The next step

Keeping with the tradition of updating this blog every few years, I'm back to write about my latest move. 😅

I'm moving on from Spendesk after creating GemWorld Studio as President and Full-stack Developer.

Spendesk was a great place to work, and full of a lot of really cool, talented people. It hit a number of firsts for me.

• It was my first time working fully remote (which I'll admit, worried me going into it)
• It was my first time in nearly 20 years working outside of the game industry in a more "traditional" software engineering role
• It was my first time in a proper senior role amongst a team of hundreds of people, with a proper path to Staff Engineer and onwards

To say I learned a lot is an understatement. Everything from process to tech to culture to people skills, it was an undeniable growth experience. I'm especially grateful to the various Staff Engineers from whom I learned an absolute tonne of stuff, and to my Team Leads, Clément and Ben, for developing my "soft skills" and helping me grow as a person.

Initially on the Onboarding team, I moved to Growth with the goal of turning the product from a Sales-led experience, where customers required an extensive hands-on experience, to a Product-led experience, where customers could self-serve and get up and running quickly. The original goal was to expand Spendesk amongst A- and B- tier companies, which could self-serve, freeing up Sales and Customer Success to focus on C- and D- tier companies.

While there, outside of squad duties, I led a revamp of the SAML integration and the Roles and Permissions system. Both systems were legacy with the only people who actually remembered them being implemented was the CTO, neither had any real documentation, and the latter was integrated throughout the product and had already gone through two different attempts at being rewritten. The SAML integration was simple enough from a code point of view once I understood what actually needed to be done, but required 4-6 months to actually complete, as each company using it had to migrate on their end. It was a good experience in terms of rolling out updates and migrating customers in real life. The Roles and Permissions system was hugely complex, but, with the help of the CTO, we settled on a simpler solution and most of the work became making sure all functionality was reflected in the new system and releasing updates without breaking anything. It was a great learning experience.

So why did I leave?

Spendesk at the time still weren't profitable (though they are now, which hopefully has more to do with being a good product than me leaving 😅), and raising more money was difficult due to the financial climate at the time. They started a process of voluntary layoffs for those that were interested.

Starting my own company was always a dream of mine, albeit a somewhat vague "sometime" one. I applied - we had to submit a viable business plan in order to be deemed eligible - and the rest is history.

So, thanks to Spendesk, and the French state, I was able to start work on what would eventually become GemWorld Studio. For all everyone complains about taxes, I don't think I'd have been able to do this full time in a different country. French chômage, or unemployment benefit is pretty generous, especially if you've been laid off due to financial reasons (which I was).

With the help of two friends, Camille and Ivan, I've been working full-time on it since around Feb 2025. The actual company was founded in Dec 2025, and our first site, GemWorld, was launched shortly after.

We live in a bit of a strange time at the minute, what with the rise of AI. When I left Spendesk, it was viable to use AI to generate code, though the results were far from perfect. Since then, it's only gotten better and better, and to be fair, we also use it in GemWorld Studio. It does make the future somewhat scary though, as if this doesn't work out, will there be jobs to go back to? Coding was once considered a "safe" career path, as year by year, the number of jobs available has only increased. Now though, it's not inconceivable that coding as a career will become obsolete. Even now, there are days when I do more reviewing code than actually writing it. Perhaps when AI companies no longer subsidise the price of their services, the price to hire an "old-school" coder will be lower than paying for AI to do the same thing. More likely though, with open-source models, once they're capable of running locally, the need for multiple developers will drastically decrease.

Already I can see the difference between a developer that can code with AI vs one who can't. In the right hands, the productivity increase is massive. It's not even close. If you're hiring a developer, I don't even think you'd consider hiring a non-AI-using developer unless you have specific principles or practices that you absolutely need to adhere to. You're essentially paying more for slower progress. It's brutal, but the latest modals are highly competent, and they're only going to get better. When the famous "Will Smith eating spaghetti" video was released and people were widely mocking it, one comment on a Reddit thread stuck with me: "Don't you get it? This is the worst it'll ever be."

Enough doom-writing, back to GemWorld Studio.

GemWorld

Our first site, GemWorld, was launched in Dec 2025. The idea behind it was simple. One of my partners, Ivan, has been in the online space for over 10 years at this point, with a number of different sites, currently primarily around "users answer surveys and get paid". We'd both worked together years ago on the site Buzzville. We'd long discussed the possibility of creating a site similar to Buzzville, where players play games, and get paid, while we'd make money from ads.

In order to make it more "addictive" for users, I wanted each game to work with the other games. You'd play Game A to get items that you can use in Game B, which would give you something to use in Game C, etc. The goal was a little like the Marvel Cinematic Universe - you can watch each movie individually, but you get the most enjoyment from watching them all.

While the site is still live, it's currently on pause, as we were never able to get verified for Google AdSense. Ironically, the rise in AI meant an explosion in the number of sites, and so it became harder to get validated. Other ad programs don't even bother with you if you're not validated with Google AdSense - if Google hasn't validated you, they don't need to.

There was almost certainly changes to be made to the business model as well, but we had to pivot. I'm still pretty pleased with the universe we created, including the Tales from GemWorld section, which allowed me to do some writing, which I've always enjoyed.

Gem Arena

Enter Gem Arena. It's a pure gaming site. One of the key things we wanted to do was reduce the reliance on ads, so this site integrated a shop, so we developed the meta more and more to justify it.

The premise is simple. Playing games costs energy, energy reloads every hour, every game has a leaderboard, and everyday the best player is crowned the Arena Champion. There's still lots to be done, but the site has a solid footing right now.

I love what we've done in terms of technology for both sites. Both are static sites built on Svelte and there's some neat solutions in relation to shared assets and localisation. GemWorld is translated into 19 languages, including languages that require different fonts, or RTL mode. GemArena took a simpler approach and is currently translated into 6 languages. How we handled both is probably worth a blog post in itself, which I might get around to at some point.

Why Svelte? To be fair, while I've worked on Front-End throughout my career, I'm probably more of a Back-End developer, so I wanted something simple to use. We were definitely going to use a framework, so it came down to choosing the right one. I didn't want React - the little I've been exposed to, I didn't like, and I don't like Facebook (or Meta) in general - and Angular seemed too complex for what we needed. After some research, I landed on Svelte. There are some gotchas, but overall, it's a great framework, and very simple to get started with. Each component is pure HTML, CSS, and JavaScript (TypeScript), so it's very easy to understand and debug. It kind of makes me want to redevelop this site (for the fourth time) in Svelte 😅.

The entire infrastructure is set up through code, thanks to Camille, so releasing new versions is as simple as merging a PR or executing an action on GitHub. The biggest challenge has been GitHub collapsing somewhat more regularly than I remember, but overall, it's been very smooth.

The work never ends, but it's been a learning experience no far - not even including all the admin stuff that you need to do as President. I want to get Gem Arena to a point where it's self-sustaining, then see if we can bring back what we've learned with GemWorld in order to get it validated as well.

While this is happening, I'll probably need to look for contract work to keep the lights on. Whatever happens, it's always on to the next step though.