Future of Hardware, Project Ara, and Market Pilot
The Project Ara developer conference 2 began today at Google Plex in Mountain View, San Francisco. It consisted on a series of keynotes to denote the plans for Project Ara to both developers and interested consumers. Google shared details about the roadmap of Ara, which we’ll cover here, and the more developer-focused details about the newer Ara development kits and inner-workings, as well as business information to get module developers up to speed on what to expect from their new endeavours.
The conference began with keynote speaker Regina Dugan, Vice-president of Engineering at Google, ex-director of DARPA and named one of CNN’s Top 10 Thinkers. She began by describing 1969, the year the United States took to the Moon and ended the space race of the Cold War that began with Sputnik. While the technology developed thanks to this competition between the Soviets and the West was crucial for plenty of the things we see today, she noted that there’s one more development from that year that is overlooked: the connection of two remote nodes on October 29, and the first online conversation.
She noted that developers have always had the ability to recognize the significance of events like these: quiet, but massive, technological advances. “[Developers] recognize that deep sense of importance when it happens”, she said. And she emphasizes the innovation and the abstraction of these ideas. “Innovation is bigger, innovation gets faster, and it allows us to help each other (…) amplify our voices and our power as a group”. It is no secret Google has an affinity for technologies like the internet, and it is made strongly clear in the words of Dugan.
A New Paradigm
She began talking about the innovation in microelectronics, smartphones and biology, and how abstraction and group thinking helped push them forward. Then she noted an interesting example of an abstraction combining the three: Fold It, the game that allows you to play protein biologist by folding proteins to help gather research for real scientists. “Harnessing the minds of the many” is what she said allowed Fold It users to solve a complex biology problem, by incredibly modeling the enzyme key to the Mason Pfizer monkey virus. Said problem stumbled biologists for more than a decade, and these “protein folding savants”, as she called them, solved it in just 10 days.
The 21st century is the century of matter, for both Google and Tom Knight (computer engineer and biologist) who was quoted in the keynote. And with 3D printing on the rise, the accessibility of new technologies – like Ara – will rise with it. Regina puts extreme emphasis on these new platforms that “become the stage, creativity and innovation of many minds that solve real needs” – like advanced microbiology.
They wanted to touch on this to introduce us to an idea they want to spread far and wide – “we are going to have to make hardware design more like software design”. Google wants the paradigm of creating hardware to change. The process of hardware making has remained relatively unchanged for half a century. Other aspects like software have been getting better, their patterns and the way they are developed, on which platforms and with which methods, are in constant evolution. The current engineering paradigm Google speaks of consists of grabbing the whole, dissecting into analyzable units, figuring out the mechanisms of each bit, then putting it all together. But integration is still done through PowerPoint design reviews or expensive prototype builds. “[For 50 years,] engineering change request numbers look like a rollercoaster”, she noted as she showed us a graph of the change requests skyrocketing on the prototype finalization of spaceships and smartphones alike.
Her answer to this is more computer simulation and computer aided design, perhaps better platforms to develop new technologies with those two incorporated, to test whether what is being built would work in the real world. “We have to virtualize much more of the design cycle – to verify design to a higher level of abstraction” she said, and later compared it to “compiling the design”. Like with Fold It, these platforms would democratize technological development and allow new technonauts to dive deeper into the engineering and scientific aspects. It’d allow for a more equal battlefield for startups and the big-fish OEMs that monopolize sectors of the industry.
Breakthrough advancements are complicated because there’s critical new information at every turn. Dugan noted that “our desire to predict the future exceeds our ability to do so”. This is a core aspect of Google’s gamble, as Ara could very well go the road of Google Glass if it doesn’t play its cards right. “Ara might not work for reasons we might understand – in which case we’ll swing it for 2 more years… for reasons we don’t understand, we’ll take a name”. But the future is what one chooses to build, and Google chooses to build the things they believe in.
Paul Eremenko took the stage to talk about Ara’s roadmap, a few of what’s changed, and what to expect soon. He said he thinks that Ara is the “hardware-analogue” to Android, and it serves its purpose in democratizing the mobile hardware. They are executing the Ara projects with fast and agile allied industries, with over 30 partners in 30 countries.
