textures or 3d models of basic props (random objects in the scene like chairs, trees, etc)? I don't think there's a clear answer for that. You might have an answer you think is correct, and I might have one as well, but I think it will be difficult and time consuming to achieve consensus in the game development community. So, the most efficient answer for now is to have any generative AI be a disqualifier.

Yes, it is indeed embarrassing that you insist on defending AI. My sister insisted on using an AI this year to generate our secret santa. She didn't get a gift. Ahahahaha. AI is strictly for stupid people that can't do good work on their own. No reasonably intelligent person is using AI to make a draft that they can then correct, because it will always be more practical to just make it yourself.

Well, you see, the "everyone" you are referring to are the same stupid masses that already don't deserve respect on the macro level. The same stupid masses voting in politicial officials like Donald Trump.

Bullshit. “Everyone” in your comment was a clear “appeal to popularity” argument, made to _other_ anyone not on the bandwagon. It deserves to be called out for what it is.

Uh... Sorry but no, LLMs are definitely fast enough. It works just like auto complete, except sometimes it's a genius that pulls the next few lines you were about to write out of the ether, and sometimes it makes up a library to do something you never asked for Mostly it works about as well as code completion software, but it'll name variables much better

I mean they mostly used it for textures, right? Those are often generated from a combination of noise and photography, it's not like they were building the game out of lego bricks of other people's art. I don't see how it's significantly different than sampling in music, it's just some background detail to enhance the message of the art. Obviously modern AI is a nightmare machine that cannot be justified, but I could imagine valid artistic uses for the hypothetical AI you described.

I believe you that genAI is being used for code suggestions, but I wouldn’t call it a genius. This is anecdotal but over the last couple of years I’ve noticed Visual Studio’s autocomplete went from suggesting exactly what I wanted more often than not to just giving me hot garbage today. Like even when I’m 3-4 lines in to a very obvious repeating pattern it’ll suggest some nonsense variation on it that’s completely useless. Or just straight making up variable and function names that don’t exist, or suggesting inputs to function calls that don’t match the signature. Really basic stuff that any kind of rules-based system should have no problem with.

I really hope their next game smashes it too after this statement. Has potential to pave the way for the rest of the industry.

IDEs use LSP dumbass go setup neovim.

VSCode autocomplete is literal AI now, they don’t have intellisense Others will follow, making “no AI” software really difficult to prove

I'm not surprised they ultimately felt like GenAI isn't useful to what they're trying to do. Game dev has known about this type of generation for a while now (see https://en.wikipedia.org/wiki/Model_synthesis and the sources linked at the bottom) and it takes a lot of human effort to curate both the training data and the model weights to end up with anything that feels new _and_ meaningful. If I shuffle a deck of 52 cards, there is [a high chance](https://math.stackexchange.com/questions/671/when-you-randomly-shuffle-a-deck-of-cards-what-is-the-probability-that-it-is-a) of obtaining a deck order that has never occurred before in human history. Big whoop. GenAI is closer to [sexy dice](https://www.scribd.com/document/605580945/Sexy-Dice-Game-Printables) anyways - the "intelligent work" was making sure the dice faces always make sense when put together and you don't end up rolling "blow suck" or "lips thigh". It's very impressive that we're able to scale this type of apparatus up to plausibly generate meaningful paragraphs, conversations, and programs. It's ridiculous what it cost us to get it this far, and just like sexy dice and card shuffling I fail to see it as capable of replacing human thought or ingenuity, let alone expressing what's "in my head". Not until we can bolt it onto a body that can feel pain and hunger and joy, and we can already make human babies much more efficiently than what it takes to train an LLM from scratch (and _they_ have personhood and thus rights in most societies around the world). Even the people worried about "AI self-improving" to the point it "escapes our control" don't seem to be able to demonstrate that today's AI can do much more than slow us down in the long run; [this study] (https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/) was published over 5 months ago, and they don't seem to have found much [since then](https://metr.org/research/).

You can literally customize the vscode intellisense right now...