Philpax

So, here we are: a mere ten years and twenty days since my last blog post. That's one-and-a-quarter AAA game development cycles. That spans the rise and fall of cryptocurrency in the mainstream. And, most painfully, that's two full website rebuilds without any new content. Let's see if I can fix that this year.

It is 2025. Technology has always had a dramatic impact on the world, but it's hard to overstate to what extent that will be demonstrated over the next five years: we are on track for a 2030 that will redefine what our everyday life looks like, facilitated by inventions that have been the realm of science-fiction until recently. I am, of course, referring to AI, but I also believe that XR will "come of age" in this timespan¹Quite literally, given that 2030 will mark 18 years since the Oculus Rift. On the other hand, I made a five-year bet in 2021 about the same thing and I'm on track to lose that, so we'll see., and the two technologies are likely to combine to create a fascinatingly different future.

My goal is to document these changes, speculate on what's next, and demonstrate interesting projects along the way. But before I do that, let's clarify where I stand and what I've been up to.

A Droll Mission Statement

As with most people of my archetype, I have spent the majority of my life in front of a computer, and much of that as an active participant of the internet. My adolescent years were spent on the "golden age" of the internet - the years everyone is nostalgic for these days - and while I miss it, I also recognise that time only moves in one direction, and we are but mere passengers on its journey.

Because of the internet, I have been a member of many communities, met many people, and forged many relationships. It has given me a fundamental appreciation for the ability of technology to help us transcend our fundamental limitations and to bind us together in friendship and more.

Of course, the negatives are many: we have seen self-radicalisation, algorithmic manipulation, unparalleled financial exploitation, insidious propaganda, unsustainable growth, walled gardens of everpresent decay, and horrific acts of destruction. You don't need me to mention all of the monstrosities the internet has enabled: you live it every day.

But that's not what I intend to write about. People more well-researched and more on-the-ground are already on that beat, and they're covering it better than I ever could.²Consider 404 Media. I disagree with them on quite a few things, but they're very good at capturing the technological hellscape that we've been shuffled into. No, my focus is rather more straightforward: I have a deep love for computers, despite the deep level of damnation that permeates them, and that's imbued me with the ability to understand them and what they're truly capable of. At the same time, however, I have a deep love for people, humanity, and the dream of a better world.

I want to explore how we can use technology - including, and perhaps especially, that which belongs to the dominant class - to create a better future for everyone. More conceretely, I want to see unnecessary labour reduced through automation while ensuring the benefits are collectively shared, and to build systems that enhance human connection and creativity rather than exploiting them.

The challenge isn't the technology itself - it's ensuring these tools responsibly serve the many rather than consolidating power among the few. To do that requires us to acknowledge what that technology is - to know your enemy, as it were - which is something I fear people of my particular alignment are failing to do. To be rather more explicit: I am a leftist, and I subscribe to Kellyian techno-optimism.

Technology is fundamentally not apolitical. It cannot be considered in isolation, as technologists as myself are keen to do. It is impossible to uninvent, especially in a world where information can be shared at no cost and the costs of reproduction shrink by the day. Our best hope is not in denying its existence: it is to subvert its use by those who wish us harm and use it to improve the lives of all people, creating systems that serve collective interests rather than enriching oligarchs.

Fundamentally, I want to be part of a world where we build bridges, not walls. And while I don't expect I'll move the needle much, nor do I believe much of my output will have any material impact or be particularly relevant, I refuse to go gentle into that good night.

The Last Ten Years

University (2014-2019)

To understand the last ten years, we must first look at what I was doing before those ten years: I worked on the multiplayer mod for Just Cause 2, an action-adventure title by Avalanche Studios. I worked on the scripting, user interface, public relations, supporting tooling, and other things that no longer come to mind. This was my first "large" project, and it ended up fundamentally defining me. After running a sequence of increasingly chaotic, large-scale public beta tests, we released it to the public for everyone to enjoy on the same day I received my high school results: it was a very eventful day, night, and following morning!

I continued working on the mod for some time, but most of my attention was taken away by the newly-overriding concern in my life: my time in university. I had foolishly chosen to do two degrees, and had even more foolishly made one of those degrees an engineering degree.³My other degree was a Bachelor of Science in both Computer Science and Pure Mathematics, so I double-STEM'd up. Terrible, terrible mistake. By the time I realised it wasn't for me, it was much too late to do anything about it, and I decided that I should instead see it through. Simultaneously, I also worked near-full-time at my first two jobs in the tech industry, and took up contracting jobs to fill the spaces between. This was not the most productive period; I spent most of it attempting to stay above water, limiting the time I had available to cultivate intellectual curiosity in the things I cared about.

Still, it wasn't the worst turn of events; those degrees are great for getting visas, which is about the extent of what I use them for today. I also gave my first international talk at DConf 2016 in Berlin about the software ecosystem I'd developed for a hobby CPU architecture of my own design. And, to be fair, I closed out my engineering degree with a final year project working on the vision system of my university's Formula Student team's first autonomous racecar, which was pretty damn cool, all things considered. In any case, I maintain that I would have been much happier swapping out that engineering degree for an arts degree.

