Building an AI game studio: what we've learned so far • Hacker News Board

Re: Building an AI game studio: what we've learned so far

#301

Post by password54321 » Thu, May 23, 2024, 11:31 AM UTC

charlie90 wrote: post #265

password54321 wrote: post #28

Earlier quoted context omitted.

Word spreads fast on good games due to Reddit and Twitch. Look at the success of Balatro and how much marketing that had. The number of games you listed just highlights how little Steam filters.

"Word spreading fast" means there is marketing going on. Thats kind of the genius of marketing to reddit/youtube/twitch as a "normal guy dev", it doesn't feel like marketing, but thats why its so powerful (see social media influencers). You could be even more cynical and look into guerilla marketing tactics they might use like fake views and upvotes/likes. And again, survivorship bias.

You are trying hard to be literal for the sake of sounding smart. It was obvious by marketing we meant paid. Everyone else caught on.

Re: Building an AI game studio: what we've learned so far

#302

Post by kibbi » Thu, May 23, 2024, 1:35 PM UTC

Hammershaft wrote: post #298

kibbi wrote: post #283

Earlier quoted context omitted.

Surely you're not claiming that a game with fewer than 1000 reviews is "languishing"? Any indie game with 1000 reviews is very successful. Picking your first example, Deadeye Deepfake Simulacrum has about 460 reviews, roughly 20,000 copies sold, an estimated revenue of about $120,000 while it's still in early access, and apparently a developer team size of 1. I'd guess that this game will net the developer several hu…

Deadeye deepfake Simulacrum has been in development for ~4 years and is still getting updated, so even at this estimate we're looking at a take home pay of 30k - 40k per year before taxes.

You're right. Three things:

1. He isn't working on this exclusively. In the update notes, he mentions that he's making the game in parallel to a job. He recently finished his degree and sometimes takes time off for his next project. In 2021 he said, 'As I’ve mentioned before, game development is not a full time job for me, it’s just a hobby that consumes a lot of my nights and weekends.' Obviously this must be considered when evaluating the revenue.

2. After he finishes the game, it will still have a long tail of sales over the following years. He can develop another game, whose long tail will overlap with the first one, then with a third game, etc. But I don't know how to estimate these long tails after the first couple of years.

3. It seems to be his first game on Steam.

This is actually a very interesting case. I think he does most things right (though he doesn't seem to do enough marketing on Twitter). I assume that he'll arrive at a place where he has the option to go full-time sooner or later.

Re: Building an AI game studio: what we've learned so far

#303

Post by chrz » Thu, May 23, 2024, 6:48 PM UTC

ToucanLoucan wrote: post #205

fabiopolimeni wrote: post #180

Earlier quoted context omitted.

Uhm…, even you use something low-code, or even no-code tool, you still have to learn how to code (former) or learn how to use some kind of visual scripting tool (latter). If you could simply type in plain English what are the mechanics you have in mind instead, then this new (simpler) way of creating game logic it may has the potential to unleash creativity from a broader audience. It is like saying why would I need…

Something that I never see addressed for this: So, let's just grant that someday, the tech will be mature enough that this is possible, and let's even say it goes beyond videogames to movies, to visual art, to graphic design, to writing, etc. Let's say that AI gets to a place where any joe blow can put in a prompt, and get a competent, and even let's be generous and say good product out of it. A solid 8/10. So... who…

Making a game is not big problem. Making it FUN is. Theres no formula for this. Lets say AI can make a game from prompt just like that - I dont know if its going to be fun to play.

Re: Building an AI game studio: what we've learned so far

#304

Post by somenameforme » Thu, May 23, 2024, 6:51 PM UTC

Hammershaft wrote: post #299

somenameforme wrote: post #281

Earlier quoted context omitted.

I'd drop the max reviews down by quite a lot. Sales:reviews tend to be around 60:1. So a game with 1000 reviews has sold tens of thousands of copies. We can look at something like the PS2 for some reality checks. It sold 1.5 billion games through 4000 titles. So an average of 350k sales. And of course that's going to follow something like a typical 80/20 ratio, which would leave us with a median sales in the ballpark…

The only source I can find for a 1:60 review to sale ratio is this: https://gameworldobserver.com/2022/11/15/how-to-count-game-s... This article explicitly mentions that that ratio only holds for hit games, and mentions that the ratio for smaller games is roughly 1:20, which is still a high ratio compared to figures I've heard in post mortems. Assuming a game with 1,000 reviews that takes ~3 years for 1 developer to…

