Announcing Eternal Dusk, an Audio MMO Attempt

Eternal Dusk

Having just been resurrected by the substrate, a plantlike neural network 100 miles below the ground, you wake up to find yourself on a giant Alderson Disk, an artificial megastructure, the last thing in an old universe. Long ago, the universe ran on boring physics, but the ancients broke the universe's anchor. Now, history itself is mutable, and belief has impact on reality. As their last and greatest act, the ancients managed to build a bastion--a giant circular disk with a sun in the center, and an amalgam of souls underneath thinking the world into stability. Here, if you tell legends for long enough, they might just turn out to be true.

This is the world of Eternal Dusk, an attempt at an MMO for the blind by the blind. This post introduces just what this is about, who I am, and why and how I'm doing it.

The Big Idea

If you're blind and want an online game, you'll stumble on a few options--primarily muds, text-based games, and a few MMO-like things. Dreamland and Mist World are probably the biggest names, but we also have World of Warcraft mods and a grab bag of other smaller projects.

Muds are great except for all the ways they aren't. You have to be super technical to get it working in the first place, you have to be super good at the screen reader to keep up, and even among the blind they are a dying breed for the same reasons they long ago died among the sighted. You can only do so much with text.

As for the other blind-specific MMOs, well, let's just say they have issues. Dreamland literally plays itself for you, as do a few others. Mist World has decent mechanics, but takes pay-to-win and gacha to the absolute extreme. Play it if you want to learn why China feels the need to crack down on videogames.

So what about making sighted games accessible? Among other things, I help maintain Factorio Access, which does exactly what it says. We've done an amazing job: you can very much win Factorio, as long as you are the kind of blind person who can hold a lot of spatial information in your head. This is the best case, as we are working on an engine with full, official mod support that even takes API requests. In World of Warcraft, engine limitations prevent mods from getting everything. Also, owing to a substantial amount of automatic navigation, blind players can barely walk around without seeming like bots. Even if we got over these hurdles, being sighted is a massive advantage that one cannot overcome simply by adapting games not produced for the blind--at least in the RPG/MMO arena.

What I want to do is capture the best parts of the sighted MMOs, the best parts of what we have in things like Mist World, and push the frontier in ways that it can only be pushed if everyone is blind. We can absolutely create complex level design, complex classes, all of that. Whether Eternal Dusk succeeds or not, we can certainly do better.

My Motivation: A Tale of a Career, AI, and a Job Loss

I'm a professional coder with 20 years of experience in various forms. I've spent my career primarily working on performance and networking projects, in low-level languages such as C++ and Rust. I spent 4 and a half years at a company called Backtrace, which is a platform used by a lot of the biggest game devs in the world to debug their games. While there, I spent a lot of time working on our custom database, which for a variety of technical reasons is very much like a game. I'll probably talk about that one of these days. Suffice it to say that there's less difference between querying a terabyte of data and figuring out which characters are dead than one at first might assume.

But then I left. I bounced around for a bit. Then Dec 20th, 2024 happened. What was special? OpenAI announced O3. Not the variant you have today, but a variant that runs for hours at a time. It was a better coder than myself. It was a better coder than everyone I know with only one exception.

It couldn't do large projects, but the writing is on the wall. Coding as a career is changing or dying. I don't know how fast, but it's not deniable anymore. I happened to be doing the job hunt during that time, so I changed direction.

I decided to start taking risks. The first risk I took was joining a small stealth startup of 4 or so people. That lasted for 6 months, and ended for non-AI reasons.

During those 6 months we got the release of Opus 4. Opus 4 isn't a better coder than me, but it is faster, and it is capable of doing projects all by itself. Overnight, coders with access got a huge productivity boost.

If you know what you're doing, you can now do small to medium size software projects without writing code. Give it a good prompt and have lunch. Indeed this blog was deployed by AI, and much of the MMO has been written by writing up a few pages on exactly what I want and how I want it, then going off and letting Opus do it.

It is quite possible these are some of the most important things to happen in human history. We have gone from no AI coders to junior level in a year. In a normal world my reaction to the job not working out would have been getting a job at Google or something; in a normal world I'd probably not have even taken the risk. But in this one, it is entirely possible my career is dead in 5 years.

Maybe AI doesn't happen and this is an interesting diversion and we get a neat game.

But if I'm going to go out, I'm going to go out blazing. I've got some savings, the right expertise, very powerful tools, and even my own audio synthesis libraries. There's also enough evidence one can make a company around an MMO for the blind; quite simply, it's been done on games more basic than the one I want to build.

To that end I gave myself August to derisk the project. While we don't have a public demo yet, there were concrete milestones to make sure this was feasible. I hit them. So unless life throws a curveball, this is now a real project with tentative timelines.