After that five-and-a-half-year chore, I finally graduated, and was ready to enjoy what the world had prepared for me; I was ready to rock. Just one small problem: the year was 2020.

Avalanche (2020-2022)

Given the circumstances, I was left with some time to think about what I wanted to do next. At this time, I was still at my second job, and getting on with the motions: there were interesting technical considerations, and the team culture was fantastic, but my interest in the company's work was fundamentally limited. It just didn't seem like the right place for me.

In 2015, the Australian members of the JC2-MP team (including myself and the other core developers) were flown to Stockholm, Sweden, to meet the Avalanche Studios team. Nothing immediately came of it for me - being that I was still in university - but, as a result, Avalanche ended up hiring the lead developer of the mod. By 2020, another member of the extended JC2-MP team had also joined Avalanche, and I decided to take the opportunity that I saw blossoming before my eyes. I was able to secure a referral through both of my friends, and sometime later, I began to work for Avalanche Studios as a contractor in Australia.⁴Astute readers may have realised that the timezone difference between Stockholm and Melbourne ranges from 8 to 10 hours. My workday usually ended around 18:00 CEST - or 04:00 AEDT. I'm a night owl, but I can't say I'd recommend doing that for extended periods of time. Your social life becomes somewhat difficult.

Around a year passed, and Australia's restrictions had eased sufficiently to enable me to leave the country. After considering it for some time, I decided to commit: I was going to move to the other side of the world, to Sweden, to see what it had to offer for me. Yes, it would be a paycut. Yes, I'd be leaving behind everything I knew. Yes, I'd have to leave my wonderful old man cat with my parents.⁵I did consider taking him with me, but as he was already north of ten years old, it didn't seem a good idea to put him through that much stress. Luckily, he took very well to my parents, who doted upon him and were in turn doted upon by him. Cats: truly wonderful creatures. But I needed to explore the world, one way or another. After jumping over a few bureaucratic hurdles, securing an apartment on the other side, and undertaking a goodbye tour, I hopped onto the first-and-probably-last-I'd-ever-take business class flight to Stockholm via Qatar.⁶A friend sold his flight upgrade to me and I figured I'd take it up as a once-in-a-lifetime thing. It certainly lived up to the hype - I don't think I've ever slept so well on an international flight before - but my hangover on arrival was entirely my fault. Refills may be offered, but that does not mean you should take them up on it.

Working at Avalanche in-person was both as I expected, and not as I expected. It's not a particularly large organisation; the team I was on was approximately 150 people, and the whole company was around 500 people. I felt like I'd gotten a decent sense for my coworkers, and I very much enjoyed working with them, but it dawned upon me that it wasn't the way I wanted to work. I've always been somewhat free-spirited when it comes to development: I enjoy the freedom to create and develop my own solutions, but as part of a studio, development needs to align with broader production goals and established pipelines. And there's nothing wrong with that! I knew that going in: but living it, I realised that I wasn't getting what I needed, and my performance was faltering as a result.

Ambient (2022-2024)

After just shy of a year as a full-time employee at Avalanche, I quit. I needed to do something where I could put my skills to full use, and I was willing to gamble with my employment and my migration status to find it. Luckily, after a short interview, I found my lucky break: I joined up with a tiny ragtag team in Stockholm building a multiplayer 3D game creation environment in need of someone with my exact skills and experience. The team, and the product, were called Dims at the time, and I stayed with them for the better part of 3 years.

This journey was varied, exciting, and chaotic, but - as most startups do - ended not with a bang, but with a whimper. I might write that up in more detail some day, but the journey took us from a closed-source game creation platform ("Dims") to an open-source multiplayer game engine targeting the web ("Ambient")⁷ Ambient. In some ways, I consider this a continuation of my JC2-MP work: I worked on many of the same problems and explored more of the design space that I was unable to last time. Aside from the business and user interest issues, there were also fundamental technological issues that we hadn't considered. I actually ended up giving a talk on these problems, which I will eventually link in a Talks section, but it's the kind of topic that merits a full blog post. , to one of the first attempts at an AI gamedev platform ("Braindump Inc")⁸ Braindump Inc (archived). Another thing that merits a blog post, but given that there are other companies in this space, I'm inclined to wait some time before discussing it. , to a second attempt at the same, to open-ended AI-related experimentation.

At the end of last year, the company could not come to a clear agreement on its future, and it was decided that the company should be wrapped up and the money returned to investors.⁹I was on holiday in Paris when I found out about this. Had I been in town, I would have... well, I wouldn't have done anything, because it was fundamentally not my ship to steer, but it's nice to believe you have some kind of control over your life, you know? That has once again left me unemployed, but I'm in a much more stable position than I was last time.¹⁰i.e. there are companies willing to pay for my services at this time

Open-source (2022-present)

Anyhow, that's not the full story. After my time at Avalanche and before Dims-Ambient-Braindump, I occupied myself by returning to my roots: game modding and open-source. Over the last three years, I have kept myself busy in my free time across several domains:

• Game modding and tooling:

