On the Internet’s Next Act
The internet’s next major act will be augmented vision – smart glasses. Svelte, stylish, unnoticeably tech, and capable of delivering a new reality based on what is around you. Today, we are bound by a 6″ screen. Tomorrow, the internet will be all around us, absorbed visually – and invisibly – in ways that make smartphones seem like a camcorder of times past.
Below I will outline how this will come together, as we are now very rapidly arriving at the power and performance profiles required to allow a lightweight glass extension which we wear to distribute processing to a much more powerful phone we also carry. This will enable a future internet that is intelligently bounded by objects, places and people in ways we can now only imagine.
Much discussion around smart augmented reality glasses happens across the imaging dimension. The visible light boom of cameras only goes so far however in building AR’s future. What people don’t see is that invisible light – or wireless radio frequency – is going through a corresponding boom. New SoCs, buoyed by step function increases in digital signal processing and perception math, decrease the chaos and entropy of atoms in the invisible RF spectrum.
Such advancements are only superficially visible to folks working outside of semiconductors. One way to think about Ultra Wideband1 wireless today, is that it’s where LiDAR for autonomous cars wants to be—commercially available, cheap, and bullet proof. Ultra Wideband / UWB is effectively wireless ‘light’. However, while light suffers 100% from occlusion and requires high processing power to capture and analyze, Ultra Wideband wireless signals do not.
Wireless is a complex beast. How do you connect and pair devices, what is the advantage of each protocol, and what other functions does wireless serve in the internet of the future? The truth is way more than what’s talked about today.
Much breath is wasted over 5G, which is largely a mirage… The real innovation in wireless is happening at the localization and perception levels of the stack. Ultra Wideband gives a 100x increase in fidelity for localization. By using standard handshaking at the physical layer of the 802.15.4z standard, novel properties have emerged for not only localization of devices, but also permission, access control and commerce. In fact, UWB will replace Bluetooth (10x the throughput) and subsume NFC—both transferring of high bandwidth data (phone to glasses) and short range payments will standardize around UWB.
And while visible light and cameras serve up a sort of ‘permissionless persistence’ where pixels are stored and can identify people and environments, perception and localization from UWB is a permission-based layer that can work in the background and is effectively computationally free.
This is an important distinction. Why?
The analogy goes like this: talking is cheap but listening is expensive—a little bit like human beings. In order for Ultra Wideband to localize in space, small cheap devices will send pulses of information—these tags can last years on a battery and standalone UWB SoCs cost dollars. Phones that can receive UWB signals will be able to listen and filter accordingly and that ‘listening’ is expensive; but since these devices (phones) will be charged daily, it’s okay for the listening device to budget a power expense. The key to UWB is that the spectral efficiency for ambient sensing applications will be two orders of magnitude greater than it ever was with previous generations of wireless.
How is this distinction critical for AR and peripheral devices connected to your phone? Glasses will only catch on if incredibly small and lightweight. For such a device to understand what’s happening it’s not okay to do a bunch of neural network image inference on-device. Having a camera ‘on’ all the time on smart glasses is simply untenable. You would never turn on the camera, record a set of pixels and pre or post process all of this information to localize where you are. That is fine in a Tesla with 75 kWh of battery capacity, but in small lightweight smart glasses the battery will be tiny, so turning on the camera – and uploading pixel data to the phone – will be incredibly expensive.
It’s also important to explain how Ultra Wideband increases accuracy by reducing entropy at the carrier frequency level, since this is its fundamental breakthrough.2 UWB uses time of flight, looking for a signature that is in sync between when a packet was sent, when a response was sent, and when it is received. This capability to leverage physics or time-of-flight is similar to the reflection happening with LiDAR.3 Today with Bluetooth a BMW key fob validates that you’re near the car (using some secure keys). With UWB, the only way a thief can replicate being 1 meter from the car, is to be 1m from the car. Any other layer that is retransmitting introduces delay. In this way replay attacks are solved – UWB gives absolutely unbreakable proof of location – even if someone intercepts the signal, they cannot use it to attack. This type of time synchronization stuff is not new; what’s new is that time of flight can now be commercialized at a scale once thought unimaginable.
What does this mean from the application perspective? Say you’re a surgeon in a hospital room using smart glasses where every room looks the same. All you need is a UWB mapping of rooms to localize that the surgeon is really in room 403—this could never be accomplished with an image-based AR Cloud in a computationally effective way. The same is true for smart home applications or any indoor space.
From a business perspective there is a fundamental need to verify that wireless transactions are precise and reliable – to create digital confirmations that two atoms are next to each other. In the world of moving atoms and moving bits, having a secure and precise way to verify this changes everything. The need and the appetite has been there for many years, but the tech wasn’t ready. It was clunky and unreliable. Ultra Wideband will usher a renaissance for an entire set of localization use-cases.
Imagine you are wearing smart glasses and there is a payment terminal in the store – and you are in front of it. There is no risk that someone else can copy your transaction. But security actually goes much further… We live in an era of emerging deep fake generated video and photos. Recorded video – regardless of the input device being a phone or smart glasses device – can use proof of location. If someone claims to have recorded a video of Trump using deep fake ML techniques it’s impossible to analyze this in the pixels. But with localization, you can.
