"I want to see applicants who are obsessed with a hard technical problem and relentlessly chase it."

From running one of Runescape’s largest private servers as a teen to building AI that understands space and time, Pim de Witte knows how to push boundaries. We asked him more about the challenge and what he thinks a real breakthrough looks like.

Pim de Witte is the co-founder and CEO of General Intuition, an frontier AI research lab focused on spatial-temporal reasoning. He is also the co-founder and former CEO of Medal, the world's largest platform for capturing and sharing gaming moments, and Highlight, an AI assistant that captures insights for users across every app. Pim's love of video games and his expertise in reverse engineering both started early and have shaped his trajectory. At 14 years old, he started and ran the largest private server on Runescape. This initial success allowed him to support humanitarian aid work with Médecins Sans Frontières, including the co-creation of a crowdsourced mapping tool, MapSwipe, at the height of the Ebola crisis.

Why must this initiative exist - now?

There is an insane amount of AI talent in Europe right now - it feels like the epicenter for frontier researchers. But talent can only go so far without the infrastructure to support it. The other piece is the type of research that SPRIND will support: "leapfrogging to the next frontier" with new model classes, modalities, and agentic systems. The world is waking up to the idea that the next generation of intelligence won't come from scaling LLMs, but these ideas still need support and financial incentives. LLMs have consumed a lot of the compute and the attention cycle. Smaller, frontier AI labs need a mechanism that rewards long-term thinking and novel approaches.

Beyond money: what's the real 'operating space' teams get here?

It's hard to overstate the importance of peace of mind for deep research: imagine how locked in you could be if you didn't have to worry about paying for compute, raising a next round, or hitting milestones tied to a contract. SPRIND will give winning teams the room to test every idea, be wrong, rethink fundamentals, and fight for their big idea. Also, it's a chance to be a part of the first cohort of brilliant peers. I think we'll look back on this in five years or a decade from now and see it as a turning point in European-led frontier research.

"A Manhattan Project for sovereign compute"

What would a real breakthrough look like?

A real breakthrough wouldn't just be measured in the outputs of the winning teams, but also by the infrastructure we create for them to succeed. A non-obvious suggestion: SPRIND catalyzes efforts to create a truly European CoreWeave equivalent - a Manhattan Project for sovereign compute.

What responsibility comes with building foundation models?

You're creating the backbone that others will build on and push into new domains. The models you develop are going to shape downstream systems in ways you can't fully predict. This isn't a reason to decelerate progress or set less ambitious research targets, but it's an obligation to take evaluation, interpretability, and failure analysis really seriously. There's also a responsibility to choose partners you're ethically aligned with, because your design and deployment decisions could impact a whole ecosystem.

If we're going to see a golden age of open science, openness also has to be a part of that responsibility. Frontier progress accelerates when ideas are shared and challenged, but this breaks down when only some members of the community take advantage of the free flow of information and do not contribute - it becomes the prisoner's dilemma. There are of course aspects you don't share for security reasons, but secrecy as default won't breed innovation.

What's the biggest challenge for Frontier AI in Europe right now?

It's never been more important that Europe creates the conditions and infrastructure that can support frontier AI. We learned the hard way that developing this muscle is critical for defending Europe and the weight of this technology is very clear. Europe needs to believe it can originate frontier AI, and then support that belief with compute, energy, funding, and institutions.

Who should apply - and who shouldn’t?

I want to see applicants who are obsessed with a hard technical problem and relentlessly chase it. SPRIND Challenge winners should be able to convince us that they have bridged their vision with details: ideas will only come to fruition with careful attention to detail, but if your idea works, what does it mean? The milestones on your research roadmap, the team you've assembled should always link back to the high-level vision for your research and the overall mission.

Contact person

Pim De Witte

Co-founder & CEO, General Intuition