The disproportionate cost and time-to-market of physical grounding
Part 1: Must intelligence be grounded in the physical world?
The assertion that intelligence must be grounded in the physical world. This isn't a simple 'yes' or 'no' question; it's a rich philosophical and practical debate with significant implications for how we design AI and even how we define consciousness.
The discussion around AGI often circles back to a fascinating, yet potentially misleading, notion: that "true" intelligence must be physically embodied. While the scientific and philosophical merits of embodied cognition are undeniable in the long run, this perspective can be a costly distraction.
My first view is simple: in the current landscape of AI development and deployment (and for the foreseeable future), the most impactful, scalable, and profitable AI will be largely disembodied. Here's why you should direct your resources and strategic focus away from the siren call of general-purpose humanoid robots and towards the immense opportunities in abstract, software-based AI.


