It started with a kid named Neil who was always first.
About eight months ago, Nathan T. Bernard fell into an alt-right rabbit hole on Twitter, and ended up on the page for Twitter user @NeilTurner_. Neil’s profile picture was a pouty mirror selfie of a skinny white kid, wearing what was clearly a Photoshopped “Make America Great Again” hat. The self-proclaimed warrior against #PoliticalCorrectness and #WhiteGenocide claimed to be from Mississippi, maintained a healthy 20-something-thousand followers and had been mentioned in publications like Fortune for his hateful, white supremacist commentary on Twitter. But most notably, he was often the first to reply to tweets from Donald Trump and Hillary Clinton — a historically coveted, high-exposure spot in online social networks and comments sections.
“I kept seeing him pop up continuously,” Bernard told me when I recently spoke to him at his apartment in Bed-Stuy, Brooklyn. “And I was like, ‘What the hell, there’s no way that this person is getting multimedia content as a first reply, day in and day out.’”
Bernard, a sardonic 26-year-old self-taught coder and app maker, began to examine the metadata of Turner’s tweets (and the metadata of tweets from others in Turner’s cohort, including Jason Bergkamp and the recently suspended @WhiteGenocideTM). He discovered that they ran a script typically used by Twitter bots that allowed them to reply within seconds to beat out millions of other users and be the prized first reply to both presidential candidates’ tweets on the platform’s mobile app. (Twitter’s mobile app organizes reply tweets chronologically, while the desktop version often reorders replies based on a number of factors, including user verification.) In other words, whoever was behind Turner’s account had enough technological wherewithal to game the platform so that his pro-Trump or anti-Hillary messages would be first. Even though Bernard thought that the content Turner and his friends tweeted was often “racist, anti-Semitic, or misogynistic,” he was impressed by their craft.
“So I started looking into Neil and trying to figure out who he was,” he said. “I got a little bit obsessed.”
That Bernard found himself lured into Twitter’s vast community of anonymous Trump supporters is understandable. In an already unconventional presidential election in which at least one candidate was accused of being the Zodiac Killer and in which a mere sniffle from Clinton can spawn a week’s worth of conspiracy theories, the platform has played a major role in shaping the 24-hour news cycle.
Energized by Trump’s crude Twitter persona, the most reactionary members of his base have banded together to form loosely coordinated networks of automated bots, aggressive trolls, and “cyborgs” like Neil, whose accounts post a combination of automated content and human contributions. “The sheer scale of bot activity being perpetrated on behalf of the Trump campaign is completely unprecedented in American politics,” Samuel Woolley, director of research at the Computational Propaganda Project, told Politico last month. In an environment where almost every niche interest group is attempting to break through the noise, these users have focused their strategy on flooding the platform and being tweet-adjacent to the most prominent voices in the campaign. By simply being the first account to appear below tweets from Trump or Clinton, they are able to piggyback off of their robust followings and earn millions of impressions. Sort of similar to how a nobody named Kim Kardashian went clubbing with the mega-famous Paris Hilton in the mid-aughts, and the paparazzi had to take photos of her: exposure by association.
But where some people saw a cheap tactic to influence the online conversation, Bernard saw an ingenuity he couldn’t help but admire. After spending months watching Turner and his cohort interact online, he built a bot that automatically replied to Turner’s every tweet with teasing messages along the lines of: “Hey Neil, what are you talking about?” Or: “Hey Neil, this content doesn’t seem very relevant.” He then made a website to gauge interest in a Turner-themed podcast and used his auto-reply bot to promote that website.
“I was like, you know what? If I learn about this guy, I could actually learn about all the other people that are following Trump,” Bernard said. “There has to be something more to this than just: ‘We’re racist.’”
Before the 2016 election, Bernard hadn’t been very politically engaged. He began paying attention around 2013, when Cruz read Dr. Seuss’s Green Eggs and Ham during a marathon filibuster against Obamacare. (“I was like wow, I pretty much just hate this guy.”) But while spending so much time on Twitter, he’d also bonded with a community of creators — like Vic Berger, Christopher Price, and Todd Dracula — who were generating Vines, GIFs, and memes based on the already-absurd presidential campaign. They encouraged him to turn his obsession with Neil into something, and eventually the #WhoIsNeil podcast was born.
The series is recorded and produced in Bernard’s living room, which is sparsely furnished with a couch, a small dinner table, a healthy house plant, and a copy of Carl Sagan’s Cosmos. A scrap of paper is taped over his laptop’s web camera. Since he started the show, there have been some security breaches. (Once, someone took control of his computer’s cursor and he rushed to pull out the plugs to his router and computer to stop the attack.) The podcast focuses not only on discovering Turner’s identity, but also demystifying the world of Twitter bots and the nuances of the white nationalist movement. Its tone is decidedly cheese-noir, full of long, jazzy interludes and narrated by Bernard in a slow and concentrated detective voice. Sometimes he re-reads entire DM conversations, representing both sides by using a voice distorter to represent a troll. I was surprised when he met me in front of his apartment sporting short blonde hair, a fuzzy gray sweater, and gray slacks — I half expected him to be wearing a fedora and a trench coat.
