Emerging Technology and Media, May 13th Edition (and MVP)
What is this newsletter?
This is the first edition of what will hopefully become a recurring newsletter covering emerging technologies and media. The purpose of the newsletter is to provide practical information and analysis related to emerging technology and media. I want readers to understand the present reality of these trends and how they are developing so that they are better oriented around their future.
I’ve been interested in these topics for a long time and I’ve wanted to find ways to channel this interest into something shareable and useful for others. As an entrepreneur steeped in the prevailing wisdom around lean startups, this is a minimum viable product. I’m sharing this in order to learn from the experience and collect feedback.
The Big Idea
One of the themes that cuts across today’s technology trends is that technology is becoming more human in terms of its behavior, its capabilities, and its appearance. Historically, computers haven’t been capable of many of the things that are taken for granted by humans. We’ve recently grown accustomed to computers beating humans in activities where raw horsepower can outstrip the complex cognitive capabilities of the human brain and the training we accumulate through life experiences. But progress towards computers emulating more organic activities, such as conversations, has not been as publicly visible. While assistants like Alexa sound human, I don’t believe it is common for an adult to mistake it for a human being.
This week saw the demonstration of Google Duplex, a technology that allows Google Assistant to imitate human speech and carry on specific types of conversations autonomously with unwitting human counterparts. The technology is capable of processing the human side of the conversation, interpreting it, and coming up with an appropriately human-sounding response. The demonstration is striking not because we shouldn’t have expected this but because the technology is already here. If I were on the receiving end of that phone call, I don’t think I would catch on very quickly to the reality that it wasn’t another human being on the other side of the phone.
It is logical to take this demonstration and extrapolate to a future where we won’t be able to tell that we are talking to a machine unless we do unnatural things in the course of the conversation. When I’m talking to a salesperson they’ll often try to relate to me by talking about shared interests such as sports, the weather, or current events. While a technological assistant will eventually be capable of this, adding in that information and getting to the nuances of conversation unrelated to the assistant’s task is probably going to be a tough challenge.
Some people reacted to Duplex with fear. A future of human-like assistants making robo-calls is scary for some. My livelihood isn’t staked on my ability to call people to accomplish tasks, so there is no direct existential threat for me here. But I can appreciate that many others’ livelihood is related to this kind of task.
The idea of computers making all of our calls for us is also scary and a tad dystopian. Shouldn’t calling people remain an inherently human activity? I, for one, am comfortable with technology stepping in. These aren’t the kinds of calls I relish making and would gladly eliminate them entirely. But if people stop calling mom and instead have the assistant do it, I’d argue we have progressed to the point of dystopia.
It isn’t the assistant’s ability to make the outbound call that excites me as much as it is the ability for the assistant to receive them. While businesses and scammers will rapidly adopt this technology as a tool to enhance the productivity outbound marketing and sales, the greater win is for everyone to have assistants to cope with the increasing amounts of inbound signals we have encroaching on our attention. I actually think it is fun to imagine hordes of computerized assistants carrying on quirky, task-oriented phone calls with each other. That’ll leave us humans to socialize and do whatever work remains to be done. It will be very nice when I have an assistant to automatically reject all inbound robo-calls, and to say no to people I don’t know who want to meet with me. We’ll be able to write out the logic for how our assistants interact with inbound stimuli.
Ultimately, both sides of this technology will have extremes we humans find unpleasant or unnatural. The controversy over Google Duplex is just a sign of things to come.
And while we’re on this theme of technology becoming more human, Boston Dynamics released a short video of the latest activity one of their robots has picked up. Give it a quick watch if you haven’t seen it already.
On to the news…
Artificial Intelligence and Machine Learning
Google I/O happened this week in Mountain View and gifted us with a lot of announcements and demos to digest. As I mentioned above, the most polarizing and interesting demonstration was Google Duplex — an addition to Google Assistant that allows the assistant to make a phone call on your behalf and secure an appointment. The polarizing element was that Duplex did this by impersonating human speech, and it did so brilliantly. The assistant used human speech quirks and could react to fairly nuanced replies to its inquiries. Watching the demo itself is probably the best way to appreciate it. The internet reacted by debating whether or not we should be told by the assistant that it wasn’t human. A Google representative later responded by saying that the assistant will announce itself as non-human.
