USA – Wildfires used to have a season. From May to October, much of California—plus parts of most states west of the Rockies—would be on alert for dry conditions. Cold, wet November meant the threat was over.
Then 2018 happened.
This year David Tikkanen, a twenty-year veteran of the California Department of Forestry and Fire Protection, worked a thirty-four-day shift, from July 19 to August 21. “Back when I first started, it was the eleven-day mark,” he says. “When I started to cross the eleven-day mark on a regular basis, it was the twenty-one, twenty-three-day mark that I would start getting stir-crazy. And then after I did that for a long stretch of years, it was thirty-, forty-plus days.” His record is fifty-three.
The fire season grows. California faced its biggest fire ever this year—the Mendocino Complex, which lasted fifty-four days a hundred miles north of San Francisco—and collectively, California fires charred an area twice the size of Rhode Island. That was before the ongoing Camp Fire, which just became the deadliest in the state’s history.
Nevada lost 1.1 million acres. In Oregon, Idaho, and Utah, 1.8 million. And it’s not just out west: Nearly six hundred thousand acres burned in Oklahoma.
The earth’s changing climate contributed to the increase, meddling with precipitation patterns and hastening suppressing wildfires—which creates a buildup of natural fuels. But so has the push of people into nature. Wild states like Arizona, Colorado, and Utah have exploded in population. There are many ways to start a fire, and save lightning, they all begin with us. Firefighters in urban areas are a crucial line of defense, and they don’t have it easy, either. New synthetic building materials burn hotter and with higher toxicity than wood and brick.
Well, 2019 brings hope. A firehouse in Silicon Valley is pioneering ways to put drones to work for emergency workers. Scientists can better predict how a fire will burn. Jamie Hyneman, the former MythBuster, has engineered a firefighting tank that can be platooned and piloted remotely or even autonomously.
After his record fifty-three-day hitch, Tikkanen rolled into camp and told everyone he was done. The higher-ups gave him two paid days off. “I got to go home, pay the bills, kiss the wife, dogs, and kids,” Tikkanen says. Then he was back. In the future, maybe he’d be rotated onto drone duty. Or VR-pilot Hyneman’s tank. Maybe he’d telecommute. One thing’s for certain: In the future, if he’s in the field at all, he’ll be safer than ever. In this special report, Popular Mechanics has found exciting new methods and technologies that will help firefighters around the country for years to come.
March 2016. On a sunny Bay Area Friday, Jamie Hyneman is alone at M5 Industries, a tinkerer’s dream of a workshop in south San Francisco that was for thirteen glorious years the de facto headquarters of MythBusters, one of the most successful reality shows in television history, of which he was the walrus-mustached cohost. He’s alone here a lot these days, now that the show is over and the crew cleared out. He owns the space, and a lot of the MythBusters stuff is still here, pieces and parts and gear stacked in neatly labeled boxes from floor to ceiling.
He’s giving an impromptu tour to a couple of kids, showing them scary masks from his days doing Hollywood special effects and a mechanical spider that’s taller than they are.
And then, in a well-lit workroom toward the back, he pulls out a drawing of what looks like a tank. He gets serious—he’s always serious, but something in his tone reveals that this is special. The tank, he says, is unmanned. It is saddled with massive water tanks and outfitted with a remote-control system that will allow it to be piloted directly into wildfires—the kind that can rage in this state and others in the Northwest for many months each year—spraying water and saving the lives both of firefighters and homeowners, a remote-control robotic first responder.
Nothing like it exists, he says.
Just something he’s working on, he says.
One afternoon in 2003, Hyneman was standing on a film set on Hermosa Beach, in Los Angeles, wearing full scuba gear and holding a remote control. He was surrounded by a crew filming a 7 Up commercial, for which Hyneman had been hired to build a robot. “They wanted to have a 7 Up machine that was mobile and had tank treads on it and would bring 7 Up to you, and the thing gets a little aggressive by pushing its product on people by shooting cans out the slot,” he says. “So I got a little carried away and I actually made a fully automatic soda-shooting machine gun that actually propelled them about 400 miles an hour.”
