Within a year his team had developed these models, as well as the tools for designing and deploying new ones faster. Before, it had taken Quiñonero’s engineers six to eight weeks to build, train, and test a new model. Now it took only one.
News of the success spread quickly. The team that worked on determining which posts individual Facebook users would see on their personal news feeds wanted to apply the same techniques. Just as algorithms could be trained to predict who would click what ad, they could also be trained to predict who would like or share what post, and then give those posts more prominence. If the model determined that a person really liked dogs, for instance, friends’ posts about dogs would appear higher up on that user’s news feed.
Quiñonero’s success with the news feed—coupled with impressive new AI research being conducted outside the company—caught the attention of Zuckerberg and Schroepfer. Facebook now had just over 1 billion users, making it more than eight times larger than any other social network, but they wanted to know how to continue that growth. The executives decided to invest heavily in AI, internet connectivity, and virtual reality.
They created two AI teams. One was FAIR, a fundamental research lab that would advance the technology’s state-of-the-art capabilities. The other, Applied Machine Learning (AML), would integrate those capabilities into Facebook’s products and services. In December 2013, after months of courting and persuasion, the executives recruited Yann LeCun, one of the biggest names in the field, to lead FAIR. Three months later, Quiñonero was promoted again, this time to lead AML. (It was later renamed FAIAR, pronounced “fire.”)
“That’s how you know what’s on his mind. I was always, for a couple of years, a few steps from Mark’s desk.”
Joaquin Quiñonero Candela
In his new role, Quiñonero built a new model-development platform for anyone at Facebook to access. Called FBLearner Flow, it allowed engineers with little AI experience to train and deploy machine-learning models within days. By mid-2016, it was in use by more than a quarter of Facebook’s engineering team and had already been used to train over a million models, including models for image recognition, ad targeting, and content moderation.
Zuckerberg’s obsession with getting the whole world to use Facebook had found a powerful new weapon. Teams had previously used design tactics, like experimenting with the content and frequency of notifications, to try to hook users more effectively. Their goal, among other things, was to increase a metric called L6/7, the fraction of people who logged in to Facebook six of the previous seven days. L6/7 is just one of myriad ways in which Facebook has measured “engagement”—the propensity of people to use its platform in any way, whether it’s by posting things, commenting on them, liking or sharing them, or just looking at them. Now every user interaction once analyzed by engineers was being analyzed by algorithms. Those algorithms were creating much faster, more personalized feedback loops for tweaking and tailoring each user’s news feed to keep nudging up engagement numbers.
Zuckerberg, who sat in the center of Building 20, the main office at the Menlo Park headquarters, placed the new FAIR and AML teams beside him. Many of the original AI hires were so close that his desk and theirs were practically touching. It was “the inner sanctum,” says a former leader in the AI org (the branch of Facebook that contains all its AI teams), who recalls the CEO shuffling people in and out of his vicinity as they gained or lost his favor. “That’s how you know what’s on his mind,” says Quiñonero. “I was always, for a couple of years, a few steps from Mark’s desk.”
The US has pledged to halve its carbon emissions by 2030
Is Biden’s pledge feasible? For now, there’s no specific road map to reaching this new target, but the White House is expected to release sector-by-sector recommendations later this year. To meet it, the US will have to radically overhaul its economy and drastically cut the use of oil, gas, and coal. Specifically, President Biden will need to push through a set of ambitious policies to spend $2.3 trillion to tackle emissions in high-polluting areas, such as cars and power plants, and accelerate innovation in clean energy and climate technology.
The reactions: Nat Keohane, head of the Environmental Defense Fund, an influential US NGO, tweeted that the new US target “meets the moment and the urgency that the climate crisis demands. It aligns with the science, pushes global ambition & accelerates the shift to a stronger, clean economy.”
“After years of US federal inaction to address its role in the climate crisis, today the Biden administration has presented all of us with significant reason for hope,” says Rachel Cleetus, policy director and lead economist in the Climate and Energy Program at the Union of Concerned Scientists.
How a tiny media company is helping people get vaccinated
More than 132 million people in the US have received at least one dose of a covid-19 vaccine, and as of this week, all Americans over 16 are eligible.
But while the US has vaccinated more people than any other country in the world, vulnerable people are still falling through the cracks. Those most affected include people who don’t speak English, people who aren’t internet-savvy, and shift workers who don’t have the time or computer access to book their own slots. In many places, community leaders, volunteers, and even news outlets have stepped in to help.
One of those groups is Epicenter-NYC, a media company that was founded during the pandemic to help neighbors navigate covid-19. Based in the Queens neighborhood of Jackson Heights, which was particularly hard hit by the virus, the organization publishes a newsletter on education, business, and other local news.
But Epicenter-NYC has gone further and actually booked more than 4,600 vaccine appointments for people in New York and beyond. People who want to get vaccinated can contact the organization—either through an intake form, a hotline, a text, or an email—for help setting up an appointment.
Throughout the vaccine rollout, the group has also been documenting and sharing what it has learned about the process with a large audience of newsletter readers.
We spoke with S. Mitra Kalita, the publisher of Epicenter-NYC, who was previously a senior vice president at CNN Digital and is also the cofounder and CEO of URL Media, a network for news outlets covering communities of color.
This interview has been condensed and edited for clarity.
Q: How did you start setting people up with vaccine appointments?
