DePIN’s Solution to AI’s Biggest Blind Spot

Nils Pihl, the CEO and founder of Auki Labs, is a multifaceted professional. He is an entrepreneur, a behavioral engineer, and a social transhumanist. His expertise lies at the crossroads of contemporary technology and human behavior.

Approximately 65% of American grocery shoppers spend more than 30 minutes in the store during each shopping trip. On average, one out of every three visits results in an unpurchased item due to being unable to locate it. However, without significant advancements in how robots and computers perceive and interact with the physical world, it’s unlikely that a humanoid robot would perform better in this scenario. A crucial development for self-reliant robotic agents might be a decentralized machine perception network, which could potentially enhance their ability to identify and navigate their surroundings effectively.

Spatial computing and privacy.

Just as humans rely on memory or directions to move around, machines do the same. For a long time, humans and machines have relied on satellite systems such as GPS for guidance. Yet, with urban development progressing rapidly, the GPS system is becoming outdated.

Despite often taking it for granted, GPS functions as a line-of-sight technology, necessitating clear communication between your device and multiple satellites. Consequently, its performance can be subpar in densely populated urban areas and enclosed spaces due to the obstruction of this necessary line of sight.

As a data analyst, I’ve studied the evolution of location services and one of the earliest innovations was the implementation of signal strength measurement by mobile phones to nearby Wi-Fi routers. Over the years, through intricate triangulation techniques, companies such as Skyhook and Google have been successful in generating rough location maps encompassing numerous Wi-Fi routers globally. This is why navigation apps like Google Maps encourage users to enable their Wi-Fi for enhanced accuracy.

Over the past ten years, WiFi triangulation has faced numerous privacy criticisms and legal disputes. Sadly, in this case, privacy unfortunately did not prevail. However, there is some solace in the fact that WiFi triangulation can only provide an approximate location, roughly within a few meters, which isn’t precise enough for a robot to accurately identify the aisle for fetching ketchup.

As a researcher in the field of geolocation technology, I’ve observed an intriguing development: the emergence of visual positioning systems (VPS) spearheaded by innovators like Niantic and Snap. In essence, VPS functions by cross-referencing the real-world scene captured by a device’s camera with a memory of that same scene, housed in a centrally controlled cloud maintained by big tech companies. Effectively, it’s a two-way exchange: you share what you see, and they provide your location.

As a researcher studying visual positioning systems (VPS), I can confirm that these technologies offer remarkable accuracy under optimal conditions, with measurements as precise as centimeters. In less than ideal situations, such as in public urban spaces, the accuracy drops to within a meter. Despite this, VPS technology’s unmatched precision is what makes it an attractive investment for tech giants. They believe that VPS will be instrumental in advancing robotics and augmented reality (AR) glasses technologies into the future.

As a careful analyst, I urge you to consider the implications of this new technology. Looking back at the past, we’ve faced numerous privacy concerns with mobile social media. So, what might await us when tech giants have the ability to see the world from our perspective through augmented reality glasses, and peek into our homes and private spaces via smart devices?

Corporations need privacy, too

If you stroll into a grocery store and begin filming the shelves, you’ll likely be asked to leave. Retailers strategically position items at eye level to increase sales, so they put considerable thought into the arrangement of their merchandise. Consequently, the layout of visual merchandising in stores is closely guarded as a competitive advantage.

From an analytical standpoint, I can understand why the retailers are reluctant to disclose the layout of their stores to a centralized service. It goes beyond mere logic that our robot could just drop by a store and magically identify the location of every item, for this would infringe upon the proprietary information that each store holds.

As a researcher, I would advocate for the implementation of advanced product querying and navigation systems within stores. Ideally, these systems should be self-hosted and securely managed, allowing them to respond effectively to robot queries regarding specific products and providing accurate directions for AI and AR glasses without jeopardizing corporate security.

A careful observer will notice that DePIN asserts it will surpass the industry leaders of the Web2 era and provide us with ketchup while safeguarding privacy.

In contrast to humans, robots and computers have the unique ability to share spatial data with one another, enabling them to collectively understand and perceive the world. This collaborative approach to spatial computing empowers machines to enhance their navigation skills by accessing external sources of information. Within a Web3 Decentralized P2P Infrastructure (DePIN) framework, this data exchange can be financially rewarded and secured through cryptography.

Decentralized machine perception networks

Picture this scenario: Our grocery shopping robot’s ability to locate ketchup swiftly without jeopardizing corporate security is an adorable thought experiment. However, the potential repercussions of decentralized machine perception are truly awe-inspiring. When self-driving cars communicate and share real-time traffic data, we’re in for a significant transformation in transportation.

In Beijing, which has more vehicles on its roads than there are residents in Los Angeles, approximately 1,000 years’ worth of human productivity is wasted daily due to traffic congestion. Decentralized machine perception could enable these cars to flow harmoniously together, releasing hundreds of years of productivity each day.

In the future, decentralized machine perception enables privacy-protecting AR glasses with a compact design. This is possible as some complex spatial computing tasks are handled by nearby positioning servers, reducing the need for bulky hardware within the glasses. The impact on human interaction could be just as significant as the discovery of writing or the telephone. With over 100 billion intelligent entities in our civilization within the next two decades, decentralized machine perception networks will assist each entity in navigating their role in the world, whether here on Earth or in space.

Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc. or its owners and affiliates.