This is another AI problem for the gamedev industry. Soon anyone will be able to give AI a video clip, a wiki link, a screenshot of a game, along with maybe a brief description and have a nearly identical clone they can play or try and monetize themself.

it's not. if you think that what is on the right is the same as what is on the left, or there's a very minor difference, i got a bridge to sell

So then stop building games with ai that take one prompt and start building games that take 6 to 12 months WITH ai automating everything.

That is what I expect. We won't feel the results immediately, but as AI models like this proliferate we are going to see downward pricing pressure as the market gets flooded and development costs drop.

People underestimate making games.

Can you be specific about how you think this is a problem?

yes. but its also not >that easy< to have an end2end ENGAGING experience built in couple prompts. at least for now.

I'm concerned that even now indie games could go from months or years to develop down to weeks or months. Maybe not single-prompt to ship, but even so I expect the economic consequences of this are going to be problematic.

yeah that is absolutely already the case.

there are no ai problems, only ai solutions. the industry is the problem.

What's the "problem" here? Brother, get serious. You are trying you make up a strange panic here "OMG, IT'S OVER FOR EVERYONE!". All I can see here are possibilities to make some stuff much faster. And only cheaper ass clones. You still have to come up with ideas, new things and polishing polishing polishing. Which takes time. ...I saw one prompt once "make it insane" and "make camera smooth". Brother, an AI doesn't know what "smooth" and especially "insane" means... 😅 You guys are really traveling on a strange boat 😸

If I could clone the game myself, why would I buy anyone else’s clone? Guess this means the only market left is data centers and prompts, nothing else is marketable. We are all self realized? Sometimes I analyze why I or other players games since I wonder what a game is even for, and it seems we want shared experiences and we’re headed toward isolation. No one is going to care about my accomplishments in a game that has 500 variations. This is the PC Tower of Babel. But as a creator, I love to play with it cuz gamedev is a game, but to what end?

Why is it a problem? Creativity will win out and consumers will have more options than ever. You cant cut corners in game dev anyways. After a prototype there is marketing, polish, UI/UX, editors, storytelling, great gameplay mechanics, performance - even with AI these things take hard work and effort.

The problem is one of economics. More options creates more competition for attention, making marketing more expensive and driving prices down. It seems likely to become increasingly non-viable to develop games as a business.

"nearly identical", and its not even 20% of the original

All these type of games are gonna be free. This is the worst AI will be and the costs for these capabilities in 2 years time will be pennies.

Sorry but I do disagree. AI will be able to copy low-effort games. That's for sure. But It won't be able to do more than that. And low effort/asset-flipping games were already an "issue" (cue 85% of steam and mobile games).

Imagine Hollywood spending $300 million on a movie then someone just cloning the movie from a script on wikipedia and trailer.

Fast follows for movies are also going to be a problem. IP is somewhat of a moat there, but not by much as you can make a knockoff version of anything while keeping major themes the same.

Well, to be fair in the opposite direction a LOT of games have misleading ads that show gameplay NOTHING AT ALL like the real game. My first iOS game Cylinderis I made because of exactly that like 2 years ago. A video that wasn't at all like the game, so I made the game.

I’d argue we actually start improving games to the point where they can’t be copied in a single prompt with these tools. Does anyone even want to play that shit game on the left anyway? If a game can be copied in a single prompt what does that say about the game?

But if you look closely, the details of the ai game make it slow, boring and basically want to quit immediately. Whereas the original game looks action packed, more animated, more entertaining. Close or 80% doesn't count. It will get there maybe but don't fall so quickly.

If AI can do even 80% of the work, studios need 80% fewer developers, or can produce games in 20% of the time releasing 5x more games than before.

Correct as arithmetic, wrong as a model of software production.

In my experience developing games, fewer people are actually far more efficient per person. There is less overhead from project managers and producers, fewer meetings and approvals. 10 programmers on a project is much better than 100 programmers if productivity goes up by enough.

Totally agree mate!

I find it unbelievable how many of the comments seem to think this is a good thing. This will make the market even more over saturated by under skilled people who think this makes them a game dev. Actual game devs won't be able to make any money and may find new jobs. This sucks.

You don't think ads for games that dont exist that get you to download an entirely different product is a little more predatory than getting AI to recreate said games that dont exist?

The problem is your wish to gatekeep what is apparently gameslop. I think we’re gonna be just fine.

I guess this will get rid of the scammy games that do not match the ads, and will replace them with overmonetized games that do at least match the ads?

We have laws on false advertising. As long as the fine is smaller than the profits it will continue. Just fucking ban them.

I think this is a good thing.

I can see why in this specific case, but it applies more broadly, not just to misleading advertisements.

This has already been a problem, sadly, pre A.i

Hey, does anyone remember OpenAI trying the same with Sora? What happened to it? Ah, right... they shut it down after a short while because they were losing too much money on it. Folks, you are wasting your time on an prohibitively expensive marketing stunt.

innovative gameplay mechanics have patents. Studio can sue them. You know like nintendo.

The problem is the novelty will wear off. It’s been months of people one shotting clones and trying to quick ship to market with 0 xp. Quick buck at itch, sometimes sloppy steam but after 1-200 purchases the trend seems to be refund, review and shame. We can keep that cycle up. As it becomes easier to prompt results, it becomes harder to unclone as the entire scaffolding is built around the use of the reference with very little wiggle room for expansion. Using the newer models, it’s still significantly harder to build anything novel and not just another clone of a clone unless you know what you’re doing still systems wise and scalar.

what is the problem ?

>describes something awesome >calls it a problem

It is in fact a problem.

I wish, still producing a good experience across multiple different kinds of devices is not that easy, and agents, because of how they are trained, are lazy and don’t care about specifics, they just do what the prompt says to do.

The 'lazy' is getting a lot less prevalent in extremely recent models. This week Fable 5.1 and GPT-6 Astra are huge steps forward. It does still take a lot of context to get specifically what you want from AI, but that versus building it by hand is a massive time difference now.

Yep, I already use both of these models myself in personal and work projects. They still have built-in “laziness” like previous models, but there is an improvement.

I feel like the new ones have more the opposite problem. They tend to overcommit and get into loops where no meaningful progress is happening and burn lots of tokens. If you're still getting short outputs it could be an underspecification problem. Or maybe just try /goal.

If the AI can reproduce your game I have some bad news for you and the game you made. The "game" in the video is a usual AD scam and not the real game it promotes. The AI did just a simple game loop..

Oh, I thought you were trying to show how the ai made game just doesn't look or feel good.

This is a bit forward looking. Yes right now AI games don't look quite as good in most cases, but that is changing rapidly. What happens when that isn't true anymore?

This should be treated as copyright infringement with nothing less then a 10.000 dollar fine or 5 years of prison. I'm so sick of scammers trying to make a quick buck by copying.

Does this mean I'll actually be able to play the game that those scam ads claim it is?

You can show AI an ad, or for that matter a few screenshots and description of a game and ask it to make the game. For most simple games you will probably get something pretty close even today, and the complexity of what AI will be able to produce only goes up from here.

Wouldn’t this be copyright infringement?