A: It began with two areas of outreach. First, when I had to register my own parents for a vaccine and found the process to be pretty confusing, I immediately wondered how well elderly residents, their friends and neighbors, manage this process. I just started messaging them.
The second was when a restaurant [from our small business spotlight program] reached out and said, “Do you guys know how to get vaccines for our restaurant workers?” Because I had been navigating some of this for the elderly, I started to help the restaurant workers. There started to be a similar network effect. One of the workers at this restaurant has a boyfriend who is a taxi driver; when I helped her, she asked if I could help her boyfriend; then the boyfriend texted me with some of his friends; and it kept spreading in that way.
Q: How is Epicenter-NYC filling gaps in vaccine distribution right now? What is your process like, and who are you helping?
A: We’ve had between 200 and 250 people reach out to volunteer. The outreach efforts range from putting up fliers, doing translations, and calling people to literally booking the appointments.
I don’t care if you’re a Bangladeshi taxi driver in Queens and your cousin is in New Jersey. We’re going to help both of you. A woman on the Upper East Side who’s 102 years old who is homebound and needs a visit is absolutely going to get Epicenter’s help.
What we’re doing now is continuing the route of connecting people to each other and opportunities. There’s a lot of matchmaking going on. We can sort through a list of about 7,500 to 8,000 people who said they need help, and then find places in proximity. We’ve become this wonderful marriage—a centralized operation that also embraces decentralized solutions.
Q: We know that vaccination rates lag in many communities that were hit the hardest. Why is that? What issues and barriers are people experiencing?
A: Just before the latest Johnson & Johnson pause announcement, I said, “We’re at a point where everybody remaining is a special case.”
I think we’ve leapfrogged to vaccine hesitancy without solving for vaccine access. We don’t see a lot of hesitancy, but we do see a lot of concerns over some issues. Number one would be scheduling. We’re dealing with populations that are working two, maybe three jobs, and when they say “I have this window on Sunday at 3 p.m. until maybe 6 p.m., when my next shift starts,” they really mean that’s the only window.
Q: People have been asked to prove who they are, where they work, and where they live in order to qualify for a vaccine. This was especially true when eligibility was more limited. How did you help people face barriers around getting the documents they needed?
A: New York State has been explicit in saying you can still get a vaccine even if you are undocumented. But that messaging doesn’t really match the on-the-ground reality.
Police in Ogden, Utah and small cities around the US are using these surveillance technologies
One afternoon, I accompanied Heather West, the detective who’d been perusing gray pickups in the license-plate database, and Josh Terry, the analyst who’d spotted the kidnapper with the Cowboys jacket, to fly a drone over a park abutting a city-owned golf course on the edge of town. West was at the controls; Terry followed the drone’s path in the sky and maintained “situational awareness” for the crew; another detective focused on the iPad showing what the drone was seeing, as opposed to where and how it was flying.
Of all the gadgets under the hood at the real time crime center, drones may well be the most tightly regulated, subject to safety (but not privacy) regulations and review by the Federal Aviation Administration. In Ogden, neighbor to a large Air Force base, these rules are compounded by flight restrictions covering most of the city. The police department had to obtain waivers to get its drones off the ground; it took two years to develop policies and get the necessary approvals to start making flights.
The police department purchased its drones with a mind to managing large public events or complex incidents like hostage situations. But, as Dave Weloth soon found, “the more we use our drones, the more use cases we find.” At the real time crime center, Terry, who has a master’s in geographic information technology, had given me a tour of the city with images gathered on recent drone flights, clicking through to cloud-shaped splotches, assembled from the drone’s composite photographs, that dotted the map of Ogden.
Above 21st Street and Washington, he zoomed in on the site of a fatal crash caused by a motorcycle running a red light. A bloody sheet covered the driver’s body, legs splayed on the pavement, surrounded by a ring of fire trucks. Within minutes, the drone’s cameras had scanned the scene and created a 3D model accurate to a centimeter, replacing the complex choreography of place markers and fixed cameras on the ground that sometimes leave major intersections closed for hours after a deadly collision.
When the region was hit by a powerful windstorm last September, Terry flew a drone over massive piles of downed trees and brush collected by the city. When county officials saw the resulting volumetric analysis—12,938 cubic yards—that would be submitted as part of a claim to the Federal Emergency Management Agency, they asked the police department to perform the same service for two neighboring towns. Ogden drones have also been used to pinpoint hot spots after wildland fires, locate missing persons, and fly “overwatch” for SWAT team raids.
This flight was more routine. When I pulled into the parking lot, two officers from Ogden’s community policing unit looked on as West steered the craft over a dense stand of Gambel oak and then hovered over a triangular log fort on a hillside a couple of hundred yards away. Though they’d never encountered people on drone sweeps through the area, trash and makeshift structures were commonplace. Once the RTCC pinpointed the location of any encampments, the community service officers would go in on foot to get a closer look. “We get a lot of positive feedback from runners, hikers,” one officer explained. After one recent visit to a camp near a pond on 21st Street, he and the county social service workers who accompanied him found housing for two people they’d met there. When clearing camps, police also “try and connect [people] with services they need,” Weloth said. The department recently hired a full-time homeless outreach coordinator to help. “We can’t police ourselves out of this problem,” he said, comparing the department’s efforts to keep new camps from springing up to “pushing water uphill.”