The Module Developer Kits allow you to make a module from scratch, and they also now include reference designs. They are also including a full release of Android firmware in the MDK for developers to get familiarized with the default software. Ara will be strong across the industry, Paul said, as they are arranging OEM contracts and manufacturer services to “help [developers] instantiate a module quickly”. New additions and projects will be provided to allow developers to reason about the design of modules and their interconnectivity despite the domain language differences and other constraints.
Paul mentioned business incentives to those who enter module development. They will provide early “assurance” to those who commit with certain types of modules, for example. To help developers capitalize their modules and users reach the product, Ara will feature a Marketplace platform with its own Ara configurator app. They also aim for physical retail store sales and they are already working with carrier partners for this. He also said the packaging for Ara products will allow them to send from a single module to a fully configured phone.
As for the manufacturing of Ara products, he said that, sadly, 3D printing at production values wasn’t going to be ready for Ara like they once previously expected. Instead they opted for a parallel path: sublimation of 3D plastic parts to decorate the polycarbonate shells and perform said decoration at near the time of sale.
There are modules in development from Nvidia and other big-names, and some other companies are coming up with innovative sensor additions like a pollution sensor. He talked very briefly about the current Spiral 2 platform model, but soon focused on Spiral 3. The new iteration will add new functionality and aims to have 20 or 30 modules available to match state-of-the-art smartphones in battery life, connection speeds, and camera. He said one of the improvements was in their adoption of inductive data transfer between the modules and phone, with a 150 micron air wall in their inductive gap. Multiple antennas will be able to connect to the same module, and antennas could also be added to the endoskeleton – which they anticipate would help significantly with performance in low-frequency bands. As for battery, they attracted developers outside the current mobile circuit, that offer energy densities much higher than those found in current OEM phones. This was done to accommodate for the reduced size a battery would need to fit into an Ara module. They plan battery hotswapping to be fully working on Spiral 3, for phones that are already powered up.
Next he talked about the market pilot for Ara. “We have questions that need to be answered through data”, he stated. Google wants to find out what consumers want, and what is missing. With Ara, they will face, head on, the paradox of Choice: “Consumers crave choice, but when they face it they get stressed out, and when they make a choice, they face remorse about making the wrong one”.
A main goal of the market pilot is figuring out the pricing models. What to monetize and how, what the elasticity of demand is, and the tradespace between subsidies and monetization. The data has to come from real consumers buying real devices, so Google has planned the market pilot to get a good estimate of all these fronts. Other factors like secondary markets for used modules are also interesting to Ara developers and Google. As for branding and marketing, the market pilot will help Google decide what road to take. Do they present the module brands? The Ara brand?
He next revealed the location of the market pilot: Puerto Rico. He gave a few reasons as to why they chose this location. For starters, cellphones are the source of internet for 75% of Puerto Rico’s users, and the carrier landscape is extraordinarily diverse and competitive. Puerto Rico is well-connected from an import, logistics and data perspective – which allows them to bring a lot of modules in for testing. They also explained that the vast and vibrant ecosystem of artists and entrepreneurs would be beneficial for Ara. They will have partnerships with Open Mobile and Claro Puerto Rico, two of the biggest and most competitive carriers in the region, Claro being one of the largest in Latin America. Finally, for a flexible retail experience, they are designing a “sales” truck for the market pilot.
If the market pilot is successful, if they navigate the paradox of choice, get a pricing model that works, and gather their data, where do they go with Ara? Paul talked about the intuitive sense of a two-sided market – for the platform to be appealing to the users, there has to be a number of modules for them to choose from. Contrariwise, for developers to have an incentive to develop modules, there has to be a strong user-base in place. Google figured out that a good starting point would be 2 to 3 module options across 10 or so categories. They worked with MIT researchers and plotted the trajectories of users and modules, and the data was used to determine the plan and help Google reach an equilibrium.
A few questions were raised during the Q&A, and no information was given on the base price of the phone. The Ara market pilot devices will be locked to Puerto Rico, but you can import one to use as a Wi-Fi only device anywhere in the world. We also learned that third-party endoskeleton development might happen in the future, but not without rigorous quality assurance. The time between the market pilot and global launch is still unknown, as the market pilot is there to figure out how to approach it.
An interesting idea that was mentioned was that of Ara modules being used for different devices or platforms that do not necessarily relate to Google or smartphones. Things such as plugging a camera module on your car, or a sensor module on your fridge. Definitely an interesting proposition that would be intriguing to see explored further.
We will report more on Ara in the following days, stay tuned for our coverage!