What he didn’t expect, back when he was helping define 4G, was that video streaming would explode in popularity the way it did and become the most obvious benefit of the new network technology. “4G made it possible to do video streaming, which was not conceivable before,” Grilli told Engadget at Google I/O this week.

The biggest users of 4G’s video streaming capacity today are Facebook and YouTube, according to Grilli, something the world wouldn’t have imagined back when the standard was being drawn up. The increased bandwidth paved the way for Snapchat, Instagram, Facebook and Twitter to add video posts to users’ feeds, and the explosion of mobile video spurred generations of flagship phones designed to display and capture high-quality footage.

During his 20 years at Qualcomm, Grilli has worked on technologies for UMTS, 4G and now 5G standards. He thinks 5G will spur the same sort of unanticipated trend the way 4G did for video streaming. “Maybe there will be a moment when some developer will come up with an application that doesn’t run well on this generation that will work well on the next to become the killer app,” he said.

As for the “killer use case” for 5G? That depends on developers, and that’s why this week’s announcement that Android Q is designed to let apps detect the 5G network performance matters. Basically, Qualcomm and Google extended an existing Android API tool that provides developers with network performance information to work with 5G as well. Since the calculation at the modem level is different for 4G and 5G, Grilli said, some work had to be done to enable Android Q to deliver this information to developers.

Armed with that data, developers can code different options into their app. “I can enable some new attributes of the app that I couldn’t use before because it wasn’t feasible,” Grilli explained. Say you find you’re experiencing high throughput and fast speeds. A video streaming app, for example, can choose to immediately push through high-res video at 60 fps or 120 fps, while loading just 720p clips at 30 fps on slower networks.

YouTube is one of the biggest users of this tool, Grilli said, as it checks a device’s downlink speed to adjust buffer rates. Grilli also gave an example of how game developers could use this feature to create a “5G mode.” Players on faster networks can compete against users with similar latency and speeds, instead of having an advantage over others on slower connections. This advanced mode could also serve up higher frame rates for smoother gameplay.

Grilli and his colleague Ignacio Contreras spent Thursday morning at a “Thinking in 5G” workshop with Google’s vice president of marketing, platforms and ecosystems Bob Borchers (and close to 100 developers). After Borchers and Grilli gave a brief presentation, developers broke into groups to come up with proposals on ways to use 5G. Each team’s favorite idea was submitted into a pool, and participants were given stickers to vote for the best suggestions.

While Grilli believes some popular uses for 5G will involve 4K and VR streaming, many of the ideas developers came up with at the workshop had to do with live sports experiences and broadcasting over multiple cameras.

There were some unique, intriguing ideas, too. One participant from Korea suggested using 5G and AR to leave video restaurant reviews that could be overlaid on top of the actual location. Future guests could stream the clip over 5G and see the reviewer inside the restaurant talking about their experience.

A woman with a passion for music suggested tapping 5G’s promised low latency for live jam sessions with remote instrumentalists. Playing in sync over existing networks has been challenging since there are delays between one musician hitting a note and it actually sounding on the other end. If 5G’s latency is low enough, it could enable real-time remote jam sessions with no lag.

Whether 5G truly changes our lives hinges on how we use it. Developers are a key piece of the puzzle, and at this Google I/O, they certainly appeared to be thinking about the next-gen technology. As networks come online and compatible devices begin to proliferate, the next step is clear: apps must evolve and make full use of 5G’s promised benefits. “We need to encourage and explain to them what 5G is and why they need to pay attention to it,” Grilli said. “If they don’t, someone else will and they’ll be left behind.”

Dark Mode

The most obvious new feature, which got a loud, welcoming response during the keynote, is Dark Mode. It’s not just a night-friendly mode that saves your eyes; it saves your phone’s energy, too. Go into Settings and search for “Dark” and select the Dark theme (as opposed to the Light alternative). This flips the color scheme from dark text on a white background to the reverse throughout the system and in supported apps.

It’s important to note here: not all apps support the dark theme. In this beta, at least, supported apps are mostly Google’s own, like Photos, Messages, YouTube and Dialer, but these already had a version of the dark interface. Calendar does a strange thing where its startup page will be black or white depending on the mode you’re in, but the actual interface is the same light background in both cases.

Google said that it’s developers who need to make their apps display a different color scheme when dark mode is activated, and it has tools to help them with that. One of them shows developers what their apps would look like dark mode simply by flipping its colors. If they like the way that looks, the tool can cut down many of the steps needed to add support for dark mode. Otherwise, there are other tools to help them figure it out.

For now, Dark Mode on the Android Q beta 3 is not yet truly system wide, but it’s a nice start.

Don’t expect the beta to be the same across all compatible Android devices, though. While the beta is available across 21 different phones, it’s up to manufacturers to decide which parts of the UI they want to implement. On a Huawei Mate 20 Pro, for example, I tried to enable dark mode but saw no difference when it was on or off. I also couldn’t find gesture navigation or dark mode on some of the other handsets on display, like an Essential PH-1 or a Nokia 8.1. Only the LG G8 I was using had all the new features. Basically, your experience with this beta may vary depending on the phone you have.