Here's a search for you with about a zillion results. [1] It's definitely way above 20:1. Your link references that number only for games with less than 1000 sales, and provides the data - which is super biased towards games with very near 0 sales, but tons of reviews. It could just be the games were terrible, but another potential issue is shilling. Many developers buy fake positive reviews. To my knowledge nobody has ever been banned for this (though devs have been banned for naively writing fake reviews themselves), but I expect these games end up getting some sort of a deranking or soft shadow ban, because it'd be absurdly easy to detect on Steam's side, and the games that do it seem to often end up buried after what's presumably an extremely low sales:review ratio. Here's an article that also includes some of the games that are hiring shills. [2]

[1] - https://search.brave.com/search?q=steam+reviews+to+sales+rat...

[2] - https://steamcommunity.com/groups/Sentinels_of_the_Store/ann...

Re: Building an AI game studio: what we've learned so far

#305

Post by viccis » Thu, May 23, 2024, 7:32 PM UTC

antifa wrote: post #252

viccis wrote: post #134

Wonder how well this will do given that, going by reactions to even fairly anodyne AI usage in games, gamers really really loathe gen AI used to make games and game assets.

I can't think of any high effort, non-cash-grabs, non-asset-flips that were dragged through the mud for using AI.

Palworld was a big one. Tribes 3 is another.

Re: Building an AI game studio: what we've learned so far

#306

Post by Hammershaft » Fri, May 24, 2024, 3:44 AM UTC

kibbi wrote: post #302

Hammershaft wrote: post #298

Earlier quoted context omitted.

Deadeye deepfake Simulacrum has been in development for ~4 years and is still getting updated, so even at this estimate we're looking at a take home pay of 30k - 40k per year before taxes.

You're right. Three things: 1. He isn't working on this exclusively. In the update notes, he mentions that he's making the game in parallel to a job. He recently finished his degree and sometimes takes time off for his next project. In 2021 he said, 'As I’ve mentioned before, game development is not a full time job for me, it’s just a hobby that consumes a lot of my nights and weekends.' Obviously this must be consid…

Yeah it's definitely one of the better cases here, as with all the devs on this list I just hope he succeeds and can make more games.

Re: Building an AI game studio: what we've learned so far

#307

Post by Hammershaft » Fri, May 24, 2024, 4:12 AM UTC

somenameforme wrote: post #304

Hammershaft wrote: post #299

Earlier quoted context omitted.

The only source I can find for a 1:60 review to sale ratio is this: https://gameworldobserver.com/2022/11/15/how-to-count-game-s... This article explicitly mentions that that ratio only holds for hit games, and mentions that the ratio for smaller games is roughly 1:20, which is still a high ratio compared to figures I've heard in post mortems. Assuming a game with 1,000 reviews that takes ~3 years for 1 developer to…

Here's a search for you with about a zillion results. [1] It's definitely way above 20:1. Your link references that number only for games with less than 1000 sales, and provides the data - which is super biased towards games with very near 0 sales, but tons of reviews. It could just be the games were terrible, but another potential issue is shilling. Many developers buy fake positive reviews. To my knowledge nobody h…

Do you specific recent sources that back that up? The source I originally linked as well as this writeup[1] suggests that the review to sale ratio has shrunk over the years and is shrinking for games with lower sales. The sales & owner estimates that steamDB offers also generally pegs the games I listed has having a 10:1 - 30:1 ratio. Here's deadeye deepfake simulacrum as an example:

https://steamdb.info/app/1545990/charts/

Note that many of the highly rated games I shared had sub 300 reviews, so even if you were to have a great 60:1 ratio for a game with 300 reviews priced at $15 you're looking at a $185k gross after steams cut and before taxes for potentially 3+ years of development and support.

I really hope I'm wrong here because I would love a meritocratic indie game scene to keep flourishing, but the greater trend I see is a bifurcation between indie games that synergize with streaming and social media to become breakout hits and excellent games that prove more difficult to market that struggle to tread water.

[1] - https://vginsights.com/insights/article/further-analysis-int...

Re: Building an AI game studio: what we've learned so far

#308

Post by somenameforme » Fri, May 24, 2024, 6:31 AM UTC

Hammershaft wrote: post #307

somenameforme wrote: post #304

Earlier quoted context omitted.