If it's not an epilogue to coding then we got an MMO. If it is, it's a pretty good epilogue.

So What's the Game?

This is probably the question you're here for and the answer is that it went from being a when-I-retire project to a right-now, as-fast-as-possible project. Specific design decisions are still being made and I don't have all the answers yet. But I do have some key pieces figured out.

You will need as few as nine keys to get started. This copies one of the good things from Mist World: WASD to walk, j to attack, KOL; as an arrow pad for targeting to the left/right/etc. That's certainly not even close to all of the keys, but you can handle complex navigation and basic gameplay with just those keys. To that end, we're grabbing the idea of wall tones from a game called Manamon, which uses synthetic wind sounds to let you walk around complex environments by clearly indicating walls. We will also be grabbing a feature from sighted games, a sort of guide to your next quest objective. It doesn't give you all the answers and it doesn't explore the world for you, but if you are looking for something to do and don't feel like wandering around, it serves that purpose. You'll log in and choose a class (I intend to launch with 3), then specialize at various milestones.

I want to reward the slow mastering of complexity. What I mean by this is that the game starts simple and you can carry on that way, but if you want to be top tier you slowly master more and more mechanics. We're borrowing the job system from Final Fantasy in a way. Instead of multiclassing or having alts, you level all classes separately and can switch between them at the press of a button outside combat. Permanent mistakes aren't fun and being able to change roles means that you can get groups together even in a small playerbase. You pick what you're doing by setting your class. Plus, you're never stuck with boring. I'll add some small cross-class bonuses to encourage and reward playing them all--more on that at a later date.

You will make a decision at level 1: which kind of theme do you want, physical/warrior, priest-like, or magical? Then at level 15 and whenever the other milestones end up, you begin choosing two things: what kind of complexity you want and how active you want to be. I can reveal two of the more advanced classes right now:

• The Clan of the Ironclad (working title) believes in their armor so much that they can reinforce it through willpower. Through long, repeated rituals over the last thousand years and their reputation as stalwart defenders among the public, they have empowered their armor. Their abilities are defensive, with short but large damage bursts. To play it, you use two mechanics: fixed chains of skills each building on the next such that casting them all in the right order provides a large boost to the last one, and a quickly regenerating health shield powered by the condition of the equipment you're wearing.
• The runecaster (working title) seeks out runes. It draws them like cards, similar to Slay the Spire. Instead of planning, the runecaster reacts to what's in their hand at any given time. They can bestow powerful buffs and perform great works of magic, but only if the runes align. As a runecaster, it's not about knowing the order of things, it's about making the best of what you've got right now, while building up hands for more powerful abilities. But as you build your hand, you have fewer choices, and maybe you can't get the one you need to survive.

In a more mechanistic/engineering view, this is how I think of it: you pick something on the planning-to-reactive spectrum. Do you enjoy thinking about things ahead of time? Pick something on the planning end. Do you enjoy thinking on your feet? Grab runecaster, the reactive end. Ironclad would be somewhere in the middle--you don't plan, but you do the same things and make decisions before the last second. The other decision is your level of activity. Not everyone wants to press 10 buttons every couple seconds. I don't have a class designed for that yet, but the other dimension is how often you need to do things. If you don't want to have to pound the keyboard, take a class that's about long-term party-wide buffs. Unfortunately these two dimensions are not entirely independent, but my goal is to make them as independent as they can be.

As for the world, I teased it above. The name has meaning: the universe is dark, but this last sanctuary has lasted for a long time. For some reason, you were brought back instead of moving on. The afterlife is 100 miles below your feet, and for lore reasons it needs the dead to support the world. You are one of them--undying, always returning to life at the soulsprings. They say history is written by the winners. In Eternal Dusk that can be quite literal. Everyone forgot how this megastructure came to be for too long, and now to remember is to decide. The substrate sits below, mostly pulling the world into reasonable consensus reality that can't drift, but push against it long enough with story, and it'll slowly give in. And of course, what happens if it dies? Or gets attacked? It is the nature of MMOs that putting this into a single story just doesn't happen; there's always more content tomorrow. But I have ideas for all of these and more, and internal worldbuilding docs for writers which do more firmly ground the stories. As with technical information, expect more posts on that as the project moves forward.

And yes, there will be crafting. I could write a whole essay on why crafting is tricky and won't make launch day, but since people always ask: it's on the roadmap.

My goal is to publish progress reports of some form once a month, and to launch some sort of small alpha game using the engine before Christmas. I am hoping for a launch in early-to-mid-spring 2026 with content up to level 30. You will see more traditional marketing as that gets closer.

The Pay to Share Model

People have complicated feelings when I mention this is my shot at making a living, because that means money has to come from somewhere and I feel it is worth talking about that upfront. I usually wouldn't get into monetization this early, but let's be real: the current MMO landscape for the blind is pretty predatory.

