E.T.’s Home Phone

Around the time the movie Contact came out in 1997, Kevin D., a governmental IT support and procurement employee in Toronto, saw a notice on a technical news site about a piece of software that was being developed by researchers at the University of California, Berkeley. The scientists were interested in SETI, the Search for Extraterrestrial Intelligence, and courtesy of Contact, so was Kevin D. The moment he heard about the program that would eventually come to be called SETI@home wasn’t as dramatic as Jodie Foster’s portrayal of Dr. Eleanor Arroway discovering a message sent across the universe, but it would make a major impact on the next two decades of his life. It also signaled the advent of a productive and unprecedented citizen-science project that continues today, 20 years after it launched in May 1999.

Kevin D. aspired to be a scientist as soon as he could read, but financial difficulties forced him to drop out of university, which put an end to that dream. “I could have probably gone with student loans and a few years of eating ramen, but I wasn’t in the right frame of mind anymore,” he says. “SETI@home and other distributed-computing projects have filled that need nicely, allowing me to contribute to science on a scale that would have been unimaginable just a few decades ago.”

SETI@home was the brainchild of a UC Berkeley grad student named David Gedye, who came up with the concept of using personal computers for scientific purposes in 1995. “That was the point where a lot of home computers were on the internet,” says Berkeley research scientist David Anderson, Gedye’s grad advisor and the cofounder of SETI@home. “Also the point where personal computers were becoming fast enough that they could potentially do number-crunching for scientific purposes.”

Gedye thought using computers to comb through data recorded by radio telescopes in search of signals sent by intelligent extraterrestrial life would both appeal to the public and demonstrate the potential for public participation to boost the scientific community’s processing power. He and Anderson joined forces with multiple partners in the astronomy and SETI fields, including Eric Korpela, the current director of SETI@home, and Dan Werthimer, the Berkeley SETI Research Center’s chief scientist. Werthimer was a SETI veteran who had been hunting for alien life since the 1970s and oversaw the SERENDIP program, which piggybacks on observations that radio astronomers are already conducting and scours the results for evidence that E.T. is phoning our home. SERENDIP supplied the incipient SETI@home with data from the venerable Arecibo Observatory in Puerto Rico, which until 2016 featured the world’s largest single-aperture radio telescope.

Fueled by $50,000 from the Planetary Society and $10,000 from a company backed by Microsoft cofounder and SETI enthusiast Paul Allen, Korpela and Anderson started designing software that would split that data into chunks that could be distributed to personal computers, processed, and sent back to Berkeley for further analysis. By the spring of ’99, SETI@home was ready to launch, despite the difficulty of making it compatible with all kinds of computers and dealing with pre-broadband internet. But its creators weren’t prepared for the outpouring of public interest that propagated through word of mouth and posts on forums and sites such as Slashdot.

“The biggest issue was not the people on dial-up connections,” Korpela recalls. “It was just the sheer number of people that were interested in SETI@home. When we started SETI@home, we planned or thought that maybe we could get 10,000 people to be interested in doing this. The day we turned it on, we had close to half a million people show up.”

In 1999, the public portion of the internet was new enough that going viral was a nearly unknown phenomenon. But Korpela says that within a month or two, SETI@home had attracted a couple million active users, which overwhelmed the modest equipment underpinning the project, causing frequent crashes. “We were planning on running our servers from a small desktop machine,” Korpela says. “That didn’t really work.” Sun Microsystems stepped in to donate more powerful hardware, and SETI@home users helped the perpetually underfunded program defray the cost of bandwidth, which was expensive at the time. In 1999, Korpela says, Berkely was paying $600 a month for each megabit per second, and SETI@home was guzzling about 25.

On the plus side, the uptick in processing power was immediately apparent. “The main benefit of the SETI@home–type processing is that it gives us about a factor-of-10 increase in sensitivity,” Korpela says. “So we can detect a signal that’s 10 times smaller than we could just using the instrumentation that’s available at the radio telescope.”

As SETI@home spread, a few of its more zealous acolytes ran afoul of the workplaces where they installed it, which the program’s creators advised users not to do without permission. In 2001, 16 employees of the Tennessee Valley Authority were reprimanded for installing the software on their office computers. (I know the feeling; my mom wasn’t pleased about the electricity costs she claimed I was incurring when she spotted the screensaver on my own early-2000s PC.) In 2002, computer technician David McOwen faced criminal charges and was ultimately put on probation when he installed SETI@home at DeKalb Technical College in Atlanta. And in 2009, network systems administrator Brad Niesluchowski lost his job after installing SETI@home on thousands of computers across an Arizona school district. (Niesluchowski, or “NEZ,” still ranks 17th on the all-time SETI@home leaderboard for data processed.) Korpela has made several SETI@home sightings in the wild, including on point-of-sale cash registers and, once, on a public computer at an Air Force base (which wasn’t Area 51).

