Post

Indie devs are in a tough spot: don't use AI, and watch everyone else ship games x5 faster than you or use it and be damned. I don't think there are many coders left that are not using AI agents for coding right now, it's just that effective (been coding games for 28 years, professionally for 20)

This post is unavailable.

2 reposts26 likes

By the way I'm not pro-AI, I wish we could put the genie back in the bottle, but it's real and it's a massive change on how we do things and sticking your head in the sand is not going to do much for you, that's my entire take

Some people describe it as the same that happened with moving from asm to modern programming languages where the compiler does a lot for you, but it's definitely more removed than that

Right. Moving from Assembly to high-level languages is still fundamentally programming. Directing an LLM isn't. With the former, the code you produce is still yours, and the claim that you-yourself wrote it holds up.

I consider writing code absolutely the least important part of my job. Architecture and interface definitions are 100 times more important, and I'm much more focused on that now than before

I don't know. You don't improve at something by not doing it. And I'm curious where we'll be in a decade if people continue to not write code. Where we'll be after onboarding a generation of new developers who haven't written any meaningful code in their life. Who aren't able to.

Yeah that's a big worry, especially for people just learning computer science now, I genuinely have no answer for that, I also strongly believe that it should free us to do more creative stuff, not do that for us, but it seems we'll just collectively drown in slop...

Honestly this is the most reasonable take. Yeah, lots of shitty uses of AI but there are a few exceptions and honestly I don’t blame small devs for at least dipping their toes. It’s the Squares of the world I am worried about. The opportunity to exploit workers and abuse a system is too high.

But its a massive change to everything it touches. We don't know how things are going to develop. Saying "I don't want this and am going to do what I can to navigate around it where possoble" is actually a very reasonable take.

I'm not against AI for programming, but the idea that people who don't use AI to make their games will be left behind is nonsensical. The games I most care about aren't interchangeable, so a game made more efficiently won't make it more appealing. More power to devs who can make great things w/o AI!

How do you think the bubble popping and all those retirees losing their entire savings due to index funds having gone hard for the AI company IPOs is going to impact what a toxic fucking brand AI already is for consumers, and thus its impact on your future sales if you use it?

I never have and never will use AI for coding, I don't need or want it in any way. Thankfully it's banned at work, and when I write code at home I want code that I understand and know. There is a fallacy that most coders use it or want it, where I see the opposite.

Your worries are unfounded. The people relying too heavily may ship faster, but they won't be hand-constructed passion projects from the ground up. I don't need games faster, just quality. A few months after relying on AI for coding, you'll learn you don't know how to code...

And then you're stuck paying for a bullshit "AI Agent" so long as you want to keep making games, no matter how expensive they make it. If anything, I think this only highlights how important it is not to allow the skill to be lost. Let the slop makers make their slop, I won't be buying it.

We do not use for it in our in-house projects. I don't feel ambivalent about this at all -- I enjoy programming. If I wanted to make unappealing compromises, I wouldn't be an indie.

I'd settle for a keeping everyone safe and avoid crunch award, not saying that AI is needed for that, but probably that's why the people at Spryfox are using it, not going to condemn them for trying out ways to make work easier if it's not slop impacting the game's quality or gen ai art

I think the word "ship" bears examination. Everyone in the genAI space talks about "shipping" and "velocity" and I keep wondering what the framework of that conversation actually is. "Shipping a repo", what does that mean? ? It's decent for making internal tools but that's not "shippable" to me.

The single biggest concern I have is simply the supplicant nature of the thing. I don't pay a subscription fee to my compiler, my IDE, my debugger. Actually, none of my tools, and none of them change perceivably over time without my say so. genAI is like fostering an addiction.

Yeah you nailed my main concerns too, I don't even pay a subscription for Adobe products myself, I have an old Paint shop pro cdrom and clipstudio paint 😅 The whole token thing invites a "use it but not too much" approach that's just doesn't work too, local models are going to be more interesting

I agree it's a very tough arena to be dogmatic in right now, there is an industrial reality to things. But to me this is less a choice between velocity or "falling behind" (btw it's not a race, no?), it's a choice between "I like this world" and "fuck this shit forever"

This is all that people never used and don't know these systems have to say, it doesn't work that way, if you think all it does is scan the web and copy stuff you are talking with no idea of reality

"After completing the study, developers estimate that allowing AI reduced completion time by 20%. Surprisingly, we find that allowing AI actually increases completion time by 19%--AI tooling slowed developers down." arxiv.org/abs/2507.09089 RIP to your grandma but I'm different

arxiv.orgMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer ProductivityDespite widespread adoption, the impact of AI tools on software development in the wild remains understudied. We conduct a randomized controlled trial (RCT) to understand how AI tools at the February-...

I've seen people SAY they used some tool layer on top of it to spit out text-edited variations of other people's games and fuck that type "edit gorillas.bas" like the rest of us.

It may be the reality that it is faster, but ethically GenAI is grey at best. What I see isn't so much knee-jerk reactions against AI, but more the "complying in advance" or those saying the ethics don't matter or ignoring ethics entirely that really piss people off. Also, famously, faster ≠ better.

So, really, what it all comes down to is, simply: Capitalism is, by design, unethical. Capitalism dictates AI is good because it makes an adequate product, at the cost of humanity. That human cost is what most people are not okay with. But, capitalism, as a feature, doesn't care about humanity.

Cheers! I have several reasons that I don't think using AI is a good idea (beyond ethical reasons), but one that's maybe not obvious is that I'm a game designer whose medium is code: for me the act of writing code is inseparable from thinking about the game in detail, deciding what to do next, and..

..having a good understanding of the technical, design, production implications of those decisions. Abdicating responsibility for the code to an AI agent would give me a much blurrier view of where things are at.

Same here. Been programming since forever, over a decade in games. I see absolutely no need for it - creating worse code faster is absolutely not my goal.

My local bubble of London friends is mostly people in AAA, I'm now one of the few indie devs I know, my circle is a mix of people going from "I hate it but they're making me use it" to "I hate it but I must admit it works" to "I'm never going back to code without it"

My circle is UK/EU AAs and indies and there are some who use it, but they're the minority. "Everybody else uses it to ship games 5x faster" is also quite factitious I'd argue - if that was the case we'd see AAAs being released left right and centre (maybe they're "just around the corner" but 🤷‍♂️)

One of the "not many" checking in. Forced to use it at work, where it takes all joy out of it and makes my work faster, but I can rarely answer any questions about a feature a week later, even though I test and manually vet the output before committing. Never going to use it for any of my games.

Yep that's my feeling about it too, and the same of most of my industry friends (with some outliers that I guess just don't want to codemuch anymore or are clicking with it more), at the moment it's a mix of FOMO and real improvements but difficult to quantify

Oh the FOMO is INTENSE in software dev (I don't work in games), though I'm not quite getting it: If the tech makes everything so easy, then I could learn at any point and not be at a disadvantage. Business folks love that shit because it's a temporary speed boost with unknown future consequences...

And if I could afford to quit over this, I _would_. AI fosters resentment in many of your employees, be aware of that. This is not the job I once loved, and forcing this shit upon me certainly hasn't helped.

Honestly, I'm very skeptical of those worried about shipping games faster Just because they're faster doesn't mean it's more efficient or even secure, especially if online features are involved And, imho, coding, like any other game-related discipline, has to take time to be done properly