What's next for inaglobe (2026)

The last post on this blog is from August 2024. It said, honestly, that we had paused operations and turned everything we had learned into an open resource. That post has been the site's last word for nearly two years. This one changes that.

We're bringing inaglobe back, but not as the same organisation. We're rebuilding it as an agent-native one. The plainer way to say it: we're agentifying inaglobe.

Why now, and why this way

We think the world is tending toward a particular kind of collaborative innovation, and that it's worth building for directly.

For a long time, real collaboration across distance was expensive. If a community in rural Tanzania had a problem, and the students who could work on it were in London, and the expertise sat somewhere else again, the cost of holding those three together, across languages, timezones and context, was enormous. That coordination cost is why most good cross-border collaboration never happens. It's also why inaglobe was hard to run the first time.

That cost is now collapsing. A lot of the connective work that used to need a full-time person, the translating, the chasing, the keeping-everyone-in-sync, the not-losing-what-the-last-team-learned, is exactly what agents are good at. Which means a way of working that used to be reserved for well-funded institutions is becoming available to a small, mission-driven one. A community, a group of students, and a specialist can actually work together on a real problem without a coordinator burning out in the middle.

That's the future we believe in, and inaglobe is how we intend to practise it instead of just talking about it.

What agentifying inaglobe means

The mission doesn't change. Students learn by working on real problems with real communities in resource-constrained contexts. What changes is the layer underneath.

A partner describes a problem once, in their own language, and it becomes a structured, curriculum-ready brief. Partner relationships stay warm on a reliable cadence instead of going cold between semesters. Project knowledge lives in a workspace that survives across cohorts, so each new team starts from where the last one got to rather than from scratch. Agents run that connective layer. People do the work that matters: the judgement, the relationships, the actual engineering.

There's a second half to it. Students on inaglobe projects, we hope, will learn to build with AI responsibly, on problems where the constraints are real. Limited bandwidth, limited budgets, and accountability to people who will actually use what they make. We think that's what engineering education should look like now anyway, and there are very few places to learn it on real stakes.

The honest history

We're not starting from zero, and we're not pretending the first attempt was a triumph.

Between 2017 and 2023, inaglobe ran more than 25 live projects with 124 students across four universities, with over 40 partner organisations, and mobilised more than €500,000 for partner projects. The learning worked. A few projects, like Paige Braille, became ventures of their own. What didn't work was the economics, for exactly the reason above: the coordination needed people we couldn't afford. So in 2023 we paused, and open-sourced the whole methodology rather than let it disappear. The thing that ended the first run is precisely the thing we can now dissolve.

What's already live

• The project directory is now public. Every project from the 2017-2023 run is browsable at inaglobe.org/projects/directory. It used to sit behind a login. The track record is the proof, and proof shouldn't need an account.
• The open resource stays free to use. The pedagogical model, the innovation playbooks, the templates, and the honest lessons.
• Partners can propose problems again on the partners page.
• Students and alumni can register interest for 2026-27 on the students page.

What's coming

Over the next few months we're building the tooling that makes this real. A brief engine that turns a partner's problem into course-ready briefs. A public directory of problems any university course can adopt. The continuity layer that keeps projects alive across cohorts. We'll ship it piece by piece and write about it here.

If you're a course leader who wants real briefs for your students, a partner with a problem worth working on, an alum who wants back in, or someone who thinks this is the right direction for collaborative innovation, get in touch. This is being built in the open, and it will need the same thing it always needed. People who care.

---

inaglobe's full methodology stays open under the inaglobe resource. The 2017-2023 run, including what didn't work, is documented there. We think the honesty is the most useful part.