The AI Era and Geospatial Modeling
As I mentioned here and here, AI is steadily moving toward building a digital replica of our physical reality.
A new project from the creators of Pokémon Go
The team behind Pokémon Go means business. They announced the development of a new AI model designed for navigating the physical world.
Technological foundation
The technology is built on data gathered by millions of players. The concept of a Large Geospatial Model (LGM) is clearly inspired by the Large Language Model (LLM) approach.
System components
In this way, people and their smartphones, wearables, robots, rockets, and so on become agents of a future operating system mapped directly onto the space around us.
Evolution of the technology
This shift from text-based models to ones grounded in metric 3D data reflects a broader trajectory in the advancement of AI capabilities: moving from language comprehension and generation to interpreting and modeling three-dimensional objects.
Essentially, this means predicting what objects look like based on partial information, such as earlier observations.
Alternatively, AI can reconstruct and communicate details about new spaces even when they are only partially visible.
Predictive modeling potential
This example shows how we can begin predicting what already exists, and perhaps even what is yet to emerge in physical space.
Impact on governance and design
If our present shapes the future and vice versa, we can consider how AI might help guide our actions and decisions not just conceptually, but at the level of designing spaces about which little or nothing is yet known.
Practical applications
Beyond gaming, large geospatial models will find widespread use, from strategic territorial planning, which we successfully specialize in, to logistics and remote collaboration on projects that previously demanded on-site presence.
_Living the good life, as they say._
#spatialdata
#urbanenvironment
#territorialplanning
#spatialmodeling