Over the decades, SETI@home’s user base has dwindled to between 100,000 and 150,000 people, operating an average of two computers and six to eight CPUs per person. But the remaining participants’ computers are hundreds or thousands of times more powerful than they were in 1999. “When we started, we designed our work units—our data chunks going out to people—to be something that a typical PC would be able to finish computing in about a week, and a current GPU will do those in a couple of minutes,” Korpela says. SETI@home is now available via an Android app that’s used by about 12,000 participants, and even smartphones smoke turn-of-the-century desktop computers in processing speed.

The SETI@home software has evolved along with the hardware that hosts it. In the early years, the program ran as a screensaver, which served multiple purposes. First, screensavers were popular, so the software filled a need. Second, the graphical representations of the program’s activities fed users’ scientific curiosity and reassured them that the program was working as intended. And third, it functioned as eye candy that entertained users and caught the attention of anyone within visual range. Now that screensavers have fallen out of favor and more people prefer to turn off their monitors or computers when they’re not in use to save power, Anderson says, “We’ve kind of moved away from the screensaver model to the model of just running invisibly in the background while you’re at your computer.”

A shortcoming of the original SETI@home software led to a much more significant change—and, indirectly, the greatest legacy of SETI@home, at least so far. In the program’s initial form, the signal-processing logic and the code that handled displaying the screensaver and receiving and transmitting data were a package deal. “Each time we wanted to change the algorithms, to change the scientific part, we had to have all of our users download and install a new program,” Anderson says. “And then we would lose some fraction of our users each time we did that.”

The solution was separating the science part from the distributed-computing part by building a platform that could update the algorithm without requiring a reinstall. Better yet, that platform could act as a conduit for any number of alternative distributed-computing efforts. In 2002, Anderson built and released that system, which he called Berkeley Open Infrastructure for Network Computing, or BOINC.

SETI@home, which migrated to BOINC in 2005, has thus far failed in its primary purpose: to detect intelligent alien life. But it’s succeeded in its secondary goal of demonstrating the viability of distributed computing. Other researchers have emulated that model, and BOINC, which is funded primarily by the National Science Foundation, is now home to 38 active projects that are doing useful science, including investigating diseases and identifying drugs that could combat cancer, modeling climate change, and searching for phenomena such as pulsars and gravitational waves. Research conducted by BOINC-based projects has generated 150 scientific papers (and counting), and the network’s collective computing power—about 27 petaflops—makes it more powerful than all but four of the world’s individual supercomputers. Anderson, who believes volunteer computing is still underutilized by the scientific community, says it’s especially “well suited to the general area of physical simulations where you have programs that simulate physical reality, which scales anywhere from the atomic level up to the entire universe.”

In other words, the sky is not the limit.

For SETI researchers, the space beyond the sky has kept its secrets close. Modern SETI experimentation started in 1960, when astronomer Frank Drake (of Drake-equation fame) scanned two nearby Sun-like stars for signals of intelligent origin. Drake’s search, of course, came up empty, yielding only a false signal that turned out to have emanated from a military project. That set the tone for the next several decades: With one well-known unexplained exception, every candidate for evidence of intelligent life has led back to our own species.

“It’s hard to get too excited, because we detect lots of signs of intelligence and it’s all human intelligence,” Korpela says, adding, “If we do see something that’s interesting, it happens so often that we can’t really get excited about it until we’ve looked at it for a few weeks, and then invariably it turns out to be something that was not interesting at all.”

Korpela isn’t discouraged that the decades-long search hasn’t struck life. Our radio telescopes are still of modest size, and detecting a signal would require E.T. to point telescopes at us continuously, which means that the aliens would need to know we’re here. Because human radio transmissions started only a century or so ago, any life farther than 100 light years away would have no way of knowing we’re here. And any life farther than 50 light years away wouldn’t yet have had time to send a message back to Earth.

The universe is so vast that the failures of our first feeble efforts to find life hardly make it less likely that there’s life to be found. Korpela puts the odds of finding a real signal at about 1 percent per year, which translates to a 50/50 chance of success in the next century. “Before SETI@home came around, we had looked at one drop of water in the ocean,” Korpela says. “And since SETI@home came along, we have looked at about a small glass of water out of the ocean. … We’ve still got a whole lot of ocean to sift for E.T.”

The SETI community is trying to sift more efficiently. In 2016, the Berkeley SETI Research Center began Breakthrough Listen, its most comprehensive sky survey yet. That project generates so much data that SETI@home handles only 1 percent of it. Anderson is also an estimated six months away from completing a project called Nebula, which he says is “basically the second half of SETI@home.”

For its first 20 years, SETI@home has been dissecting data to identify blips of energy at particular frequencies. Many of these blips, however, are produced by radio-frequency interference, or human-made noise. Nebula will tackle the tricky task of filtering out the interference from radar, cell phones, and other devices, reducing false positives. It will also accept the challenge of, as Anderson says, “looking through this huge database of momentary signals and identifying ones that are potentially repeated”—beacons that come from the same place in the sky over a long period, potentially pointing to an extraterrestrial sender.