A slew of other announcements were also noteworthy, though not as novel and intriguing as Duplex. Gmail will be gaining a Smart Compose feature which will offer auto-completion suggestions for email. Google Photos will have suggested actions to automatically touch-up images as well as recognize friends and suggest sharing photos with them. Google Assistant will be gaining the voice of John Legend, among others, later this year. The Assistant will also have a Pretty Please mode that rewards politeness. Lastly, Google Assistant Continued Conversations will remove some of the friction from longer interactions. I imagine this came about because developers were sick of saying “Hello Google” thousands of times during quality assurance.
I point out these many feature announcements to illustrate how machine learning is being deployed and how products using it are adapting over time. It is these changes that will inform us as to how the technology is being used and commercialized as well as where it is going in the future.
Over the past few weeks, a number of developer-oriented machine learning frameworks and services have been released that are worth noting. Google ML Kit is a new SDK for machine learning and is an addition to the Firebase developer platform. It features APIs for computer vision use cases such as image labeling and face detection. Mobile developers across both iOS and Android will be able to use this SDK, though I imagine the association with Firebase will not be ideal for many developers. Facebook AI open-sourced it’s deep learning framework, PyTorch. Lobe.ai entered the news cycle with its introduction of a visual interface for building, training, and shipping deep learning models. They struck a chord in the community as they received over 10,000 sign-ups for their closed beta in a single week.
The significance of these developer-oriented services is that they will simplify and accelerate the adoption machine learning. The technology is nascent and many smaller developers won’t explore it or experiment with it until robust, accessible tools become available. What remains to be seen is how much the largest tech companies will come to dominate this emerging technology given their advantages. Structured, high-quality volumes of data is a prerequisite for machine learning and many smaller companies don’t have it or worse, don’t have a pathway to obtain it. Using the tools provided by the tech giants may further entrench them as the market leaders.
Virtual and Augmented Reality
Google also unveiled some AR features at I/O that I’ll address here. Google Maps Street View is going to show directions through your camera. Google Lens also gained some noteworthy features. The first is the ability to highlight text seen through the camera and copy it. The second is called Style Match and it is the ability to point your camera at something and Google will attempt to tell you where to buy it. Finally, Lens operates in real-time now. These aren’t earth-shattering developments but they are very practical improvements to Google apps that improve or extend the user experience. These features suggest how Google will use AR in it’s core business.
The hardware releases this week were more interesting to me. The Oculus Go was released and garnered some very positive reactions. It is the first brand-name, relatively high-end, standalone VR headset to come out at an affordable price. While it is limited in degrees of freedom and display resolution, it has the right price and good enough features to attract customers. I’m very interested to see how well it does and I’ll be experimenting with it on my own.
Lenovo released their own standalone VR headset, the Mirage Solo. While it is technically superior to the Go, it has a weaker ecosystem of apps and content available through Daydream. Reactions seemed less enthusiastic relative to the Go. Lenovo also released the Mirage Camera, which is a camera that utilizes Google’s VR180 format to shoot VR photos and video. The price-point on the Mirage Camera seems very reasonable and I’m interested to see what impact it will have on VR content production. In my opinion VR is lacking both killer apps and killer content right now so facilitating more experimentation and adoption by content creators is worth watching.
Amazon Alexa and Smart Speakers
Amazon caught my attention with a critical change to its Alexa development ecosystem. They’ve opened up in-skill purchases to all third-party Alexa developers. While Amazon has reimbursed third-party developers, this came out of their own pockets and was indirect. Alexa app developers will now be able to emulate mobile applications in how they monetize, which I’m confident will attract more developers to the platform. The number of skills (another term for app) on Alexa devices has already been growing steadily.
The other announcement I’ll mention is actually a lack of an announcement. Facebook was reportedly going to announce their own smart speaker product at F8, but due to their ongoing issues stemming from Cambridge Analytica opted not to. I’m curious to see how they will seek to differentiate themselves in the increasingly crowded smart speaker market. With its ownership of the social graph, I imagine Facebook will lean into capabilities that the other companies can’t match. For example, messaging and status updates seem like a natural fit.
Podcasts and Digital Media
Apple released some eye-opening figures a couple of weeks ago that illustrate the rise of podcasts as a digital medium. As of March 2018, Apple Podcasts passed 50 billion all-time episode downloads and streams. If 2018 matches 2017, there will be somewhere around 14 billion downloads this year. As someone who adopted podcasts in 2006 to survive long commutes, I’ve long believed in podcasts as an emerging medium. Considering these numbers, I don’t think it is controversial anymore.