“I’m getting in fights on Twitter and making podcasts about it,” Bernard later said, when I brought up his recording persona. “It’s already like weird and kind of crazy in the first place. So, doing the voice distortion, and the detective investigative journalism voice — I don’t take myself that seriously.”
In early July, it dawned on Bernard that he could use the same promotional tactics as Turner to get more podcast subscribers. During a summer getaway with his family in Maine, he and his 24-year-old brother created a script that beat out Turner and his alt-right community for the first reply to both Clinton’s and Trump’s accounts. He began plugging his project.
“The first two months of me using the bot, I was using it every day and switching out content and being obsessive and stuff,” Bernard said. “And I got like 50 million views. So you can really fucking push the message.”
Bernard was by no means the first crafty coder to exploit this strange Twitter loophole for a little spotlight, but he certainly was one of the most dedicated. After friends like Berger, Price, and Dracula began asking to use his script to troll Trump, Bernard decided to build an app that allows anyone to easily plug in accounts they want to first-reply to, along with the content they intend to use to do it. When he showed me the interface during my visit, his queue for Trump was a playlist of tweets that included quotes from the candidate’s recently leaked Access Hollywood video, paired with suggestive photographs of Trump with his daughter Ivanka.
Bernard’s script was so effective that people began noticing. A field director for the Democratic Party reached out to ask about his podcast, and Turner’s identity in particular. Both of Trump’s adult sons and the Trump campaign’s director of social media, Dan Scavino, blocked him. Strangers began DMing him, offering him thousands of dollars to use the script. But most notably, Turner, Bergkamp, and @WhiteGenocideTM began direct messaging him on Twitter. They called him a “cuck,” warning him not to dox Turner and in turn threatening to dox him. They labeled him a shill for Clinton, banded together to flag his content on Twitter so he would be temporarily shadowbanned, and even criticized his lack of creativity in promoting the #WhoIsNeil podcast. It was clear that Bernard’s technical skills had earned their attention.
Through his investigation, Bernard had quickly learned that Turner was not the person from his profile picture. In fact, some research into the origins of @NeilTurner_’s profile photo yielded a handful of accounts with versions of the same mirror selfie (sans, of course, the very fake Photoshopped Trump hat). Any person or group of people could be behind the account, Bernard reasoned, and he figured the only way to learn who they were was to talk to them. Eventually, he began having debates with Neil and his friends via direct message, promising “no judgment,” in exchange for a civil discussion. Often they pushed conspiracy theories about Clinton’s untrustworthiness and health, or explained their theories of racial hierarchy. Occasionally, Turner asked Bernard for feedback on a website he designed or inquired about his bot.
“When we first started, it was pretty confrontational,” Bernard said. “I was trolling them, they would be pissed at me, we didn’t know each other. Then when we started DMing and everything, we all got chill. Sometimes we’d get in public feuds but we’d talk about those privately and everything. It was almost staged in a way. It wasn’t like we were actually fighting.”
Eventually Turner opened up to him, at the request that Bernard not reveal his identity on the podcast. Turner claimed he was a junior studying computer science at a public university in the South. Though he didn’t want to work in politics, he launched his first-reply scheme because, as Bernard put it, “he realized he could have an actual influence in everything.” Though Bernard was gratified to have established a mutual understanding with Turner, he remains skeptical.
“They could just be trolling me,” he said. “I always assume that nothing is true, I don’t trust anybody or anything.”
Meanwhile, Bernard’s first-reply app — inspired by these anonymous trolls — has the beginnings of an actual business. The day I visited, he was planning to meet with the VP of marketing at the daily fantasy sports site FanDuel. He’d given them a free week trial of his first-reply app, which the company used to quickly respond to the official NFL and Giants accounts during football games and earn engagement. The satirical news site, Cafe, had also tested out his services during the vice presidential debate. He and his brother charge rates comparable to Twitter’s in-house ads: around half a cent per impression. The rates vary for other metrics like link clicks, followers, or subscriptions, but they’ll offer a discount if a company wants to do a bulk upfront purchase of, say, $10,000.
“We’ve made a couple thousand dollars off it,” he said, shrugging. “I wouldn’t call it a real business at this point.”
Though their startup, called Impressiv, isn’t technically violating Twitter’s terms of service, the two have already set their sights toward building an ad-tech company that goes beyond the platform. Bernard envisions using a type of marketing chatbot to work on platforms like Facebook Messenger for everything from brand campaigns to elections. Like Turner, he too imagines his technology could have a real impact. He mentioned the Colombian hacker who was recently jailed for rigging elections across Latin America — proof, he says, that people might not realize how powerful social media is.
“I want to keep this thing going for the next four years so we can build a legit business. So the next four years from now, Hillary is running again, someone is running again, and I have a fucking team of tech people doing this,” he said. “And we’re blowing stuff up.”