Artifacts

Everyone says their AI can write a great brief.

Fair enough.

We figured we’d stop talking about it and just publish some.

Today, we’re launching Artifacts: a growing gallery of living publications generated inside dedicated APEx Whitespaces. Each one has its own job. One follows global geopolitical shifts. Another continuously evaluates the economics behind the AI infrastructure boom. Another studies how ideas spread through culture. Another tracks the changing human condition. They’re deliberately a little eclectic. Mostly because those are the things we genuinely want to read every morning ourselves.

The interesting part isn’t that they’re automated — it’s that they’re all running in parallel, live, and automated. Each Whitespace has its own research context, sources, assumptions, publishing cadence and purpose. Some publish every day. Others every week. Others only when there’s actually something worth saying.

Even better, they’re not all using the same model. Some currently run on GPT-5.6. Others on GPT-5.4. Others use Kimi, GLS or Faible. Turns out that once you’ve built a good thinking environment, model choice becomes much more of an engineering decision than an existential one.

That’s encouraging.

The publishing frontend behind Artifacts is custom-built, but everything else is just Ape Space. The same Whitespace architecture, the same version anyone can use today at space.apesonfire.com.

In other words, these aren’t handcrafted demos hiding behind special infrastructure. They’re Whitespaces tuned to do sustained cognitive work.

Over time we’ll share more of what’s behind them: the prompts, research setups, publishing logic and design decisions. Some will work brilliantly. Some will inevitably need to evolve. That’s part of the point. We’re less interested in proving an AI can answer a clever prompt than in seeing what happens when you give it something worthwhile to think about for months.

In the meantime, hopefully, you’ll find a few of the artifacts worth adding to your own routine.