January 2023


Welcome to our first blog post of 2023!

The new year has already been a wild rollercoaster and so I have a lot to cover.

Let's dive in!

Getting Down to Business

My year took off to a fantastic start with purchasing several online courses designed to help me succeed in business and the games industry. I was so inspired by these courses that I realized I can and should make my own courses using the knowledge I've obtained from spending 15+ years creating games.

Previously you might remember I did a panel at some local anime conventions on creating your own visual novels. The panels were very well-received, and in retrospect I should have developed it into a full online course. Maybe I can do that now.

But I think the first course I would want to release is a course on how to write murder mysteries. I've already started writing the course and it's at 13,000 words already. I know it's a small niche, but the content comes so naturally to me it's the easiest one I can do first. Plus, if you're a fan of Detective Butler and you want to know how you can write mysteries like me, I think you would definitely appreciate it.

After that, I'd aim to do courses on game development, software development, cryptocurrency, and who knows what else.

I always enjoy learning and I also enjoy teaching people, so I think these kinds of "information services" really suit me. If I can distill my knowledge into games, books, or online courses, then many people can benefit from the lessons I've learned over the years.

On that note, I've begun updating my YouTube channel more frequently. My home movies channel was a test run to see if I could upload at least one video per week (and how to get good at thumbnails and titles). Now on my main channel I'm uploading at least twice a week, with the goal within the next few months to be uploading one video per day.

I'm still trying to discover my niche on YouTube. At the moment, I've been uploading myself playing through my old games and giving commentary. And I definitely plan on doing how-to videos on software development and using technology. But if I ask myself what I enjoy making videos on, it's abstract topics like entrepreneurship and self-improvement. I might just do some combination of it all until I see what sticks.

Last but not least, I've updated the Goldbar Games website to include a mailing list form. Please enter your name and e-mail in the form to be notified of when our new games and demos are available and other important updates. The mailing list is extremely important because it's the one way we can stay connected directly with you. Whether we get deplatformed off social media, or the platform itself gets taken down, or the algorithms change and suddenly you're not seeing our posts anymore, the mailing list is the one way we can always let you know about our projects.

A Shocking Discovery

Two weeks ago, one of the outlets in my house suddenly sparked, and as a result all of the outlets in my bedroom, plus a few in the living room and kitchen (including the refrigerator) all died a painful death. Fortunately, nothing caught on fire, and none of the data in my computer was lost. However, this was still a major setback, as now I had to relocate my development operations while trying to find an electrician who was both willing and able to repair the problem.

The issue was that this house was wired with aluminum rather than copper, and so it degrades over time and becomes more prone to sparks and fires. This is particularly dangerous at the connections to the outlets, and in our case we actually had copper-rated outlets connected to aluminum wiring.

Apparently it is quite hard to find an electrician who is both willing and able to solve this issue. Of course we wanted an electrician who is licensed and insured, and among those only so many are able to service this area. And among those who are willing to service this area, they want to avoid working with aluminum because of the potential liability issues.

Eventually, we were extremely lucky to find one, and they did a phenomenal job fixing everything.

The only other option would have been to rewire the the entire house with copper wire, costing anywhere from $15,000 to $30,000. If I were to hire an artist for Detective Butler 2, the graphics would also cost within that range.

More on that next!

The burnt outlet

Becoming an Artist

Now this electrical incident does actually tie into my progress for the month. I had all my documents stored on my desktop, which could no longer run on the outlets in my room. So my choices were to either move the desktop to a different room, work on the laptop, or work on the computer in my shed.

Moving the desktop would be cumbersome because there aren't many other places to move it around here. And the computer needed to have its parts replaced anyway, so I took this as an opportunity to take the whole thing apart and give it a fresh build.

Working on the computer in the shed would be feasible. But the biggest deterrent is simply the cold weather. We've had lots of very cold snow recently and it's just not convenient going out there at the moment. Also, we've had to put dozens of boxes inside the shed as temporary storage until the electrical issue was fixed.

So the best option was to work on the laptop, which was doable because I had pulled all my data from my desktop's hard drives. But I didn't want to install everything on the laptop (and even if I did, there wasn't enough hard drive space). I wanted to install them on a new, clean drive on my desktop.

All of this is to say that I couldn't do any programming for the second half of the month. I could only focus on writing the story.

