Proof of Brain Theory & Further Optimization

Further Optimization of Proof of Brain (PoB)

logo_transparent@1.png

HIVE is a base layer governance token.

No one token can reward all things. I foresee HIVE being a "rare" token to earn in the future as more and more tokenized communities pop up and as HIVE inflation goes down. Those who accumulated HIVE early enough to get a large enough stake have a "cheat code" of sorts, a asset accepted everywhere, coveted platform-wide, to "tip" for "x" - "x" being anything your brain can think of, as HIVE will transcend all tokenized communities and barriers. It can be seen as the upvote of the internet, like visa, accepted everywhere. With it, you can have access to endless content, knockdown paywalls with the click of an upvote. It'll be like having all of your favorite subscriptions paid for, for life. It'll be like having infinite free marketing for your business at the tip of your mouse. The ability to incentivize anyone, anytime for anything will be seen and known as one of the most amazing concepts ever created in the digital world. It flips all business models on its head, it's much more frictionless than traditional tipping, and once you go PoB upvotes for tipping, you don't go back to out-of-pocket tipping. It just does not make sense.

HIVE is the "Crypto Punk" of upvotes, the rarest of the rare, the most highly regarded; we're just in the unminted phase ;). HIVE best fits as the top token to seek and an honor to be earned, unexpectedly, among your journey earning many other types of tokens, NFTs, and assets. However, alone with HIVE only rewards makes for an awkward place, as there can only be so many large curators vs. potential content creators, one far exceeding the other.

Community curators in tokenized communities are akin to gatekeepers; that's why you want as many of them as possible with the lowest entry possible to become one. What is good about having niche-specific curators is, these curators are more likely to see your car-related content in a car-related community. Niche communities offer people a way to "grow a stake" with their knowledge and helpfulness of a subject with a greater chance of being seen. Having a community with a specific theme and its own siloed reward pool means you have greater inflation to give to fewer people, which means, on average, your car-loving community members are exceedingly more likely to earn rewards in a car-themed community, etc.

Sure, you may not always be earning HIVE on your post, but the car community mods like your content, and you're loaded with car coins. You're also loaded with several other coins of hobbies that you find interesting. You now have a sense of ownership in your interest, a invisible bond with your community, a network that can form monetary value simply by the quality of knowledge in that community. Community tokens give ownership built on immutable HIVE land that can be developed, painted, molded in an endless variety of ways, giving everyone a chance to "earn" something.

Diminishing the ability to earn is not the way; web 3 is the abundance economy. People can earn money on HIVE; they do daily; that's nothing to be ashamed of; they'll earn a hell of a lot more from talking about their Splinterlands investment on LEO; hey, that's two tokens in one post!

Rewarding of content, games, and all of the rest will be done increasingly on tokenized communities. This is both game theory optimal and has already played out in practice. People in tokenized communities have started to experience greater monetary rewards than they have on Hive. Splinterlands has its own in-game token that it rewards its players with, which lives outside the influence of HIVE PoB.

One of the most important reasons to have PoB token distribution on the base layer is to get the governance token in as many good actors' hands as possible and offset the inflation rewards that witnesses & DAO contractors make. We are a staked-based electoral system. We must always make sure to keep a balanced token distribution; if not, we risk being Justin Sun'd again. 2 years ago, who would have believed what happened to us would have even been possible? We live in crazy times, and it's best to make sure we learn from our mistakes and make an effort to always distribute HIVE to as many loyal Hivers who power up and vote in governance. - This is our superpower; this is what separates us from everyone else; it's the ace up our sleeve that will benefit us down the road if we steady the course.- It starts with and ends with the token distribution on any stake-based electorial blockchain. It's first and foremost the thing that matters for censorship resistance of the base layer.

The biggest plague to BTC was China's centralized hash rate (akin to stake in PoS). Now that China has come out against BTC, supposedly, we see a big relief in the BTC community due to the further decentralization of the hash rate. How much would BTC be worth if it was known that 90% hash rate in China? It would be worth whatever the Chinese market makers make it, as no one else would use it because it would be worthless to anyone except the Chinese government.

"It's not about what attackers won't do; it's about what they CAN'T do." Like I said previously, no one wants to hear, "oh, don't worry, the tiger won't eat you..." you want to hear, "the tiger CAN'T eat you!"

We must learn from our mistakes, and everyone else's because, thankfully, we can't make every mistake ourselves, but we can learn from every mistake made.

When this technology first came out as Steem, the price of STEEM boomed, it became a top 3 cryptocurrency, and people were making thousands of dollars per post left and right. Big names would join, get thousands of dollars in upvotes, tell their large followings, the token was spread far and wide, the price blew up in value, and all was seemingly swell. Everything was new, no one understood how to exploit the system right away, and decent curation was happening, which helped create the initial network effect. People from all over flocked to this new social media that paid you to post! While it didn't always attract the "good" people, the good people it did attract ended up being priceless. This lasted for a while; lots of curating was going on, but the exploiters started to understand how the system worked, and the flaws started to surface over time.

Now, it is never the exploiter's fault for exploiting a "game." I view all of this as a game, with rules and ways to play. The object of the game is to create a balanced game, one where there are fewest exploits. So, if there is a way to exploit a system, don't hate the player, hate the game, or better yet, fix the game. I have a little sense of respect for people who are good at poking holes in games; I have a poker background, and that is all I did for a living for the better part of 10 years.