Even if Nebula doesn’t find anything, Anderson hopes it will establish that if there is a meaningful signal, it must be below a certain strength. “It’s a negative result in a sense, but it at least proves that we did something that nobody has to go back and do again,” he says. That would push the field forward, even if, as Korpela quips, “No one gets the Nobel Prize for not finding something.”

There’s more than one way to track down E.T. The SETI Institute, a not-for-profit research group unaffiliated with the Berkeley Center, employs real-time monitoring: Instead of mooching off of other astronomers’ telescope time, the SETI Institute controls radio telescopes directly. When it picks up a signal, it can investigate it immediately, reorienting the telescope to see whether the signal is really interstellar or is only interference masquerading as the real deal. Seth Shostak, the SETI Institute’s senior astronomer, says, “We wanted to follow up on signals immediately, within a minute, because you don’t want to run the risk of, well, this was a beacon, and then E.T. got bored or went out to lunch or who knows what and then turned off their transmitter, pointed it somewhere else.”

That preference for rapid responsiveness comes with one trade-off: The SETI Institute can’t study signals in the same depth that SETI@home can weeks or months after the fact. “We have to make a decision, what are the best sorts of signal signatures to look for,” Shostak says. “And they can look for a wider range of those, because they have all the time in the world and a lot of the processing power in the world.”

The two-pronged approach makes the groups complementary. Shostak says he’s often asked whether SETI@home has helped him. “Of course the easy answer is no, because it’s not our project,” he says. “But that’s not true. It’s like asking an astronomer in the 1980s, ‘Hey, does Carl Sagan help you?’ And of course if that guy doesn’t work with Carl Sagan he’ll say no. But in fact Carl Sagan has helped him, because Carl Sagan has increased the interest in the field of astronomy. … SETI@home has done much the same for SETI.”

Even so, the field faces financial constraints. Congress canceled NASA’s SETI program in 1993, and the nonprofits that picked up the slack are always searching for funding in addition to alien life. SETI Institute research scientist Jon Richards says that SETI@home “does well with the resources it has, but it needs to grow larger,” adding, “I would support a big funding initiative.” Richards notes that even if new cloud-computing companies could replicate SETI@home’s power, they couldn’t replicate the PR value of making the public part of the search.

After 20 years, the excitement surrounding SETI@home has ebbed. The software is now a source of ’90s nostalgia, but like a lot of ’90s acts, it hasn’t had a hit in the 21st century, aside from spawning BOINC. “If we tried to make a post go viral now, it would be, ‘Oh, SETI@home, I remember that,’” Korpela says. “And people would just ignore it and not re-sign up.”

Yet some loyal users’ support has never wavered. Kevin D. (a.k.a “Mr. Kevvy”) installed SETI@home the month it launched, and his participation hasn’t slackened since. He ranks 14th on the all-time leaderboard, an impressive placement considering that some of the entries ahead of him correspond to institutions with many more machines in harness. Now that his two sons have grown up and moved out, he, his wife, and his cat share a “cozy-to-the-point-of-crammed” one-bedroom apartment with 11 devices running SETI@home: seven Linux desktops, one Windows desktop, and three Android phones.

Kevin D. is an all-around SETI@home powerhouse: He’s helped organize fundraisers, made matching donations, moderated forums, and dutifully upgraded his hardware. Three-fifths of his current processing contribution comes from the flagship of his SETI@home fleet, a Linux desktop that he recently souped up with four liquid-cooled GPUs. Even though it’s performing more complex analysis than the software demanded decades ago, that computer completes work units at approximately 3,600 times the speed of the PC he was using to run SETI@home in early 2000s, when he took the screenshot below.

As one would expect given his dedication, Kevin D. is disappointed that SETI is sometimes dismissed as non-serious science. When Arizona school superintendent Denise Birdwell announced Niesluchowski’s resignation in 2009, for instance, she said, “We support educational research and we certainly would have supported cancer research; however, as an educational institution we cannot support the search [for] E.T.”

What keeps Kevin D. going is his idealistic vision of what could occur if sentient life of extraterrestrial origin were discovered, which he believes might be the most momentous event in human history. “We would be more united as one human species, finally,” he says. “No matter our differences from even the most disparate human on Earth, they would be but a thousandth as different from us as any intelligent extraterrestrial. Perhaps then there will come a day when calling another human being ‘alien’ is socially unacceptable.”

Foster’s character in Sagan’s story for Contact, which inspired Kevin D., was based on Dr. Jill Tarter, the SETI Institute’s cofounder and chair emeritus for SETI research. Tarter values SETI@home’s role in expanding the public’s involvement in science. “No question that SETI@home put citizen science with distributed competing on the map,” she says.

Tarter is still partial to the SETI Institute’s approach. “I prefer to analyze incoming data as close to real time as possible in order to follow up immediately,” she says. “On the other hand, SETI@home has an enormous amount of time on the sky at certain frequencies, but with delayed analysis. We’ll know what was the better strategy when one of us succeeds.”

If either one does, the other will be far too happy to have heard from E.T. to hold a grudge. And if SETI@home is the first to find a signal, humanity may owe users like Kevin D. a debt.