But I soon realized that writing isn't just enough. Over the years, I've learned the hard lessons of not marketing your game early (and also what effective marketing even looks like). Rather than simply tweet out how many words I've written every day, people want to see screenshots and pictures. And it's best that I set up a page that I can bring people toward, so they can sign up to be notified of future updates to the game.

So the mission became to create a landing page. And landing pages require art.

So how was I going to get art? I've been fascinated with AI-generated artwork using Stable Diffusion since last October. The shed computer (which was originally designed for mining crypto) was repurposed as an AI art generator. So I can use my laptop to comfortably remote into the shed's computer to generate as much AI artwork as I need.

I had been working on creating AI art for VN sprites since I got my graphics card in November. I've tested a lot of different methods, and I've learned how to train the AI to recognize subjects and even replicate styles of my artist friends (though none of them seemed to be very happy about it). This technology is game-changing for people like me who can only draw scribbles on pieces of paper.

I see AI art generators like Stable Diffusion as just another tool. If you're an indie game developer, you likely have a website. But you probably didn't hire a freelance contractor to build it yourself, right? You just used something like Weebly or Squarespace or Wordpress (I've done the same thing!). I could certainly make the argument that these "no-code website generators" are just as harmful to the livelihoods of freelance developers as "no-drawing art generators" are to artists. Or what about using premade game engines like Unity or Unreal instead of contracting an experienced game engine programmer? How exactly is it ethical to copy and paste code, but unethical to copy and paste art styles?

What a can of worms that opens up!

At any rate, my intent is to create what I call "AI-assisted" artwork rather than purely AI-generated. The biggest practical issue with AI art is that it can still take hundreds of even thousands of "dice rolls" to come up with a decent sprite. And even if you do, all of the sprites may still look slightly different from each other to the point where they still don't seem like they belong in the same game.

Here's my workflow: first I draw a sketch of what I think the character should look like. Then I send that sketch into Stable Diffusion and it creates a very nicely drawn sprite. I adjust the parameters and roll the dice until I find a handful of designs I really like. Then I pick the best one.

Now here's where things get fun. I print out the AI's design (on a piece of real paper!) and then sketch my own interpretation of it on a blank white sheet. Then I take my version of the AI's design and digitally add color. Finally, I run the colored image through the AI one last time to apply a filter which blends the lineart and the coloring together into a smooth finish.

My workflow using AI to generate sprite art

I used this process to create all of the character portraits you can now see on the itch.io page for Detective Butler and the King of Hearts.

On a side note, this process bears similarity to what I wrote about our own attempts at dynamic sprite generation back in 2015.

While the initial idea was to just use AI art as placeholders, I've received feedback saying that my drawings might actually pass for the real deal. I'd prefer to just use them in the final game, honestly -- it'd save a lot of time and money. And I'd be proud to see people enjoying my artwork. So maybe that's the new plan.

All of the 21 characters you see on that page were sketched and drawn in a single day. Their character designs were generated over the span of 5 days. If I was a professional artist, you can imagine how this would rapidly increase my drawing efficiency, especially if I trained the AI to draw in my own style. So to me this is proof that AI art is just another tool that artists can use to work more efficiently, and the ones who adapted and embrace the technology will benefit the most from it.


Now that electricity has been restored to my room, I can power on my desktop and get to work coding a playable demo for Detective Butler 2. The game will be coded in my custom game engine, reusing the systems I wrote for the first game. It shouldn't take long to get going, but it will take time to get right.

I've noticed that I've always had a habit of doing really well at the start of a game's development, but then as it gets to the 90% finished mark, I just give up and lose steam. This has resulted in games with rushed endings, lack of polish, and too many bugs. You can see it in the first games I programmed when I was 12 years old and you can see it in Snowball (which I made when I was 24). But the last 10% of a game's development is probably the most important, and I won't neglect it anymore.

So please look forward to a demo of Detective Butler 2, hopefully this month but most likely in March -- because of course February only has 28 days this year! The demo will include the prologue and first chapter, in which you'll be introduced to all of the characters depicted on the itch.io page. It's a massive cast for a massive game!

In the meantime, I will continue posting videos on YouTube and being more active on social media. My goal is not simply to make games, but to build a community of people who can enrich the lives of each other. And games are certainly one way of connecting people.

As always, thank you for your support, and I'll see you next time!