While Apple has done a lot to legitimize and grow podcasts as a category, Google hasn’t done as much. Google has significant influence over the adoption of podcasts, as most of the world is using an Android phone. However, the podcast experience on Android devices has not been a first-class experience unlike on Apple devices. But that is now changing with updates to how Google Search and Google Assistant handle podcasts. The highlights include the ability to play podcast episodes and subscribe to podcasts from search results, as well as aggregating podcast subscriptions in a home screen shortcut. Google Assistant can now play podcasts and will sync subscriptions and progress across devices.
Like many other Google products, Google News also received some love this week at I/O. In short, Google News is going to be more personalized and present content specifically for you. A new format called newscasts is available that uses natural language processing to collect content for requested stories. The Full Coverage feature will present even more collected content on a story, including a range of perspectives. Google seems to have learned some lessons from how Facebook aggregated and presented news stories.
Social Media
One of the surprising amounts for me from F8 was Facebook Dating. Despite the recent controversy over the handling of private data, Facebook is entering the dating category. This news immediately tanked the stock of Match Group, which operates many dating services including Tinder. However, Match went on to beat earnings and Tinder itself appears to be in fine health. I’m fairly pessimistic about this move by Facebook but I think it is a worthwhile experiment.
In other news, Instagram is gaining a slew of features. Businesses on Instagram will be able to offer calls to action such as booking appointments and purchasing tickets. Apps such as GoPro will have integrations with Instagram Stories, allowing you to share Stories from other apps. Video calling will allow individuals and small groups to chat together while using the app.
Gaming and eSports
Fortnite is dominating gaming right now and it isn’t hard to see why it is such a phenomenon. Alongside the massive success of Marvel’s Avengers: Infinity War, Fortnite launched a limited-time mode where players compete to wear the Infinity Gauntlet and play as a super-powered Thanos. Cross-IP collaborations such as this have been a hallmark of Japanese mobile role playing games, so I’ll be curious to see how well this does.
Blizzard is reportedly looking to expand the Overwatch League (OWL) with new franchises priced between $30 and $60 million each. I’m skeptical about how much interest these slots will receive at those prices. My impression of the OWL’s performance has been one of underwhelming performance, especially given the growth on the Twitch platform coming from Fortnite’s success (and the top streamer, Ninja).
DOTA 2 began the ramp up to this year’s The International with a new battle royale mode. The move to emulate the success of PUBG and Fortnite isn’t surprising and I fully expect this year’s E3 to feature a lot of forays into the battle royale genre.
Retrospective
One of my favorite podcasts is Acquired.fm, a podcast about tech acquisitions and IPOs. What makes it one of my favorites is the in-depth format in which they contextualize the history of the companies involved. This week I listened to the episode on the Microsoft acquisition of Forethought, the company which invented PowerPoint. There were many points of interest throughout the story and if you have an hour to spare, I’d encourage you to check it out.
It was surprising for me to learn that PowerPoint was not always a Microsoft product. I always just assumed it was. But it was an acquisition and remarkably good one at that. Microsoft acquired the company for $14 million in 1987. Over the last three decades that product became a cornerstone of the Office bundle and generated billions of dollars in revenue.
The original direction for PowerPoint was as a graphical user interface application to facilitate the production of presentations on overhead projectors. Back then, presentations were given using projectors and physical slides. Presentations had to be produced by corporate art departments. In the podcast they mention this scene from Mad Men where they give this style of presentation. It is hard to imagine a world where I’m not responsible for making my own presentations, where they have to be physically produced, and where presentations can’t be digitally shared.
PowerPoint, leveraging the graphical user interface and the personal computer, changed the paradigm and empowered individuals to create presentations, update them instantly, and share them over digital networks. Forethought was founded in 1983 by a team that believed in a future of computing where graphical user interfaces were the dominant paradigm. In today’s heady times, where the technology industry has become a monolithic industry touching every other industry and employing a significant fraction of the population, it is easy to believe that new technology is going to emerge quickly and commercialize quickly. The story of Forethought and PowerPoint is interesting to me because it is a story that took years to develop. Even an innovation as potent as the graphical user interface took years to gain traction and generate killer apps such as PowerPoint.
Thanks for reading!
If you’ve made it this far, I’ve done something right and I appreciate your time. You can follow me or message me on Twitter: @bguenther.