Hello and welcome to the first Logic World Wednesdays of 2026! I hope this year brings you peace and joy. It will certainly be a big year for Logic World, so let's get into the blog!
Sometimes we have multiple writers on these blogs, but this one is written entirely by and from the perspective of me, Jimmy.
New schedule!
You may recall from this video that in September 2025 I switched to working on Logic World full-time. I gave myself the deadline of January 1st by which to grow Logic World revenue enough to cover my rent and food. In the months since, we shipped the Subassembly update, the Demo World, subassemblies on the Creation Hub and many other improvements to the game & website; I also learned a lot about marketing and distribution, and we successfully got Logic World in front of new audiences, driving up sales significantly.
However, January 1st has come and gone, and I was not able to make enough money through LW for it to sustain my full-time work on it. I believe we'll get there eventually, but for the time being I'll be joining the rest of the team in working on LW part-time only. I remain dedicated to getting it done, and I'll continue to be the project lead until we hit 1.0 (+ probably also beyond 1.0).
With my reduced hours on the project, Logic World Wednesdays is switching from a weekly blog to a monthly blog. For the foreseeable future, we'll publish a new LWW on the last Wednesday of each month, updating you on the progress we've made during that month.
I'm already feeling way better about my work on this game now that I'm not staking my entire financial future on it. Although I'll be working fewer hours on Logic World, I'm more excited and energized for those hours, and I will make them count!
New music!
Markku is working on new music for the game! The first single from Logic World: Volume 2 just came out. Listen to it while you read the rest of this blog:
You can get Now, We Fly on Spotify, Apple Music, Youtube Music, and several other music places.
Volume 2 will be similar in length to volume 1, around 40 minutes in total. I think you'll love what he's cooking up for volume 2 -- Markku does phenomenal work, he's so good at driving the emotions immanent to the Logic World experience, and I couldn't be more excited to be working with him again.
Markku's music from volume 1 plays at arbitrary intervals during gameplay, but with the music of volume 2 we're looking at playing the pieces more contextually and intentionally. For example, Now, We Fly has this sense of grandeur, epicness and triumph to it. We want to really maximize the emotional impact by waiting to play it until you've spent many hours building up a huge and complicated machine. Everything has come together, your big project finally works, you're marveling at the expansive yet intricate operation of the whole thing, and BAM we hit you with this music. I think it's gonna work really well!!
One more thing: the cover art for this single was made by Markku himself! He started learning Blender specifically to make it. (Ecconia joked that it's more like "now, we swim" but Markku explained that the inverter is nervous about flight, and the music will give it the courage to fly up out of the water :))
New codebase!
We've mentioned this before, but never outlined it in detail: we are rewriting Logic World from scratch.
Among other reasons, we're doing this in order to dramatically increase performance & scalability. Logic World is about beholding and exploring the majesty of emergent computational complexity: many simple logic gates working in tandem come together to form astonishingly capable machines, and the core of LW is to apprehend such phenomena. With the rewrite, we're pushing this pillar to the next level.
Currently, Logic World can handle around 10 million simulated components before performance issues make it impossible to continue building. Our initial goal for the rewrite is to increase that number by a factor of 100, achieving 1 billion simulated components at a stable 60fps. Hence, our name for the overall effort is Project One Billion, or POB for short. However, in the long term, we believe we can get that number significantly higher than just one billion.
Here's a little preview where I'm testing the data structures we've designed for this purpose. A billion components, initialized in 30ms, using <50MB of real-world RAM.
POB isn't just about enabling singular builds with a huge amount of RAM or CPU cores or a giant display or whatever. It's also about enabling vast shared worlds for multiplayer. Our ambition is to turn Logic World into an MMO-scale experience. Think of a big computer you've seen in LW -- this one, perhaps. Now imagine you and a hundred other people each have your own computer like it in one shared world, and you've connected them to an in-game internet where you can send code and data between the different computers. Visions like this inspired me as I began development on Logic World all those years ago, and this is the kind of paradigm shift that POB will make possible.
Now that we've shipped subassemblies and the Demo World, we've done everything we wanted to do with the legacy codebase, and the majority of development effort is now focused on the rewrite. We started planning the architecture of POB in December of '23, and we started writing the code for it in February of '25. Now it's finally our full focus, and '26 will be the year that POB blossoms into glory!
I'm really, really excited about this. With POB, I think we're going to make a game that is truly magical and totally unprecedented. We've learned so many important lessons through building Logic World the first time, and our second take on this game is going to be so much better in every way. We'll share lots more on the blog this year as POB takes shape.