Gesture navigation

Another significant change is in navigation. With Android Q, you’ll be able to use a fully gesture-based system that resembles the one in the iPhone X and XS. That means you can replace the triangle, circle and square (back, home and all apps) with a swipe-based system. To be clear, you could already do this in Android P, but Q introduces some new gestures and completely removes the back button altogether.

This isn’t a mandatory or default mode (not yet, anyway). You can enable this in the settings by searching for “gestures.” When enabled, you’ll find a bar at the bottom that’s slightly wider and thinner than the “pill” in Android Pie. While you could tap the pill to go home, in the Q beta, the bar doesn’t do anything when pressed. It’s more of an indicator for where you should be swiping.

Flick it up quickly to go to the home page and swipe sideways on it to switch between your open apps. Go back by swiping in from the right edge of the screen. You can pull up the app drawer from the home or recent apps pages, though not if you have an app open. You also no longer summon Google Assistant by long pressing on the pill or home button; instead, you have to drag diagonally up from the bottom left or right corners of the display. Or, you know, say “OK Google.”

I was a little upset at losing the back button at first, but over time I realized I like swiping from the edge of the screen. Since I can do so from basically anywhere on the display’s edge, so it’s handy when I’m gripping my phone nearer the top or holding it horizontally. And if you don’t like it, you can always go back to a more familiar button-based interface.

To Google’s credit, though, it’s offering users a choice. When new Android Q devices start going on sale later this year, they’ll all have the classic three-button navigation system and the new swipe-centric system we’ve been talking about. And to be clear, I do mean all new Android Q devices will have these options.

To make the act of using Android feel as universal and consistent as possible, device makers like Samsung, Huawei, LG and the rest will have to offer these navigation systems on new phones, even if they’ve developed their own gesture interfaces. If you already have an Android phone that you eventually update to Q, your navigation system won’t change by default. (For now, that seems like the only way to keep the gesture controls that debuted with Android Pie.)

The benefits to this approach are pretty clear: if you ever switch Android phones post-Q, you probably won’t face a steep learning curve as you get acquainted with your new device. And by ensuring that all new Android Q devices have the new gesture navigation, Google is making it so that devices with smaller screens don’t feel nearly as cramped. That said, most phone makers of note have already cooked up their own gesture navigation systems in response to the (lackluster) option available in Android P, and the last thing Google wants to do is take that option away from users.

The company will strongly “encourage” its hardware partners to embrace Android Q’s navigation schemes, but the Samsungs, OnePluses and Huaweis of the world can continue to make custom skins with custom gesture interfaces for people who really want to use something different. The catch, according to a Google spokesperson is that those companies have to offer Google’s gestures right out of the box if they want to include some custom gestures.

Huang said the reason for this push to standardize navigation is because of app developers: their job gets infinitely more difficult if they have to worry about how different swipe gestures performed inside apps gets misinterpreted as something else. By insisting on one (or two) main modes of interaction, Google is trying to take some of the load off developers who might otherwise have to design their software with specific devices in mind. As far as Huang is concerned, that doesn’t lead anywhere good.

“If everyone does their own thing, Android apps are going to get worse,” he added.

And to be clear, Google hasn’t just sprung this on its hardware partners this week. According to Huang, the company reached out to major phone makers well in advance, and some OEMs have specifically asked Google to develop standard Android gesture controls. Well, wish granted, whoever you are.

With Android Q’s summer launch getting closer by the day, Google doesn’t have much time left to iron out the early issues with its new gesture controls. Based on what I’ve experienced in the last few days, though, Google has taken some significant steps forward when it comes to the ease and quality of interacting with Android. That’s only going to become more important as it tries to connect with its next billion users, so here’s hoping the coming months are well spent.

That won’t be the case for much longer. Yesterday, Google announced that a new, redesigned version of Android Auto would start appearing in cars this summer, and wouldn’t you know it: The company set up a 2020 Hyundai Kona on-site at its I/O developer conference, complete with new (but not quite final) software. As someone whose experience with Android Auto is mostly limited to rental cars, even I found some elements of the original Android Auto design maddening. Thankfully, this new software addresses a lot of those pain points, mostly because it doesn’t feel like it wasn’t designed for a car.

“After many years and a lot of customer feedback, we understood that users don’t want yet another experience they have to learn,” said Rod Lopez, Google’s Android Auto product manager, in an interview. “They would like to stick to the models they know and understand, particularly the one on the phone.”

With “Android” in the name, you might expect Android Auto to already act like a smartphone. After all, it takes what’s happening on your phone and translates it to run on your car’s center display. The reality wasn’t nearly as elegant, though: When you connect your phone to the car, Android Auto springs to life with a dated, card-centric interface that looks like it came straight out of Lollipop. If you were listening to something before plugging the phone in, it just stops. And perhaps most frustrating of all, you never really have an easy way to access all your Android Auto-friendly apps.

Those earlier version of Android Auto felt unpolished, which Lopez attributes to the team’s desire to build the proper foundation first. “I think back when we designed the first version of Android Auto: The emphasis was on making a product that was somewhat limited in functionality,” he said. “So that we could make sure it was the right experience for driving.”

The experience was right in a way, because it gave Google the opportunity to learn how to tune Android for cars displays. But it’s only now that Android Auto is starting to feel nuanced. Complete.