Based on X Space Episode 10

Watch a one-year-old learning to walk.

They don’t know they’ll be running by three. They don’t have a vision board with “marathon runner” or “mountain hiker” pinned to it. They just take one wobbly step, fall, get back up, cry sometimes, sigh loudly other times, and try again. One square of the living room floor at a time.

We know because we’ve seen it happen. We know that those shaky, uncertain steps are building toward something remarkable. But the baby doesn’t know that. They only know the next step, the one right in front of them

And somehow, that’s enough.

In Episode 10 of our AI Deconstructed series (our final episode for 2025), we had a warm and insightful conversation with Khellar Crawford, who leads AGI research and decentralized infrastructure at SingularityNET. We explored why the path to artificial general intelligence might require a similar approach: taking baby steps.

“In keynote speeches, AI is magic,” Khellar told us. “But in reality, it’s built one square at a time.”

That gap between magic and reality? That’s where most projects fail. We get so enamored with the vision of where we’re going that we forget to honor the unglamorous work of getting there.

AGI is the moonshot everyone talks about. The transformation that will change everything. And it will. But moonshots aren’t built in a moment of genius. They’re built through infrastructure nobody sees, decisions that compound, and systems that interact in ways we’re only beginning to understand.

The baby steps aren’t small ambitions. They’re how big ambitions actually get realized:

• Building infrastructure before breakthrough capabilities
• Solving real problems today while working toward tomorrow’s vision
• Understanding what people actually need, not what sounds impressive
• Making progress you can measure rather than promises you can’t verify
• Letting small decisions compound into emergent intelligence

That wobbly walk through the living room? It teaches balance, coordination, spatial awareness, and confidence. It’s building the foundation that makes running possible.

The same is true for AGI. Every piece of infrastructure. Every real-world application. Every deliberate choice about what to build next is all building toward something we can see clearly, but we can only reach it one step at a time.

Want to hear the whole conversation about how moonshots actually get built? Listen to Episode 10 of AI Deconstructed with Khellar HERE:
🔗  https://youtu.be/fnN5t7W6u1g?si=M12MJ2qSwO5XzaQY

And yes, sometimes the most revolutionary act is honoring your audacious goals with the humble work they actually require.