Russian “troll” describes interview for meddling in elections | Hacker News Reader

Russian “troll” describes interview for meddling in elections

recode.net

21 pointsboulos9 comments

At the end he makes an interesting point: that this worked in the US not because it has a lot of stupid people (which Russia presumably has too) but because it has a lot of stupid people who have not been lied to too much.

“Who really reads the comments under news articles, anyway?,” he said when asked whether the tactics worked in Russia. “Especially when they were so obviously fake.”

“But for Americans, it appears it did work,” he continued. “They aren’t used to this kind of trickery. They live in a society in which it’s accepted to answer for your words.”

When I moved to United States from a semi-communist country this was one the things that I found amazing about American Society. If you someone tells you something people just belive it. For example even CEO's have fake resumes[1], because people just believe what you tell them. In the country where I grew up, people are so used to lying and scheming that people just assume everyone is a cheat by default unless proven otherwise.

It was just an amazing feeling to be a part of higher trust society like USA. Things are just easier and less stressful, you are not constantly worried about someone trying to get you.

1. http://www.businessinsider.com/successful-executives-who-hav...

Having grown up in North America, I think I can say that the kind of higher trust society you're speaking to only exists at certain strata.

If you were born into a lower strata, upward movement will sometimes be viewed with suspicion. If you were born high, you'll not likely sink below your level.

In short, I think that's a privilege of the upper-middle classes and above.

Everyone here is a cheat too, it is just more pathological. People don't even KNOW they're lying.

That's one explanation, but it seems unnecessary. Trump was on TV all the time. He was never not on TV, for the entire 1.5 years leading up to the election. Is there any candidate in history who would not have had better results than he did, given such a massive structural advantage? The question isn't why he won, but rather why his margin of victory was so small. (Possibly because he is an odious vainglorious born-on-third-base lump of shit? But when have we cared about that?)

TV people will protest that he had to be on TV, because he said such outrageous things. That speaks to the fact that Trump hacked TV, much as Russians are alleged to have hacked social media. It seems relevant that Trump's largely old voters get more news from TV while Clinton's largely young voters get more news from social media. Yet investigations concentrate on social media: perplexing!

In case it matters and in case someone couldn't tell, let me observe that I voted for someone else and indeed would never vote for Trump. That's no excuse for believing the obvious bullshit being peddled now.

We miss that during the 80's, psychological operations were being honed, sharpened, and primed for precise effect in the mainstream of marketing and commercialization. Not only did it work, it worked so extremely well that entire departments bloomed around the practice thereof.

The weapons and technologies of mass persuasion and reinforcement were paid for and researched under a for-profit system. It took a while to realize that there were other areas they could be readily adapted for further profit.

Non blogspam link: https://www.washingtonpost.com/news/worldviews/wp/2018/02/17...

The interesting bit to me is that the guy didn’t make the cut. ML based methods are way behind humans at this point, but how far off is the day that we could train something to pass? 5 years? 3? Less? Especially for the sort of short-form comments on news articles, it seems possible to me to at least insert high-confidence “sounds American, but false” today-ish.

So this is an entirely out of my ass theory but here goes...

The reason something like ML is failing here is because what the russian trolls are trying to do is fundamentally change how discourse occurs, how people in the US receive and process information, and the tone of the conversation. ML is great are reproducing things that already exist. That's why they get accused of being racist...because they are reproducing patterns in data. The russians were trying to cause things to happen by changing the conversation.

That's fundamentally different and I haven't seen a lot of ML systems where one can dictate a goal like that and have the system work out the response. From what I know, that seems like one of the differences in the 'older' schools of AI and the newer schools that let the field actually have such an explosion of success nowadays...they got way more comfortable with backward vs. forward looking learning.

Reply on news.ycombinator.com