  • porting various games to VR
  • developing tooling for modding games in Rust
  • helping formalise and run the Dalamud modding ecosystem for Final Fantasy XIV ¹¹ You'll notice there haven't been any new commits recently. As it turns out, process is only as good as people's willingness to follow it, but I'm still proud of what I pulled off and continue to facilitate. Also, free amusing fact: I have not played past the base game (A Realm Reborn). I have spent hundreds - possibly thousands of hours - facilitating an ecosystem for a game that I've only played for a few dozen hours. Won't that make for a fun blog post some day?
  • creating a process for timely plugin (update) approval for Dalamud

• Open-source AI development:

  • building client libraries and Discord bots
  • porting llama.cpp to Rust and building an ecosystem around it¹²Another project that was formally shuffled off this mortal coil. As it turns out, it's extremely difficult to keep up with a project of llama.cpp's development velocity, especially when you're trying to maintain a Rust-appropriate level of rigour, and efforts to do so will result in burnout.
  • designing and specifying the GGUF file format, used by all llama.cpp-compatible executors today
• and other miscellaneous drive-by contributions to other open-source projects¹³A not-insignificant part of which is documentation. I will happily copyedit your documentation and address your linguistic mishaps, because if I'm bothered by it, other members of your audience will be, too.

Of course, there are other private projects that I haven't mentioned here that I have also been working on, but I am reluctant to discuss them as they're in various states of completion. I hope to finish them sometime this year and bring them online.

To put a bow on it, the last ten years have ended much how they started: unemployed, a pile of projects at my virtual feet, and a bright future ahead. En avant!

Why Blog Today?

So that brings us to the here and now. The era of blogs is, charitably, long past. Much of the writing that happened on blogs now exists on social media, in another medium, or behind paywalls. So why blog today?

As I mentioned before, I think the world is going to change in dramatic ways, and I believe I'm well-placed to cover that from an explicitly-leftist technologist's perspective. However, I want to emphatically avoid the curse of the average writing technologist: I am, to the best of my knowledge, not an asshole. Many technologists either don't care to explore the ramifications of what they're hawking, or have explicitly given into the lure of Winning At Capitalism™. Not I, I say.¹⁵This may be difficult for me, and potentially you, to square with the fact that I often work for VC-backed startups. It does bother me, yes, but being actively involved in these spaces gives me unique insight into both their potential and their contradictions. We'll see what happens.

I also refuse to fall into the denialism of people on "my side". I have seen what these technologies are capable of; I have read the research; I have worked on these technologies. They're not perfect - both current-generation AI and XR are deeply flawed - but the progress is not stopping. An inordinate number of person-years and dollars¹⁶and yuan, too. Not a one-country race, as the USA is discovering. are being applied to address the problems that are being encountered, and it's working. So what are we to do?

With that being said, however, the secondary reason is that I'd like to formally share my knowledge and projects with the world; I have done this sporadically across a variety of mediums, but always in a slapdash form. I have considered using existing platforms for this, but they're all compromised in some way or another: they're not suitable for long-form writing, or the user experience is bad, or they're overly financialised, or they're owned by or cater to awful people. I would rather you read my content without having to go through a GDPR consent dialog, or having to dismiss a subscribe-to-see-more dialog on scroll-down, or feeling like you need to wipe down both computer and mind after visiting.

It's naive, but I believe blogging is, to some extent, a way to reclaim the internet I grew up in: a place where people could share what they were passionate about, uncorrupted by the pervasive incentives of the modern internet. Hosting a static site is easy-ish and cheap¹⁷Actually, currently free, thanks to GitHub Pages. That may change if I add more significant content., and it's a worthwhile investment on my part to make my little corner of the internet a bit less miserable. One of my idols is the late Shamus Young; while I didn't always agree with him, he was deeply influential on me as a programmer, writer and creator. I'd be honoured to have a fraction of his contribution to the greater ideosphere.

One last note: I have explicitly licensed all of the content here as CC0/PD. This is a statement of intention: I want my ideas and code - here, anyway - to be enjoyed by both people and machines. It would be kind of you to attribute me, but you don't need to. I have learnt all I have learnt through consuming the works of thousands of people: it is only fair that I pass that onto posterity, in whatever form it may take. Owning a part of the ideosphere has, and will, proven to be increasingly difficult to enforce in a world where information transmission is effectively free. The solution towards the degradation of intellectual property as a concept isn't strengthening IP laws, it's to work towards a world where a monopoly on creativity isn't necessary.

What's To Come

Realistically, I don't know how often I'll blog. It's taken much longer than I hoped to get this post out the door. But, when I do write, you can expect:

• analysis of emerging technologies' societal impact, particularly AI and XR
• examinations of how technology is reshaping work, creativity, and human connection
• deep dives into specific technical projects and the lessons they've taught me
• critical perspectives on tech industry trends and their broader implications
• explorations of how we can build and use technology more equitably
• technical tutorials and insights from my open source work

But, above else, my hope is to write interesting, technically-informed pieces that centre the human. You will get something for your time invested, or your money back¹⁸This offer is void if this work is published in a medium where you actually do have to pay for consumption. You're on your own.! See you soon.