Here's a search for you with about a zillion results. [1] It's definitely way above 20:1. Your link references that number only for games with less than 1000 sales, and provides the data - which is super biased towards games with very near 0 sales, but tons of reviews. It could just be the games were terrible, but another potential issue is shilling. Many developers buy fake positive reviews. To my knowledge nobody h…

Do you specific recent sources that back that up? The source I originally linked as well as this writeup[1] suggests that the review to sale ratio has shrunk over the years and is shrinking for games with lower sales. The sales & owner estimates that steamDB offers also generally pegs the games I listed has having a 10:1 - 30:1 ratio. Here's deadeye deepfake simulacrum as an example: https://steamdb.info/app/1545990/…

In your link, look at the median data he offers for literally every single thing except date. I think the following matches your overall selection of games pretty well:

---

Indie - 65:1 sales:reviews

$10-$20 - 65:1

>90% (review positive) - 51:1

250-1000 reviews - 89:1

Casual (genre) - ~70:1

---

The only thing that contradicts my 60:1 claim is the date datum, but I would take it with a grain of salt due to a endless quite obvious biases. For instance that article is based on the data 'leaked' up to 2018, but that leak was closed in June 2018. So you not only are missing half the year's data, but the two most critical holidays. There's also lots of different demographics that will have different purchasing/reviewing trends. For instance many people don't buy new releases, or buy them and only play/review them much later, and so on. There's just lot of biases in date alone.

On this site we often get a very unrealistic sampling of economic reality, because a lot of tech stuff is based in locations with extremely high local inflation both in costs and in compensation. For the US as a whole, median personal income is $41k, about $55k exclusively for full time workers. [1] $185k with ongoing passive income is doing wayyyy better than average even in America. And of course with independent development, you're not tied to any location. You could move to anywhere in the other 95% of the world where $185k would spend much closer to something like a million dollars. Places like San Francisco and Seattle are absolutely pricing out small scale entrepreneurship of all forms, but all that means is that we'll instead see things coming disproportionately more frequently from the other 95% of the world.

[1] - https://en.wikipedia.org/wiki/Personal_income_in_the_United_...

Re: Building an AI game studio: what we've learned so far

#309

Post by majormajor » Fri, May 24, 2024, 5:26 PM UTC

johnnyanmac wrote: post #235

majormajor wrote: post #217

Earlier quoted context omitted.

Nah, the standard will just change. A "AAA game" will be the ones that have that extra it whether that's something resulting from human curation (the person firing off prompts who knows how to tune them and knows what other people want) or style (the person assembling things to be more cohesive) or editing (the person making sure the AI puts out an absolutely amazing story instead of a meh one) or whatever. Basically…

as a nitpick, you can argue there was no "AAA title" in 1994. Games like MGS2, FF7, and RE2 in '97/'98 is where we commonly consider to be the beginning of the moniker for "AAA". But yeah, GTA 3 was over 20 years ago. A bit sad that it's still non-trivial to make a game that scope as one person, mostly because 480p assets and fixed lighting pipelines won't cut it anymore (regardless of scope).

Not something like 1994's Wing Commander or other PC games with Hollywood actors and cutscenes?

I'm not super invested in the nomenclature but it's odd to me that those are all Playstation games.

Re: Building an AI game studio: what we've learned so far

#310

Post by oriel » Fri, May 24, 2024, 7:59 PM UTC

stubish wrote: post #247

Bjorkbat wrote: post #156

I think a part of the reason why I'm kind of dismissive of AI when it comes to creative endeavors is that I haven't seen it do something that isn't already possible using a "not-intelligent" solution. For example, AI website builders. It's impressive as a tech demo that you can prompt an AI to make a website for you, but otherwise I don't think this is going to be disruptive to the web design industry since it's alre…

> haven't seen it do something that isn't already possible using a "not-intelligent" solution This is dismissing time and resources. I think things like this game maker project are extremely impressive, because it enables an individual to do in hours what it would take a team to do in months. Just '5 different cats' is impressive, because most game studios would just do one and repeat it, because their time is limite…

I've been using godot with chatgpt as a side project to see how much I can coax a game out of it, with minimal effort. Its a manual form of systems like this but definitely shows how generative tools enable independent devs or small teams to tighten their dev loops and rapid prototype ideas.

When you can rapid prototype ideas in a game search space, its more likely to get a high quality output, or at least games that hone in on fun more effectively than a AAA made with a waterfall model and a huge team with poor crosstalk abilitiy.