A 9.4-mm untethered flying robot achieves controllable flights, such as hovering, collision survival, and route adjustment.

This article isn’t about whether AI is conscious. It’s about how it behaves—or, more precisely, how it performs something that resembles thinking within a completely different geometric, structural, and temporal reality. It’s a phenomenon we’ve yet to fully name, but we can begin to describe it—not as a function of symbolic logic or linear deduction, but as something more amorphous, more dynamic. Something I call the fluid architecture of cognitive possibility.

Traditional human thought is sequential. We move from premise to conclusion, symbol to symbol, with language as the scaffolding of cognition. We think in lines. We reason in steps. And it feels good—there’s comfort in the clarity of structure, in the rhythm of deduction.

But LLMs don’t think that way.

Study finds that AI fundamentally lacks the human capability to make creative mental connections, raising warning signs for how we deploy AI tools.

Large language models (LLMs) show promise in tackling planning problems, but there’s a balance between flexibility and complexity. While LLMs can act as zero-shot planners, they struggle with complex tasks involving multiple constraints or long-term goals.

Many frameworks that address these challenges require task-specific preparation, such as tailored examples and predefined validators, which limits their ability to adapt to different tasks.