I didn’t have a lot to say about the election this year. To be perfectly level with you, this was largely because my friend Zooko precommitted to unfollowing anyone who was tweeting about it, and looking back from late November, I think he kind of had a point. I’d remarked on Twitter, well before the primaries, that the breakneck pace of overwhelming demands for attention was going to exhaust people. (“You can’t fatigue the Trumpenkrieg,” shot back Andrew Auernheimer. Good ol’ weev.) In private conversations, I ruefully predicted another Brexit, though apparently not with enough conviction to have an analysis of my own prepared. (I’ve been writing other things, but most of them have been code.)

Even now, about all I have to say is that clearly nobody, least of all the media, learned anything whatsoever from Protein World.

(“What the hell is Protein World?”)

Of course you don’t remember Protein World. That was April 2015, nobody remembers that kind of ancient history! So. Protein World is an online store that sells weight loss products and employs some of the most deviously brilliant marketers ever born. Their UK advertising campaign for spring 2015 involved purchasing these ads:

on the London Tube, sitting back, and waiting for knee-jerk anger to propel their graphic design in front of way more eyeballs than a £250,000 ad purchase ever could have done by itself. Needless to say, it worked. Never mind that two weeks into the ad’s three-week run, regulators pulled it; those two weeks of Twitter outrage and petition-shilling drove £1 million of direct sales Protein World’s way. The grapes of “They’ll win no awards for this” are extra sour when your “they” has already laughed its way to the bank.

(I pause to reflect, again, on social media’s own peculiar brand of amnesia. Flash outrages, though they may seem omnipresent from within the bubbles whose joint attention they consume, are often not that widespread in the broader scheme of things. What it is easy to excuse having missed, it is also easy to excuse forgetting about. From a more Huxleyan point of view, there’s no need to construct a mandatory memory hole when people are easily incentivized to build their own.)

Now, this may be a glass sword effect, extremely effective once but minimally so thereafter. When Protein World expanded their subway campaign to New York City as May 2015 ticked over into June, search traffic maybe rose slightly, but not distinguishably over the noise of the month following their UK campaign. That said, a TV ad campaign in January 2016 coincided with a visible increase in baseline search traffic — but the same thing happened in January 2015. Considering that the same kind of increase for “beach body” also happened in January 2014, when Protein World was just getting started, I’m going to chalk that up to new year’s resolutions rather than marketing savvy. Still, if you think for a moment that marketers aren’t paying attention, figuring out which tools are glass swords and which tools are reusable, you’re deluding yourself. (And for that matter, even if a glass sword is only useful once, it’s still useful that once.)

I leave the corresponding analysis of, for instance, breathless coverage of 300 edgelords in a hotel, as an exercise for the reader.

We live in a world where individuals, acting in concert, have the power to make literally anything important, and what do we choose to make important? The things that rise above the noise floor of our baseline level of annoyance, the more unusual or appalling the better. Of course people are going to hijack that tendency. Welcome to social engineering, the infosec praxis in which you, your thought processes, and your habitual tendencies, not your silicon or your software, are the attack surface.

We often find, in hacking, that other fields, generally perceived as less adversarial than their offensive-hacking counterparts, independently replicate certain aspects of the offensive domain. Fuzzing has equivalencies in test-driven development, and TDD is finding ways to incorporate fuzzing into its processes. Exploit development has equivalencies in type theory and logic solvers.

The mainstream equivalent of social engineering is, of course, marketing. Also PR.

The less reputable red-headed stepchild of social engineering is, of course, trolling.

When marketing and PR find ways to incorporate trolling into their praxis, we get Protein World. And we get the 2016 presidential election. And, apparently, everything that comes afterward.

The paradox of large countries is that no matter how carefully their machinery of state is designed to disempower populism, at a sufficiently large scale, populism becomes a necessary operational mechanism. If you need dozens of millions of voters to go to the polls for you in order to win, you have to find and convince those dozens of millions of people. If you don’t reach out to enough people, and don’t convince enough of the ones you do reach, that is very bad and you will not go to the White House next year. Blame whatever external forces you want, if your ground game isn’t there, you won’t get the turnout you need in order to win. This is life with 219 million eligible voters. It just seems very odd to me that anyone would think that continuing to broadcast ever more strident messages of fear to the same demographic that didn’t win the last election, while giving plenty of free coverage to something they claim to hate, will move the needle any further leftward.

George Lakoff has lamented, lately, that the left seems unable to grasp the notion of a conceptual frame. This is probably true, but it is compounded by what appears to be an intentional innumeracy. “But Hillary won the popular vote!” Which doesn’t matter, because the race is decided by a different function, namely the sum of electoral college votes. Win all the urban centers you like; if rural and suburban voters disagree with them and outnumber them statewide, those extra millions of votes in New York City and the Bay Area won’t help. It’s not a popularity contest among the voters, it’s a popularity contest among the states, and it doesn’t matter how much New York and California love you if too many other states think you suck. I’m not sure why it’s been so hard for the DNC to grasp this, but if they don’t, after 2020 there may not be much of a DNC left to do any grasping.

The other paradox of populism, of course, is that any sufficiently large popular movement attracts spotlights that bring whackjobs flocking in droves. Occupy got this in spades, with every kind of conspiracy theorist vying for the media’s attention on a leaderless movement. If anything, the alt-right is even less organized than Occupy was; as such, anything that looks like organization looks like a story, especially to a scope-insensitive audience. Forget fake news; where are the calls to action for not-even-wrong news? The grain of truth in that attention-hijack cocktail you’re slurping down doesn’t make it any better than the 100% artificial kind.

I confess I don’t see any easy way out of this. Fake news, and not-even-wrong news, hijack your attention because we got too good at detecting bots clicking on ads, to the point where it became easier for sites to compete for real users’ attention. Yay, I guess? But new problems carry with them new complications. You can’t solve a problem that exists because of a cognitive bias — a heuristic that developed so that its user can expend less effort — by asking people to expend more effort. (Tried it, only works in exceptional cases.) This hugely constrains the possible solution space, even though there’s a significant Pareto improvement at simple actions like “not sharing things you haven’t read.” That sweet spot between “not enough attention to read” and “enough attention to relay” is where troll marketing scores big.

If you’re worried about things becoming “normalized,” worry about the normalization of troll marketing. Particularly the fact that it’s already happened.