Debate vs. Determine. Consensus vs. Truth


#1

Two quotes from Friday’s call:

“To get to the truth of something you have to have an educational debate about it […], you have to be willing to see all sides of the argument […], so we can collectively agree what is the right thing we want to believe”.

“Trustory is not only a platform to figure out what’s true and not, but also a platform to share knowledge and think critically.”

Questions:

  1. If something can be debated and if there are multiple sides of argument, can an objective truth really be determined? Isn’t that arriving at a consensus view rather than a universal truth?

  2. Is the main added value and goal of Trustory to determine truths, or to encourage debate? Ideally the debate would lead to a/the truth, but is if there is too much debate and competing views on a claim does that not suggest that there might not be a clear, single truth?

  3. If so, should there be a buffer zone in the voting outcome which is ‘undetermined’, i.e. if 52% say something is true and 48% say it’s false, is it true? What about 60/40? 70/30?

  4. Is it more important, and easier, to debunk falsehoods than to determine truths? “The sky is blue” was mentioned as a truth that’s too obvious, but even that’s debatable. The sky can be gray, red, black… However, “the sky is made 98% of Oreo-flavored milkshake” is definitely a falsehood.

I think a key determinant of Trustory’s success will be to identify where it’s true value add is and where it can have the biggest impact:

  • Finding truths
  • Debunking falsehoods
  • Creating a discussion platform

Thoughts?


#2

I will answer in the order making most sense from the claims.

  1. Determining what is false is from a logical reasoning perspective very close to determining what’s true. It’s just a percentage question. If 98% think it’s false - then 2% think it’s true (excluding uncertain voters). Same thing, if 80% think it’s true - then 20% think it’s false.

From my understanding, the Trustory MVP has the main focus to determine what has a majority consensus. The implementability for this type of validation is much easier - as consensus can sometimes be very high (90%+), and there is a more obvious market to it (false information (news)).

One concern of the MVP is how Trustory handle it when the majority is actually wrong, but the proof provided by the minority isn’t sufficient to flip the stakes. The solution I understand Trustory has proposed to this is that a claim can be “re-voted” when new evidence show up/the community wants. But the emotional part of voters whom are to re-consider their previous decision is where the crux lays - since we humans don’t take into account all factors, are not rational, and have pride - even if we strive to be on our best. Thereby, even if the opposition (true-voters) re-state their claims in a more logical and understandable way to the false-believers - the false-believers won’t take full account to the new claims. A solution to this in future versions could be another type of staking settlement.

Terms I think we should be careful using (excluding synonyms)

  1. Obvious lie - just because 1% think it’s true and 99% are sure it’s false doesn’t mean its obvious lie, as history has repeatedly proved. Rather, what is an obvious lie is when the claim is broken down into pieces and the defenders can’t respond to the sub-claims and solve it by distraction or sensationalism. So obvious lies is only the type of lies where all sub-claims can be disproven - not just where the high-level claim has a high consensus.
  2. Universal truth - because science works in the way we can’t know anything with 100% certainty.

Regarding your point 1, I argue that from reasoning above - “Objective truth” should not be determined.

Trustory’s success and impact rely on many more factors than the consensus mechanisms.

Regarding impact on fighting fake information (news), studies has for example shown that when a human has viewed a fake-video, it is “impossible” for the human to forget and discount what they just seen. For Trustory to make impact - it must therefore play a role in content-feeds (social media, news, search) - which I understand is the direction Trustory want to go. However, to truthfulness-verify these external feeds (Twitter, FB, Google, YouTube) will be tough from a 1) collaboration and 2) scale perspective. The solution I understand is being worked now is importing this external content into the Trustory app - and with the scale of content and staking that requires - only very few claims will have market liquidiy enough to be verified - and thereby not very good UX. I don’t see that the market liquidiy with mass-scale verification of external content will change with time - which is why I rather see that the solution to this is using other staking mechanisms to enable cross-use claims.

To now respond to your point 2 - the main value and goal of Trustory MVP I say is experimentation. Moving to Version 1, I see there would be two goals:

  1. obtaining scale in high-consensus-false claims, which is urgent and important
  2. start experimenting in highly conflicting claims by discussing the underlying claims and thereby reaching a more complex consensus on the high-level claim

I have many other concerns for Trustory to obtain the success and impact this type of service deserves, mainly in obtaining scale and combatting social manipulation - which is why it is important we get to start experimenting as soon as possible!


#3

Excellent question @daniel! see my responses:

  1. The key to making TruStory work is focusing on the right type of content. We aren’t focused on moral or “universal truths”. We don’t even want to enter that territory because there is no point. For example, if someone says “Bitcoin is the best cryptocurrency.”… we can argue about that for 10 years and see all sides of the argument, but still, never come to a conclusion on whether that is an accurate claim or not. What does “best” even mean? How can we falsify such a claim? On the other hand, if someone says “Cryptocurrencies is mainly used for illicit activities like terrorist funding and child pornography”, we can objectively falsify this claim by proving other uses of Bitcoin.

  2. Both. You can’t really have one without the other. How can you know whether something is True if you’re not open to discussing how it might not be True?

  3. We’ve thought about this, but for MVP, we’re sticking to a simple majority. Once we see how things are playing out in the real world, we can know whether it makes sense to have the “Undetermined” state. I sense this would make sense if we want to also allow subjective claims. For now, we are so focused on getting the right type of content (i.e. objective claims) where there is little or no ambiguity.

  4. Totally. This is why we say that a story is valid if it can be falsified.

We’re in early days, and the important thing is to run small experiments and iterate until we get this right.


#4

great point! we do take this into account, by not penalizing people for changing their minds in the face of better evidence :wink: