Navigating Wave Two of AI

Superintelligence, geopolitics, bubbles, and biology: those aren’t talk topics as much as they are the dimensions of living through the second major AI surge. A decade after the first wave pulled many of us deeper into the field, we now inhabit a world where everyone has an opinion, few share the same priors, and the loudest voices often have something to sell. This post captures the takeaways behind my recent session at SOCO without recapping the talk itself.

Demand better mental models

Wave two rewards practitioners who understand how the underlying models actually learn. Without that awareness, it’s too easy to fall for claims that every scaling law guarantees superintelligence or that current language models have already hit an immovable ceiling. Grounding ourselves in the mechanics is the antidote to both doomerism and blind optimism.

Know your own biases before judging the machines

Human biases shape every decision we make about AI adoption, policy, and investment. Recognizing that fact matters more than memorizing the latest model card. The real risk is building new systems with all of our old blind spots. Taking inventory of what we want the technology to confirm—or disprove—helps us avoid being swept away by either hype or fear.

Remember that incentives drive the narrative

We are, to a large extent, in an arms race fueled by the arms dealers themselves. Vendors, investors, and even some labs strive to keep excitement at a perpetual boil because it benefits their portfolios or power. Healthy skepticism is not cynicism; it’s a way to ensure we keep our agency while deploying tools that genuinely augment our teams.

Treat AI as a co-creator, not a prophecy

Superintelligence is unlikely to arrive in this wave—and that’s probably for the best given today’s geopolitical climate. What matters right now is that well-instrumented workflows already feel like superpowers. Integrating diffusion models, copilots, and automation into creative processes let me build a cartoon-inspired deck that simply wouldn’t have been possible on deadline otherwise. The win isn’t spectacle; it’s discovering how the tools reframe our roles.

Prepare for wave three by decentralizing innovation

Scale still matters, but it peaks faster than most roadmaps admit. The next phase will be less about trillion-parameter chases and more about practical utility, agents, and the open ecosystem that smaller labs can now access. That decentralization might finally let us respond in a more coordinated way when wave three hits.

If we stay curious, check our biases, and keep iterating on the workflows that AI already supercharges, we’ll enter the next wave with alignment, not anxiety. That’s the real takeaway behind the cartoons.