Week 2: Offense

Part of:
Truesight
Instructor:
Florian

Last week we discussed the exploit of System 1 injecting assumptions into our perceptions, before they even arrive in our System 2. We started to implement the habit of automatically shelving assumptions, so that we may draw conclusions correctly. This, we called “defense”.

As you may already have deduced from the title of this week’s session, we’re now going to go on the offense. We will discuss how this exploit is used against other people, and then we will spar. This will get the concept across even better, and will get you battle-hardened. Legend says after all: “you can’t bullshit a bullshitter”.

There’s not a whole lot of theory to talk about, as it’s the same as last week: We have the issue that we automatically pattern match and assign assumptions, meaning and intent to perceptions, before we become conscious of them. To induce wrong assumptions into others, you just have to provide a convenient pattern for them to match onto or avoid breaking their defaults. I’ll give you a few examples, so that the abstract concept becomes more clear:

You can put on a uniform to automatically get pattern matched as a specific type of worker. See the videos above. You can slap logos onto an email, use some official language, and make it look legit to get people to click links. You can leave out disadvantageous information on your CV (e.g. “worked part time at this job”) to get pattern matched to the more advantageous default (e.g. “full time work”).

You already know by now that people are eager to jump to conclusions. You only have to provide the runway. It’s important to note though, that this pattern matching exists for a reason: Most of the time, the pattern matched is the correct one. How people look is most often a strong indicator of their mental state, the tribe they associate with and their general philosophy. Mails that have all the details right usually are legit. Information left out is usually what one would consider to be the default.

To find opportunities for exploits is a little bit more tricky. Generally speaking, you can always find something in a gap between how things are and how they are intended to be. Video games are a prime example: There is usually some gap between how developers intended their game to work, and how it actually works, which results in speedrun exploits (if you’re interested, Hollow Knight is a fun example because most skips in there are just normal game mechanics).

If you study any source code, you will most likely find something exploitable at the spots where the programmer made wrong assumptions (which are most likely the way the used programming language diverges from all the others, e.g. Solidity’s lack of floats). In war or competitions you can trick your opponent into difficult positions by presenting them with situations they habitually respond to, so that you can exploit the response.

Material

Homework

  • Break into a
  • Come up with ways to use this exploit. Post one way into #course-content that hasn’t been posted yet.

Cohort

(30 minutes). One of your cohort mates defends a lie from the others’ questions with truths that are formulated in a way that keeps that lie alive. Some potential lies:

  • You’ve never been to city [XYZ] and have to defend the lie that you’ve been there
  • You’re a broke college student who is somehow at a high-end party and has to keep up appearances as a highly successful professional
  • You’re a millionaire at a low-end party and don’t want people to know that you’re rich

Set timers to make sure that everyone in the cohort has some time to defend.

Example. A has to defend the lie “I’ve been to Barcelona” without ever having been there. B and C ask questions like “what did you especially like about Barcelona?” and “what do you think about [famous building]?”

(20 minutes). Discuss the exploits others came up with in #course-content. How could you prevent yourself from being exploited in these contexts, if you were the defender in that situation?

Optional activity. After class, we'll be meeting up in the General voice channel to play Mafia Among Us (the social deception game). All Guildmembers are welcome.

The Different Levels of Lies

I find it very useful to think in terms of a hierarchy of deception skills, from the least sophisticated to the most sophisticated. The less sophisticated ones are harder to justify than the more sophisticated ones, which is reason enough to increase your sophistication level.

The least sophisticated form of deception is outright lying and fabrication of evidence. A slightly more sophisticated form of deception is misdirection. You don’t lie, but you foreground a pattern of true information that is likely to lead to false conclusions. Next you get withholding of information. You don’t lie or misdirect, but you don’t necessarily share any information that you don’t have to.

Next, you get equivocation, or sharing of information in ambiguous ways. This allows you to maintain plausible deniability against charges of lying, misdirection or withholding information, and relies on the predisposition of the other party to draw certain conclusions over others.

At the final level of sophistication, you get not-correcting-others. You don’t lie, misdirect, withhold or equivocate. But when others are drawing false conclusions that you could correct if you chose to (or missing inferences that are obvious to you due to your greater skill), you selectively choose not to help them out. If they make no mistakes and miss nothing, you’ve given your entire advantage away, though.

There is progressive minimalism and (social and moral) defensibility of means in this hierarchy. I am not a lawyer, so I’d be curious to hear from the lawyers among you, about the legal defensibility of different levels of deception. Outright lying is obviously perjury, while not-correcting-others would appear to be entirely defensible.

The more sophisticated techniques are harder to use. They take less energy (both to use initially, and to cover up if discovered later), but more skill. How you use all these techniques to compete in adversarial situations should be obvious. It is harder to see how you would use any of the skills to help another party. When you get good at not-correcting-others in helpful ways, you become a good teacher.

Excerpted from Be Slightly Evil, by Venkatesh Rao.

Community Notes

This section contains links and information that Guildmembers found helpful.

No notes available.

Write a Note

You must be signed in to write a note.