Hyneman piloted the machine around the beach using the remote, shooting 7 Up at a surfer and crushing beach cruisers like cars at a monster truck rally. The broad treads hugged the sand, pivoting the robot to and fro at Hyneman’s whim. For the climax, he cranked the remote control and sent the vending machine through a volleyball net and straight into the Pacific Ocean. He was wearing scuba gear so he could retrieve it—he had made a deal with the production company that he’d get to keep the robot after the shoot.
“I GOT A LITTLE CARRIED AWAY AND I ACTUALLY MADE A FULLY AUTOMATIC SODA-SHOOTING MACHINE GUN.”
“I thought we were going to lose it,” he says. “But just for why not, once they called cut, I hit the stick all the way over for a second, and forward—and the thing came marching right out of the water.” He took the 7 Up robot back to his workshop in San Francisco, sand still clinging to those big, wide treads.
Jamie Hyneman’s Way of Working
Hyneman has one degree to his name, and it is in Russian linguistics, from Indiana University. Jamie Hyneman, creator of one of the most feared robots on Robot Wars, Blendo—a saucerful of pikelets that crossed a lawnmower engine with a wok with sharp objects—once ran a pet store. Jamie Hyneman was a Caribbean charter-boat captain. Jamie Hyneman was a chef.
He is not what you’d call a linear thinker, Jamie Hyneman. He wonders, and he arrives at something. When he looked at those treads, what he saw was surface area. And for some reason he can’t explain, he thought of grass fires, and the fact that the most efficient way to put out a fire was to stomp it with a wet blanket—wet it and smother it. He thought, Instead of some guy out there in the field dousing fire with a sprayer, maybe I could spray water on those treads with all their surface area and roll right over the flames.
This is not the traditional kind of leap to make: Maybe this thing that’s typically used for this could actually be used for that. Or at least, it’s not a leap anyone else had made. No one else envisioned a pack of autonomous tanks running down a wildfire. That’s why some people are inventors, and some people get a strange idea, the love child of a non sequitur and a lark, and dismiss it, because Well that’s ridiculous and might not even work and anyway I’m late for a meeting and what should we have for dinner tonight?
By 2015, Hyneman had an idea and was working on a prototype. What he didn’t have was financial backing. Then he met Palmer Luckey at a venture capitalist’s picnic in 2016. Luckey, of course, had founded Oculus VR in 2012, built a revolutionary virtual-reality headset in his parents’ garage, and sold the company to Facebook in 2014 for $2.3 billion. But within a couple years of his conversation with Hyneman, he was out of Oculus and getting ready to launch Anduril Industries. Anduril would be a different kind of defense contractor. The traditional model, according to Luckey, is that companies first secure a huge defense contract, then go try to build something. Anduril would instead first build things worthy of the Department of Defense, then sell them—by his estimation a better process.
“We talked a bunch about projects that we wanted to be working on, if we could do anything, and I told him about some of my crazier stuff,” Luckey says. “And he told me he had this idea for a remote-controlled firefighting vehicle that was self-cooling.”
Anduril wanted in on Hyneman’s tank in part because it was a perfect test for virtual reality as Anduril wants to apply it. Wildfires are chaotic: challenging terrain. Smoke. Heat that’s invisible to human eyes, even when it’s intense enough to cause a reignition. But what if terrain could be mapped by lidar and heat by IR camera and all of it paired to high-resolution maps, then stitched together into a seamless virtual environment that pilots could remotely navigate using a VR headset? Hyneman saw immediately how that could be great for fighting fires. Anduril saw how it could be great for other defense applications.
So, in 2017, Anduril bought in. Hyneman is a contractor. If he can hand off a physical object to Anduril, his work will be done.
Hyneman’s first idea was simply to run tracks around a giant water tank. Then one of his collaborators suggested looking at M113 armored personnel carriers, giant people-moving tanks the military has been using since Vietnam. They can carry and tow an excessive amount of weight and travel 40 mph. But there were questions about the legality of owning an armored military vehicle. That led them to a non-armored variant, the M548 tracked cargo carrier—basically a pickup-truck version. It had room in the back for water tanks, and because it was designed to be airlifted, it was relatively light at 14,500 pounds. Hyneman bought one in 2017 from an Army surplus store in Pennsylvania and stashed it in his shop in San Francisco.