Before I get into details, you won't need to spend money to play or stay competitive like you do with the current options. Also, no PVP. Instead, we will have better AI and mechanics, and perhaps indirect ways to compete.

A while ago, I realized that one of the unique things about the blindness market is that many blind people are isolated. The other unique thing is that money is not consistent, nor is it spread equally. The normal approach here is to use game mechanics to force you to compete with paid players, to make you want to buy via a perversion of the competitive spirit. Both of these are depressing facts.

But why not flip that? A lot of blind people get stuck with not much to do and few opportunities for social interaction. That's not fun. I've been there, and in many ways I still am. I also believe that if you give me your money, you should be doing so because I provide something positive. I don't think flipping the model can work for mainstream games. That market is too competitive, and anyone who doesn't race for the bottom will simply fail. But here, competition is low and rare, leaving the market wide open, and that may provide enough flexibility.

So here's what I plan to try, and what I call the pay to share model. Firstly, I don't care how premium currency enters the system or what it's spent on, but if you want to buy progress directly you will be buying it from a player who spent the time playing to make that happen. This lets free players get premium currency, and is the first kind of sharing. Perhaps some paid player got ahead, but someone did the work and got rewarded for it.

The second kind of sharing is that the more core purchases--expensive or with expirations--are all shared with your friends. To provide a few examples:

• A buff that is shared with the party but does not stack for 10% increased exp.
• Automated translation through AI at cost plus a small overhead, so that you can play with people speaking other languages.
• Guilds with various benefits, but you don't have to buy progress and can't e.g. "pay away" daily activities with premium currency as you can in Mist World.
• Voice chat channels, but you being in the same party grants it to everyone in the party.
• Fast travel items, but they leave the portal open behind you for a bit for anyone to use.

These give you options if you don't have money. Maybe only one of your friends buys it for everyone. Maybe you grind out something to sell to the paid players. But the important thing is that it makes interactions with paid players start on a positive footing. They add something to the community instead of being walls that get in the way.

Here's an ugly secret about many of the MMOs we have. If you actually interact with the people spending thousands of dollars, they're just people. It's the game forcing it to be us-versus-them, and they are doing it on purpose. Even if the paid players want to change it, they can't. Artificial scarcity among other things makes it so that there's not enough to go round unless you shell out the cash.

In either case, I have to keep the lights on and I'd even like to hire. But I can do better than the current offerings and I think I know how.

What's Next?

I can't commit to a timeline so don't take the below as hard deadlines, but it looks a bit like this.

• Fall-winter 2025: some sort of online game prototype with the engine. Like Diablo for instance, but smaller.
• Winter 2025-2026: worldbuilding, class design, and early alphas/betas.
• Spring 2026: a launch with content to level 30 of 100, probably 3 classes and 6 specializations.
• Summer 2026: get content to level 50 or so, the second place where more choices open up.
• Fall/winter 2026: level 100, and starting in on endgame content.

For perspective I expect taking your first path to level 100 to take 100-200 hours, but for bonuses from that to make subsequent paths faster. Numbers are TBD, but essentially I'm saying that the first launch is 10-20 hours of content per class (admittedly the same areas though), enough to make it to the second specialization point, and the second enough to make it to the third. As with other things I'll probably be publishing more specifics soon.

As for this blog, I aim to post once a month at minimum with at least some sort of update. I might post more often but at the end of the day I'm one person unless I succeed at an initial launch, and these take a lot of time to write up. In other words, watch this space. Later on there will be Eternal Dusk social media as well, but we aren't that far yet.

Postscript: On AI Use

If you don't care about AI, you can stop here. Many do however, so below is my disclosure and response.

I am using AI heavily, primarily for coding and secondarily for some of the writing and worldbuilding. I am not yet using AI for sound and music, and won't change that until it's at least as good as the average human. You will probably see 11 Labs or similar somewhere for voice acting. There will probably be at least some AI-powered chat moderation. If you are concerned about your privacy, be reassured that there will be proper terms and conditions before launch, outlining any third parties to which your data may be sent. I thought about whether or not AI use was optional and whether or not it was worth trying to satisfy everyone, but the problem is that it's AI or no game. If you choose not to play that's your decision. But AI is what's letting me give you something to avoid in the first place.

The best thing you can do if you want more humans to be involved is to support the game. That said I'll also be up front that as I grow the team I intend to make sure everyone has access to AI for these purposes. Between 10 humans or 10 humans with an AI sidekick, I'll take the second any day. In my opinion, so will you once the tech matures more and I prove that I have good judgment on when to use it. AI isn't a negative. It's a possible avenue to building those huge audiogames everyone always wants with unending content and big open worlds.