How to Make a Nintendo 64 Game in 2026
https://phoboslab.org/log/2026/08/xibalba64-making-ofTheir previous console had an edition made of the same metal as Anduril's lethal drones.
[1]: https://www.timeextension.com/news/2025/12/site-news-why-we-... [2]: https://kotaku.com/modretro-anduril-chromatic-attack-drone-m...
Modern toolchains, programming languages, dev cartridges and libraries from the community makes all the difference. Outside of some niceties like ICE (gotta write that GDB stub for target hardware at some point), it's not that different from what you'd experience with modern embedded work.
[1] https://github.com/spicyjpeg/ps1-bare-metal
[2] https://github.com/grumpycoders/pcsx-redux/tree/main/src/mip...
Until then, only arcades had already made the switch.
On PS2, PS2Linux, Dreamcast there was already C++ support, by the way.
I bought a MiSTer for me and my son and we played the heck out of Mario Kart 64 and Mario Golf. I think he picked up my "gaming is dead now" outlook since he doesn't really have any interest in new systems or games anymore
One gets the impression that gaming is dead when one's only contact with modern gaming consists of reading headlines about a handful of AAA games.
In reality, the gaming landscape has never been better, as game production has expanded in several directions:
- The barrier to entry into game development has been lowered; People who are talented in design but not programming now have the opportunity to create games
- AAA budgets have increased; those who enjoy large-scale productions will find increasingly large games
- There is still a market for old-school games, so plenty of (especially smaller) companies produce retromodern games
Heck, they even still produce commercial Commodore 64 games o_O
Of course, there is less room for radical innovation, but people do not play only radically innovative games. And regardless, innovation still exists today, especially in smaller-scale productions.
I hope it works out. I can see it working in the reverse: I play n64/ps1/ps2/gamecube games with my kid, they go to school and their friends are playing fortnight or valorant, and then you become the bad guy for trying to be an obstacle in that.
This is interesting, as I was about to say that gaming has never been more alive than it is now - but it's true that at least the developments in hardware are not nearly as exciting as they used to be in the past. Maybe that's what he's feeling?
Sony however decided to provide a full C toolchain based on GCC as well as a set of very high level libraries that abstracted away basically all aspects of the PS1 [1], all the way up to implementing a ready-to-go 3D engine and MIDI sequencer. This was an unpopular move as the libraries were slow, inefficient and most game developers were used to having full control over the hardware, but it also allowed inexperienced devs to get started quickly (as proven by the thousands of low budget PS1 titles that used the high level APIs) and eventually contributed to the popularity of the console over its peers with worse tooling.
Sony would eventually go on to release some slightly lower level documentation for certain aspects of the console as well as APIs that better mapped to how the hardware actually worked under the hood. It wasn't however until years later that the hardware would be fully reverse engineered at the register level [2], and it took even longer for most homebrew to move away from the official SDK (which only really started happening a few years ago as accurate emulators and homebrew SDKs started popping up).
They provide the devkit hardware, the specification technical documentation and the tool chain to compile the binary.
This makes me want to jump in. But also, I wonder how many people will ever buy and play your game.
Does Modretro's target market buy this type of stuff? Or are they the type to just download and emulator and ROM.
For the same reason it wasn't possible at the time for a third-party to make N64 cartridges themselves to avoid Nintendo's fees. But I assume these lockout chips are replicated or circumvented by now.
https://youtube.com/watch?v=bNN9XhGZ3Yw
(Also The Centre for Computing History YouTube channel is seriously underrated)
The cartridges from Modretro only work on NTSC N64s, though. I'm told the next production run should work on PAL, too.
Also,
2026 web dev: 4MB of JavaScript to render a button. 2026 N64 dev: entire FPS in 1MB
References: [1] https://en.wikipedia.org/wiki/Palmer_Luckey
But yes, the N64 isn't particularly well suited for this type of game (or anything 2D, really). I didn't know that when I started the project.
I got it to write a skill for driving the Ares emulator via GDB Remote Serial Protocol and it uses that for its feedback loop.
My point is, people get enjoyment out of game design too. It’s not just about nerding out on technical specs.
There's a time and place for everything.
I don't disagree, but I don't see it as a positive, looking at the current game releases and their performance being shockingly unoptimised, there is a lot of lost real world value in knowing the lower level details, it's like anything really, sometimes you have to do the hard/unappealing work to create your best work.
I've been a low level engine programmer in video games all my working life. I worked on a few AAA games that sold 20M+ copies, I'm not saying this to boast, I'm saying this to say I know how the sausage is made. My literal day to day job is fixing issues like "The RHI thread on the Switch takes 0.5ms longer than it should" and optimizing things as much as possible.
I think if you asked me few years ago, I would have also taken the same stance you did - that these tools make it "too easy" and we get unoptimized crap out there.
But you know what, nowadays I feel like I mellowed out a lot. A lot of these so called "friendslop" games are horrible in terms of technical work. And you know what? They still bring joy to people. They still make people laugh and have a great time with the people they like. One of my favourite memories from a game I worked on was reading comments from people who said they were looking forward to just playing the game after work in the evening with their friends. Was optimizing the IO performance on PS4 essential to make the game happen? Sure. But what mattered more was that the game was actually fun to play.
Nowadays I see the improvement in tooling as nothing but a positive. There will always be hardcore engine programmers who know how to do this stuff, I have zero doubt about it. But allowing people to just download an engine and make the thing is absolutely fantastic. To say that it's bad - to me personally - that's gatekeeping. Saying you can't make a game unless you know how to write a renderer or a physics system is so offputting to people who want to make a game but just don't know how - we should be encouraging them, not chastising them for it. Little Big Planet did it well as one of the first games of that kind, if you recall that - just saying, here's a sandbox, make a thing that brings joy to other people. Isn't that what gaming is about?
And on a more serious, business side - it's kinda crazy that Epic lets you use UE for FREE until you start making serious money on the game that you made. You can start a company, hire people, and just start making the game you want to make without paying anything for the tooling. Very few other companies in the world let you do this.
“Here’s how I’m doing the same thing, but also doing it in a way that totally avoids understanding or learning anything”
Cool story