DataDownload: Here’s how Amazon is setting up Alexa to orchestrate your life
DataDownload: Here’s how Amazon is setting up Alexa to orchestrate your life A weekly summary of all things Media, Data, Emerging Tech View this email in your browser
Hi Newsletter Friends…
This week, let’s go Sci-Fi. Air Taxis could be more than a Jetsons cartoon. I vote YES! And if you want to get a glimpse into the future of filmmaking — don’t miss the story about a group of eight cousins from Nigeria who make a Sci-Fi film with a broken iPhone. Seriously — Must Watch!
And — Amazon eats the world. Because they can.
And for a special surprise, read on to find out how a Steve-bot can change the way we consume information.
Notes, feedback, links, ideas. As always: Steve@NYCMediaLab.org
Steven Rosenbaum
Managing Director
NYC Media Lab
Must-Read
Inside Amazon’s Plan for Alexa to Run Your Entire Life
The entirety of FAMGA is aiming for ubiquity (or as Google calls it, ambient computing). While each of the tech giants has found a way to dominate a part of our lives, Amazon’s Alexa has noticeably cornered the smart home market. Alexa is able to control over 85k smart home products and execute over 100k skills. Now, Amazon is aiming for omniscience — anticipating and orchestrating your daily tasks.
Rohit Prasad, Alexa’s head scientist, recently outlined the personal assistant’s future at Web Summit in Lisbon, Portugal. The team will continue building a solid foundation, mastering basic NLP and video processing, but they’ve also begun tackling prediction and decision-making. One example will be Alexa’s ability to package skills:
“In order to follow up on a movie ticket request with prompts for dinner and an Uber, a neural network learns… to recognize which skills are commonly used with one another…. But reasoning is required to know what time to book the Uber. Taking into account your and the theater’s location, the start time of your movie, and the expected traffic, Alexa figures out when the car should pick you up to get you there on time.”
Inside the High-Stakes Race to Build the World’s First Flying Taxi
Air taxis are “very fast, very efficient, and low noise,” and in five years, a fleet of Lilium electric jets may take passengers from Manhattan to JKF in 10 minutes for $70 a ride… if all goes right. And there’s plenty that can go wrong.
Around 20 companies are competing with Lilium, including Kitty Hawk and Uber, and they have yet to receive government certification to fly commercially. Battery tech isn’t up to par, and the hardware isn’t ready to be scaled commercially. But this hasn’t stopped the investors: Daimler, Toyota, Porsche, Tencent, and plenty others have invested millions to support the flying taxi dream.
Click below for a beautifully produced video from Lilium covering their work:
For the Media
Get Me A Steve-Bot
Here’s a bit of digital-optimism. What if the current deluge of information isn’t going to last forever? What if in the future, we look back at 2019 and say…”boy, that firehose of tweets, and blogs, and video, and noise was truly terrible!” The solution might be data privacy and personal information robots — as I suggest here.
“The logic underpinning search engines is akin to a lesson from kindergarten: no question is a bad question. But what happens when innocuous questions produce very bad results for users?”
Microsoft’s Michael Golebiewski coined the term data void: “search engine queries that turn up little to no results, especially when the query is rather obscure, or not searched often.” Seemingly innocent, data voids can be manipulated to display disinformation — Golebiewski and danah boyd identified five types that could be abused: breaking news, strategic new terms, outdated terms, fragmented concepts, and problematic queries (learn more about these data voids in this Data & Society report).
Nieman Journalism Lab overviews a number of real-world examples, which are painful to stomach and difficult to get rid of, but still must be “systematically, intentionally, and thoughtfully managed,” like any security vulnerability.
AI Could Help Us Deconstruct Why Some Songs Just Make Us Feel So Good
University of Southern California researchers have recently mapped out how 74 unique song elements like pitch, rhythm, and harmony trigger brain activity, emotions, and physiological responses. They then used this data to train an ML system to predict how we respond to new music — research that’s part of a larger goal to understand how media affects us.
Once these responses are better understood, “we can try to productively use [the data] for actually supporting or enhancing human experiences,” said Shrikanth Narayanan, a professor at USC. That can mean generating playlists to help you in various contexts — sleeping, driving, or even post-breakup recovery.
What We’re Watching
Nigerian teens make sci-fi films with smartphones
An inspiring look at how The Critics, a group of eight cousins from Nigeria’s north-western state of Kaduna, have been creating viral short sci-fi films using just their smartphones, tripods, and green fabric.
Events & Announcements
Last Call: Cyber NYC Inventors to Founders Application
Deadline: November 9, Midnight
Apply for Cyber NYC Inventors to Founders Startup Accelerator! Cyber startups receive capital, mentorship, and access to Columbia, Cornell Tech, NYU, CUNY & NYCEDC networks. University startups (students, faculty, alumni) are eligible. Apply Here.
Event: She+ Geeks Out In NYC
Date: November 12, 6PM
If you’ve felt stuck as an individual contributor and unsure of what it takes to be a manager, this group of panelists are here to help. Features: Shifra Pollak (Chief of Staff to CTO, Datadog), Alisha Sedor (Sr Mgr of CX, Harry’s), Stephanie Killian (Engg Mgr, LinkedIn), and more. RSVP Here.
Event: Driving Corporate Innovation In 2019
Date: November 13, 8:30AM
Join NVCA and Bionic for an evening of dialogue on how to grow and innovate within large enterprises through the models and methods of venture capital and entrepreneurship. RSVP Here.
Event: Media Guru — Jean-François Decaux
Date: November 20, 6:30PM-8:30PM
Jean-François Decaux, CEO at advertising giant JCDecaux, will discuss the company’s strategy and operations, a culture of sustainability, “Phygital” (Physical + Digital), adtech innovation, and more. Register Here.
A Deeper Look
How Cities Are Using Hyperlocal Weather Forecasts to Save Money
For every inch of snow that falls in NYC, the city spends $1.8M to remove it. ClimaCell is one of the companies looking to use hyperlocal forecast data to determine when and where to lay down salt and sand, and how to best deploy snowplows to reduce this cost.
ClimaCell predicts temperature, precipitation, humidity, visibility, and lightning strikes for local areas like a neighborhood, airport, or city block, using existing infrastructure such as cell towers, street cameras, connected vehicles, and IoT devices.
How to Train Artificial Intelligence That Won’t Destroy the Environment With “unbiased” AI filed away under “myth” and ethics cemented as a key component of algorithm development (though, as The Outline points out, AI ethics is still not a requirement in graduate programs), the next circumvented issue is starting to rear its ugly head: AI’s environmental impact.
It’s easy to file away cloud-based processing because it’s quite hard to directly see the consequences. Fortunately, recent research is bringing the tech’s carbon footprint into the limelight:
- University of Massachusetts Amherst student Ananya Ganesh published a paper examining the environmental impact of deep learning models.
- Sasha Luccioni, a postdoctoral researcher at the Mila AI Institute, published a paper on Quantifying the Carbon Emissions of Machine Learning.
- In 2018 Kate Crawford and Vladan Joler released the Anatomy of an AI System website: “the Amazon Echo as an anatomical map of human labor, data and planetary resources.”
Transactions & Announcements
Neural Magic Gets $15M Seed to Run Machine Learning Models on Commodity CPU