When Anduril decided they wanted in, and that they would be designing a system of VR controls, digitizing the functionality of a fifty-year-old military vehicle became a necessity. A tank in a virtual environment is worthless if a computer can’t work its throttle. Hyneman enlisted Jim Newton, a former science advisor on MythBusters who went on to found the TechShop maker spaces, to devise, program, and build the network of sensors and microcontrollers that would give the tank a digital doppelgänger.
He also moved the M548 to an industrial fabrication shop in Oakland, Cooper Gray Robotics, that he’d worked with in the past. It’s the kind of place where every spare inch of shelf space—and there’s lots of shelving—seems to be stacked with scrap metal; where the crew is equally adept at customizing equipment for heavy industry and building fire-breathing zoomorphic vehicles for eccentric billionaires (not that they’re allowed to talk about it). They were perfect for working on the modifications Hyneman had devised, which were myriad—some straightforward, some eccentric. The M548 went up on blocks in the crease of two towering sets of shelves, maybe thirty feet high, which walled off different work spaces. On top, still for a dozen years, were the treads of the 7 Up robot.
Imagine this: You are a wildland firefighter battling an expanding blaze in the woods around a small town. The winds shifted suddenly overnight, and no one was able to warn the people. You’re in a wildland fire truck, which is pretty rugged, but there’s only one road into the town and it’s a wall of flames and the heat from the fire has ripped up the asphalt and layered the roadbed with smoldering debris. Even if you could get the truck to the trapped townspeople, how could you justify the risk to the driver?
doing things that are radically different,” says Luckey. When you have instead of a truck a tank and instead of a driver a remote pilot, you can charge through the wall of flames. You can crush smoldering debris like tinfoil.
“IF YOU GO IN HERE TO GET RESCUED, YOU COME OUT LIKE YOU’VE BEEN IN A SPA.”
This all puts an emphasis on heat resistance—which was a thorny problem.
“There were just too many surfaces we had to protect,” says Newton. “Too many things. You know? From wiring to batteries to controllers to electronic stuff and pumps. So many things. It was like, we can’t run a traditional heat-exchanger system on all these components. It’s just impossible. I mean, yeah, we could do it, but it would be crazy. So Jamie’s idea was, well . . . What if we just spray water all over everything?”
Hyneman calls it The Rainstorm: The inside of the tank is a maelstrom. A coolant that is a 50/50 mix of glycerin and water sprays continuously over the entire interior, so the vehicle can drive into a fire without its electromechanical guts becoming chitlins. Once the coolant absorbs the heat, it sinks down into the vehicle’s bilge—it’s amphibious, so it has one—which is located so the coolant’s heat is passed off to the thousand gallons of water in the tanks, after which it can be pumped back to the sprayers to rain again.
So what about the townspeople who have to climb inside the vehicle once it’s charged through the wall of flames to rescue them? They’ll be in the rainstorm too? “If you go in here to get rescued,” Hyneman says, “you come out like you’ve been in a spa.”
Jamie Hyneman is not precious when he talks about his inventions. The inside of the vehicle is deluged by “the rainstorm.” The vehicle’s electronics live in the “brain box.” The entire vehicle is wrapped in a tailored suit of heat-resistant fabric, which he calls “fire jammies.” The tube of fire jammies that wraps a deluge gun looks like an “elephant’s trunk,” and at one point, considering how people being rescued will pass through the fire jammies to get inside, he considers a circular hole that will cinch shut, like a kind of “anus.”
In May, there was a night when Newton stayed at the shop in Oakland until midnight, and returned way too early. A big demo was set for eight in the morning. He’d been programming the servopneumatic valves that would allow the tank to be remote-controlled. These were the six valves that would operate the tank’s real controls in the absence of human hands: moving treads, goosing the throttle. He’d built a circuit to operate the valves, tested them on a manufacturer’s test board with great success, then brought them to the shop, to the real vehicle—and nothing had worked.