How so? – if people widely adopt platforms where every phone and glasses combo cryptographically signs everything that comes out of the devices. You aren’t going to conclusively tell that an image or video that was generated was real; but you would know that someone with the private key for a device was able to record and upload it to a server at a specific location and time stamp. In that way, faking a video of the White House lawn from afar would be impossible. In this way metadata validates creation, cryptographically signed with your signature.
It’s clear that a developing narrative for secure proof of location can reduce the ability of adversarial threats. Not all threats are tied to specific locations, but with AR comes new questions for how proof of work and secure timestamping come in to play from an ecosystem perspective. We don’t know yet. The way I think about it is simple: just as proof of work is very important for Bitcoin because it’s the first solution to the double spending problem at scale, proof of location for how atoms move is the only real solution to the spatial double spend problem.
Of course, the key to any platform approach is finding new killer applications. The smartphone was a platform enabler for countless new experiences that we could never before imagine. For glasses, let’s be honest: they will only catch on if the integration of radios and sensors don’t blow the size and aesthetics appetite of consumers. And adoption must happen before the discovery of many different jobs to be done. That’s why this will be a gradual process, with smart glasses deeply tied to phones.
I would argue there are two constraints present around deploying new hardware in the form of this new internet that don’t exist purely in software, but which are going away.
From the highest level (a platform provider perspective) new consumer internet hardware needs to provide an economic benefit around owning the user, either in the form of being a profit center (they make additional money) or a cost center (there are cost savings in the overall benefit and value chain). Ideally a profit center motive – where if you spend x you get some fractional 1.x benefit.
But hardware costs at the chip level are fundamentally different than software. No company wants to invest money in a new platform if it can’t interact (via a connected secure internet) with other devices and paying users. And it’s not just the fixed costs that are high in hardware, so are the variable costs with deploying chips. The beauty of how Ultra Wideband is evolving is that consumers are increasingly entitled to this innovation layer…they are and will be getting it largely for ‘free’. Contrast this to LiDAR, where even if it worked, self-driving car companies cannot commercialize this effectively at scale.
The challenge for platform adoption of new consumer internet hardware will be appealing to the motivations of participants who will build apps, giving them distribution and the ability to monetize how and where they fit in the network.
This is where, inevitably, commerce will be a wedge for the deployment of new smart glasses. Cryptocurrencies have been talked about as a potential currency of the internet. This is a contentious point for some. Regardless of whether you believe that, there will likely be a sovereign currency of the next internet. One that crosses boundaries of physical world interactions. If a smart glasses platform wants programmability in a network for payments, it needs to support a modular ecosystem of interconnected participants and developers. Whether you believe that future is a recursive programming protocol like Ethereum, or Bitcoin, or something else. In reality, any approach that will win around commerce will need to assume base layer wireless security for physical world transactions that are inherent in secure private key storage, and in the immutability that comes for free with Ultra Wideband proof of location.
The internet is the new nation state, the connected participants its citizens. And while new internet hardware platforms are built out, the promise of these new forms of commerce are a deflationary force like few we have seen before.
People like to pontificate on the future, what will happen next in vertical end markets. The reality is we don’t know. But what I have outlined here will happen in some form.
This truth is that we are now living in a permissionless persistence world. Always on perception and localization from cameras means we live in a new frontier where data persists without you always granting access. Whereas the localization happening with wireless is being designed in a way that is secure, opt-in and will allow never conceived before trusted interactions. This is really important for user security and privacy.
If it’s not abundantly clear, the complementary layer to imaging that will make new phones and smart glasses work together securely, while preserving privacy, and enabling new AR applications and commerce absolutely requires wireless. That layer is Ultra Wideband.
And make no mistake: there will be a Cambrian explosion in the amount of connected devices that are powered by image and wireless sensors. Think billions coming on line, and in the future trillions.
How this new internet develops is largely up to the creators who will build new amazing applications on top of these new tools. And the capabilities these creators unlock will push the world way beyond anything we have ever seen.
I have been talking about Ultra Wideband or UWB <a href="https://twitter.com/stevecheney/status/826595737479823360?s=11">for 3 years</a>—see this podcast <a href="https://overcast.fm/+HgwtnUi_g/41:13">Steve Cheney: The Next Internet</a> at 41:15 for more of my thoughts on the evolution of the internet and autonomy. ↩
For a great primer on Information Systems Theory read George Gilder’s <a href="https://www.amazon.com/Knowledge-Power-Information-Capitalism-Revolutionizing/dp/1621570274">Knowledge and Power</a>, which talks about Qualcomm’s advancements to OFDM in the 90s reducing entropy at the carrier frequency vs other wireless protocols of the time (TDM). A very similar thing is happening today with Ultra Wideband in my opinion. ↩
With Bluetooth, receivers looked at the radiated field from an antenna, and attempted to detect decaying signal strength (it drops exponentially with distance). ↩