I want to end this introduction to POB by giving due credit to my colleague Zoey. She was the first to realize that something on the scale of POB is actually possible, and she conceived many of the core concepts that the new codebase is being built around. When we're all playing with gigabytes of simulated RAM in LW, it will be thanks to Zoey's efforts and inventiveness. I'm extremely grateful that I get to collaborate with someone so brilliant on this ambitious next stage of Logic World's development.
New runtime!
One thing I love about being a .NET developer is that it's fast. It gives developers an extensive and well-designed suite of high-level tools that provide huge productivity boosts unavailable in lower-level languages, but even so, apps written for .NET can match or (in some very specific cases) even exceed the performance of bare-metal languages like C++ and Rust.
Every year, a new version of .NET is released that makes it even faster. If your hyperfixations overlap with mine, you might enjoy Stephen Toub's blog posts about the performance improvements in each new version of .NET. When I started writing code for POB, it was on .NET 9, but .NET 10 has since released and so I took some time this week to upgrade the project for those sweet sweet performance gains.
In the new codebase, we're investing in robust diagnostics, including a comprehensive benchmark suite to track performance changes over time. This is useful for ensuring that accidental performance regressions don't sneak in, and also for measuring the gain from performance improvements.
I ran the benchmark suite on .NET 9 and .NET 10, and I was pleased to see a number of benchmarks with significant performance improvements! Here are the results for our most-improved benchmarks. Each of these functions may be run millions of times per second, so improvements of even a few % lead to serious gains.
Benchmark | .NET 9 | .NET 10 | Ratio |
|---|---|---|---|
SerializationBenchmarks - DeserializeStructureBaseHash | 1.03 ns | 0.16 ns | 0.16 |
SerializationBenchmarks - DeserializeStructureHash | 1.02 ns | 0.18 ns | 0.18 |
SerializationBenchmarks - SerializeStructureBase - data=Arbitrary 2 | 47.59 ns | 36.14 ns | 0.76 |
SerializationBenchmarks - SerializeStructureBase - data=Arbitrary 1 | 47.51 ns | 37.17 ns | 0.78 |
StructureModificationBenchmarks - WithChildrenRemoved - case=Remove ten children; all ten removed children had different hashes; all ten hashes still in structure after removal | 861.10 ns | 680.98 ns | 0.79 |
StructureModificationBenchmarks - WithChildrenRemoved - case=Remove ten children; all ten removed children had different hashes; all ten hashes not in structure after removal | 647.13 ns | 530.61 ns | 0.82 |
IsSemanticallyValidBenchmarks - TestIsSemanticallyValid - data=500 children, 5 different hashes | 1.37 µs | 1.13 µs | 0.82 |
SerializationBenchmarks - DeserializeStructure - data=Small | 22.32 ns | 18.48 ns | 0.83 |
IsSemanticallyValidBenchmarks - TestIsSemanticallyValid - data=10 children, 10 different hashes | 49.02 ns | 41.27 ns | 0.84 |
SerializationBenchmarks - SerializeStructureHash | 3.20 ns | 2.71 ns | 0.85 |
StructureModificationBenchmarks - WithChildrenRemoved - case=Remove ten children; structure has no children after removal | 281.06 ns | 238.65 ns | 0.85 |
SerializationBenchmarks - SerializeStructureBaseHash | 3.19 ns | 2.74 ns | 0.86 |
IsSemanticallyValidBenchmarks - TestIsSemanticallyValid - data=100 children, 100 different hashes | 473.34 ns | 407.90 ns | 0.86 |
StructureModificationBenchmarks - WithChangedBaseHash - case=Change data | 7.21 ns | 6.28 ns | 0.87 |
SerializationBenchmarks - SerializeStructure - data=Small | 29.96 ns | 26.30 ns | 0.88 |
SerializationBenchmarks - DeserializeStructureBase - data=Arbitrary 1 | 55.29 ns | 48.72 ns | 0.88 |
SerializationBenchmarks - DeserializeStructureBase - data=Arbitrary 2 | 55.15 ns | 49.09 ns | 0.89 |
This table was generated with Benchmark Buddy, a tool I wrote specifically for POB development to compare performance differences.
New freedom from twitter!
We're officially quitting twitter. There will be no more tweets from Logic World or Mouse Hat Games!
Thanks for reading! We'll see you on January 28th for the first LWW of the new monthly schedule. If you want to make sure you don't miss it, you can sign up for our newsletter.
Ciao!