What it is advisable to know
- Niantic Spatial, a brand new firm from the individuals who introduced Pokémon GO to cellular, have began down a brand new path.
- After promoting its video games division, Niantic Spatial has moved onto a path that includes creating a big geospatial mannequin to merge the digital and bodily worlds.
- Niantic Spatial sees this expertise bettering the way in which AI machines see and perceive our world, by fusing the 2 for higher help.
Niantic is reportedly shifting away from the gaming scene and into the world of synthetic intelligence.
In a press release, the corporate introduced its new path, often known as Niantic Spatial. This identify may sound acquainted, particularly in the event you’re a Pokémon GO fan. Niantic reiterates its latest sale of its video games division to Scopely. In brief, the corporate is beginning with roughly $250 million in its again pocket. As such, Niantic Spatial is popping onto a special highway, one that can lead it towards a “giant geospatial mannequin.”
Niantic Spatial states it should search to bridge the hole between the “bodily and digital worlds” by way of this geospatial fusion.
If all goes nicely, Niantic states its new mannequin will facilitate higher understanding and interactions between individuals and machines to “enhance how we dwell and work.” This huge geospatial mannequin has reportedly been developed utilizing “over 30 billion posed pictures from tens of millions of areas all over the world.” Niantic states this could support LLMs (giant language fashions) with spatial reasoning.
Basically, these fashions (in principle) will have the ability to see/perceive the place one thing is within the bodily world, and its real-world utility to the house round it.
Niantic says that the difficulty with right now’s AI is that it is restricted to understanding “textual content, code, and pictures” solely. Its geospatial mannequin goals to interrupt that open by creating a brand new “map” that allows these machines to know our world the way in which we do.
A Spatial Future
Niantic hopes to put its AI fashions in smaller gadgets like AR glasses; nevertheless, it additionally has set its sights on autonomous programs. The submit additionally highlights the Niantic Spatial Platform, which consists of the Visible Positioning System, Gaussian Splatting (3D scanning/visualization), and developer SDKs. Niantic says it’s engaged on deploying proof-of-concepts throughout varied real-world areas.
Niantic’s curiosity in AI is not something new, as its CEO, John Hanke, publicly backed a pair of smart glasses from Good Labs final yr. Not solely have been these glasses mild, however they might be outfitted with customized prescription lenses, too. Moreover, Good Labs featured multimodal AI software program for the machine’s front-facing spatial digicam.
The corporate’s AI, Noa, can hear and see the world round you. It makes use of this information to ship actionable and helpful info to the consumer, and at instances, it does this with out you asking.
