negationgame

Great mechanism design session today making design choices about 

This is trying to answer, "Does the negation game really need to only have negations?"

Great mechanism design session today making design choices about 

This is trying to answer, "Does the negation game really need to only have negations?"
https://youtu.be/HYWHGih4wvI?si=kkA9jPl7DV0HKkQD

The only problem is externality pricing.

Building https://negationgame.com

Great discussion. Here's what comes up for me: 

I'm going to introduce a metaphor to explain this. Think of a body. We're building a collective body of knowledge. Mapping the informational genome so that we can understand the DNA that causes this body to exist in the way that it exists. Once we understand the body, we can improve its health. (1/12)

Founder of Trick or Treat: https://trickortreat.co // https://brentschulkin.com

Everything you've been focused on so far is the skeleton. The bones. The most solid and strong part of the body. The skeleton defines the basic structure and shape of things. This is hard matter. It's strong because it has been tested. You can negate it, bet on it, stake against it, resolve it, and so on. Things are elegant, clear, and structured. (2/12)

But around that skeleton there's all kinds of different organs, tissue, and a lot of fat. The meat sacks of humanity are mostly made up of a more squishy, ephemeral substance that's hard to pin down and hard to make sense of. But this fleshy part is what most humans like to look at. This part represents the way they want to express themselves. (3/12)

The mission we're on with this product is mostly about creating the best skeleton. But most people are more interested in the fleshy bits. Luckily, they are deeply intertwined. To build a product that can become ubiquitously useful, I think you need to make it look, upon first glance, like a fleshy human. But once the product is embraced, people discover the skeleton and get interested in that. Your ladder of engagement starts skin deep and ends in the marrow. (4/12)

My intuition says that you want a UI experience which feels as simple as possible, as inclusive as possible, and as frictionless as possible. So imagine that the "top of the funnel" is that people can just create "beliefs". We could use a word like "argument" or "premise" or "idea" but I'm imagining the case where we lean out of this logical framing and into "beliefs" because everyone already has "beliefs" that would be easy for them to comfortably express, while that might not be as true for "logical arguments", etc. It's intentionally a bit fluffy. (5/12)

At the beginning, users see a fun, easy, and inclusive UX based on beliefs and how they are conceptually connected to one another. It's broadly appealing, bringing in more users (and ultimately liquidity). There's a 👍  button, a 👎  button, and a way to add links to other beliefs. You see interesting metadata and social signal, but no other functionality. AI does some of the heavy lifting here by automatically doing some basic sensemaking to map connections between the different beliefs, even before you necessarily introduce any of the market dynamics. It's a fun and easy way to express yourself by agreeing or disagreeing with beliefs and not (yet) overthinking it. Your activity auto-generates a profile which quantitatively describes your beliefs, plus an AI summary. (7/12)

Maybe all of these surface-level beliefs are represented in the interface as white boxes. Then someone going through the interface sees a belief that has a sparkly silver color! "Oooh what's this? My brain is intrigued by the variable reward of finding the belief with this special color!" And the silver color means that you can do more: bet on it, restake, and make money on it. You don't make any money on the white boxes, so now you get a sort of slot machine effect where people come back to look for silver ones they might make money on. (8/12)

Of course the goal would be to make the overall vibe of this thing about sensemaking, pro-social motivations, the search for truth, etc. And yet, this whole money making factor supplements that and helps people justify continuing to invest lots of their time into this society-wide sensemaking exercise. Once in a while, users will find the rare gold-colored beliefs. These represent beliefs where the question is easily resolved and you just have a normal prediction market. (9/12)

The skeleton (of silver and gold beliefs) is organized by the betting behavior itself. The negations and staking create the structure of the arguments and there is no room for AI to be messing around and influencing anything. But the outer layer (white beliefs) do not have any betting associated with them. They are too fluffy for that. White boxes include stuff like "2025 is going to be lit" or "Rico has a crush on Maria". So this is where the AI is connecting the beliefs to one another and handling most of the sensemaking. (10/12)

Such connections might be thought of as tentative "rough drafts". At any moment, the AI (or a user) could think of a better way to map connections to other beliefs, and there's some lightweight way to just edit the connections without any money being involved. And when people are engaging with the white boxes, the AI can automatically try to bridge people over to the silver boxes by suggesting that they bet on a silver box that seems related. That's how they might onboard users into the skeleton of logical arguments. This could also be where users are bridged from offchain to onchain (although maybe you just use account abstraction to give all users a wallet when they sign up anyway). (11/12)

So, getting back to the main question at hand, I do like the idea of giving users the ability to support beliefs, AND also adding a way to downvote them that is separate from a real negation. These new features would be totally meaningless from the perspective of defining the "skeleton" or finding truth. Their only purpose would be to create a fun/easy initial UX and let people quickly and easily build a graph of what they believe in, which creates a cool and shareable user profile that drives user growth. If you decide to embrace this dichotomy between the skin and bones, then I think you'll be able to make it understood that the action of upvoting or downvoting a belief is always "fluffy". There can be design affordances for making this clear, like maybe when you're interacting with a "silver" belief, the "Negate" button is silver but the 👍👎 buttons are still white. Thank you for coming to my TED talk. (12/12)

What if you design it so when you are at the level of the outer skin 100% of beliefs in the system allow for functionality X, but not Y or Z. If you go deeper into the muscle, beliefs at that level may allow for functionality X and Y, but not Z. Beliefs in the bones allow for X, Y, and Z functionality. With that in mind, imagine one way to define it could be that functionality X is creating, agreeing with, or disagreeing with a belief (disagreeing is different from a proper negation). Functionality Y could be the betting mechanism with real negation and re-staking enabled (the main thing you've been building). And functionality Z could be a traditional prediction market that only works with time-bound falsifiable things that can be resolved by a trusted oracle. (6/12)