Newton was able to replace one of the six with a valve from the test kit. One. Then he found out Hyneman had an extra valve in his shop in San Francisco, across the Bay from Oakland. He raced to get it. That gave him two working servos, enough to control one tank tread and the throttle.
And then it was time for the demo.
When their guest arrived, they positioned him on one side of the tank. They fired it up and used the remote controls to bring it to life. Tread turning. Throttle roaring. Of course, the other side of the tank was kaput.
“Oh, that’s really cool!” the guy said, none the wiser.
Hyneman’s Idiom II
The other thing about the way Jamie Hyneman talks about his inventions is that you can tell he is open to new ideas, and to success, and to failure. It’s in the way he tosses in qualifiers that begin “We may well . . .” to indicate a possibility that has occurred to him, but that wasn’t quite his intention, but that he understands may become the reality:
“I had the battalion chief for northern California down here, and he was the one who pointed out that it may well be that the most important use for this thing may be cleanup.
“So if we have something like this tank that is a range extender or a manpower extender, then that may well be the main purpose for it.
“And, you know, we may well have issues with the sensors.”
In late August, Hyneman, Newton, and two VR experts from Anduril gathered in the shop to start up the vehicle, which was nearly complete. It must have been twelve feet tall. Its surface covered in shiny aluminum. Its haunches two five-hundred-gallon tanks. Over its left and right front corners there are professional-grade water monitors, also known as deluge guns, which can raise and lower and pivot and drain those tanks completely in fewer than five minutes. It is officially called the Sentry.
Newton climbs up a ladder to the top. “Track check!” he says.
“Right track is clear,” one of the Anduril guys says. He walks to the other side. “Left track is clear.”
“I’m going to turn the air compressor on,” Newton says.
“Air on!” comes a chorus of voices, followed by the tenor thrum of the compressor.
“Fuel pump on,” Newton says.
“Fuel pump on. Engine front on. Remote control on. Starting engine.”
The roar of the diesel explodes off the walls of the shop. There’s a computer set up on a workbench, with a virtual-reality headset and two hand controls, each operating one side of the vehicle—treads and monitors.
When the joystick of the left hand control is actuated, a waterfall roar arises from the spinning of the left tread. Same with the right. When the virtual pilot looks left, the monitors pivot to follow his field of view. In the shop, the tank is up on blocks. In virtual reality, it’s in a forest of flaming trees. In virtual reality, the pilot watches from above as the tank stalks a flickering yellow prey. When the pilot pulls a trigger on the hand control—
Well, they haven’t actually rigged the monitors to fire yet. They’re still in a shop, after all.
What Is Actually Happening in His Head
“When I’m problem-solving with something, I have, effectively, a CAD program in my head that’s like a room that has specific qualities to it that I go to some deal of effort to populate. Textures and smells, something like that. With the intent of creating attachment points for my brain. Things aren’t invented in limbo or in a vacuum. They have a context. So I try to intentionally populate that context for them . . . I became interested in this just because I’ve got these big tracks left over from a commercial. What are they good for? And I start bringing them into the CAD program inside my head. What else could it do? Well, you could stamp out a fire, because that’s a lot more efficient. And then the whole thing avalanches. Less water resources, hotter, drier conditions— this becomes a thing. What could we do to minimize water use? What could we do to make sure you can optimize that? That’s what the Sentry is, over the course of a long period of time.”
In late September, a flatbed truck chartered by Anduril pulled into the shop’s lot. Hyneman drove the Sentry, now off its blocks, onto the trailer. The truck drove it down to Anduril HQ in southern California, where the VR will be perfected and field tests will begin.
For Hyneman, it’s been, roughly, ten years of thinking about it, and one year of building.
“Yep,” he says. “I’m done.”
About the Russian linguistics. It started with Slavic music. That was an offshoot from an interest in classical music. And that? “It had something to do with a girlfriend whose father was interested in classical music,” Hyneman says, in the shop to help make adjustments as the fire jammies are being tailored on a small Singer sewing machine. “That started me in that direction in the first place.” He pauses for a second, considering the long, strange chain of cause-and-effect, problem-solving analysis and revision, which some people simply call a lifetime. “Way to get access to the girl, I guess.”