If someone played badly, I exploited it and made as much as I could. If I didn't, that would be a greater shame to the player; there is no honor in holding back, only honor in a hard-fought defeat/victory in fair games of skill.

The exploits started innocently, people setting out to find ways to sell votes to others; on paper seems like a good use case, right? -- People buy up the token to sell votes to others, cool. However, in practice, once exploiters found out that selling votes were far more profitable than upvoting others (real curation) and less of a hassle than creating 10x a post a day and 100% self-voting (many still did), the shininess wore off of the system. Eventually, trending was overtaken by ads using bidbots to boost their post. The worst part about bidbots was, in essence, you pay for a upvote, you get upvoted, and you earn your crypto back. So it was like free advertisements. Few understood this at first, but it quickly got out of hand and became a free for all, with bidbot owners chomping at the bit.

It got to a point where people who bought their stake wouldn't upvote others for free. The reason was simple, stakeholders who invested enough to have large enough stakes required to boost someone to trending would be MASSIVELY diluting themselves by doing organic curation.

I learned this first hand, I bought $1m worth of STEEM (@ 1$ ea) when I made my 2nd big purchase; the 1st was around $250k worth of STEEM (@ $4) - and as I was proceeding to upvote "honestly," I realized over time how badly I was getting diluted, I had to keep buying STEEM to keep up with exploiters getting it freely.

I was interested in owning a "piece" of the protocol and have my small say in governance. So being diluted by what I perceived to be "short-term investors," I felt the pressure of playing the red queen game and keeping up; mind you, my original goal was only around 1m STEEM at the time.

It was a system set up to fail, and it was spiraling out of control by the day. It got to the point that the only way you'd reach trending is if you bought votes, it was impossible to get there organically. Not many were downvoted because it came with opportunity cost at the time; there were no free downvotes back then, so that means each downvote came at the cost of your voting mana you could be using for upvotes.

With so much money to be gained by selling votes with all your voting power, those who tried to stand for good were trampled under the weight of massive exploitation that only a drastic change to the underlining economics could solve. The idea of PoB for base layer distribution was on life support.

A drastic change was made in the form of a new EIP.

PoB has changed in various ways over the years. However, the biggest shift that created the largest impact was moving from 25/75 (author/curation) to 50/50 and adding 2.5 free downvotes per day (cost zero voting mana to use.) And WOW! What a change it was! Bidbots were run out of town overnight; trending became organic again; I've never seen such a reaction with a few tweaks than what we saw. Many people new to Hive probably don't even know what a bidbot is; that shows how far we have come.

The EIP changes made even the most exploitive change their ways and do "good" out of pure greed. Due to 2.5 free daily downvotes, it was no longer more profitable to sell votes or (only) self upvote to gain 100% of the reward from both author and curator. Now it was more profitable to be getting 50% curation rewards for doing good. Mind you, this is no ordinary feat and is unseen in the token distribution world.

Downvotes are equally as important as upvotes and are far more important in practice to be used because currently, they are drastically UNDERUSED. We need to normalize downvoting, as we would give a "thumbs down" on any content on web 2.

You currently get a better feel on web 2 platforms of the objective truth of what people think of the content, simply because they can downvote in private and do not need to think of anything else.

First, if all web 2 downvotes were public, IE you could see which account downvoted content, there would be much less downvoting, even without monetary value attached to them. Most people don't like confrontation; it isn't worth putting themselves out there to be attacked.

Once you add value, that brings things to a whole new level. Because on Hive, when you downvote someone, they take it much more personal than they would a web 2 downvote.

That is why downvotes are much more important on Hive in practice; no one complains about getting an upvote, and they're sure isn't a lack of upvoting going on. But there is a lack of downvotes, and without downvotes on PoB, it's destroyed and, in essence, makes distribution turn from PoB = PoS as everyone will upvote themselves to get both curator and author rewards. We want token distribution; not rich get richer if we can help it.

Anonymous downvotes on Hive would benefit negative curation greatly, I'm not sure how that would be done, but people would use them much more liberally if we had them. This is the "magical" thing I hope someone can figure out, maybe with the use of technology such as zkrollups.

PoB is very important and still in its infancy. What PoB has solved is an issue people have been trying to solve for decades. How do you distribute value without keeping the value for yourself? It's a difficult challenge that Hive has approached in a novel way. It's taken 5 years of being "attacked" to figure out what we know now. It's not perfect, nothing is, but it's damn good and getting better.

While downvotes are necessary for PoB to function in a decentralized way, and by adding 2.5 free downvotes, we have given plenty of ammo to the "good guys" - but inadvertently, we also gave plenty of ammo to the "bad guys." Keep in mind; we are spreading a governance token here; no one large entity should be able to censor governance distribution on a select group of people, IE "targetted bulling" - emotions should never lead to the ability to suffocate another's rewards in perpetuity without a good ability to defend vs. it.

The damage one can cause alone is on a small scale when compared to the greater ocean, but we should still try to tighten the belt there and do the best we can. And one large downvote here and there is not a big deal; the issue comes when one large entity turns to targetted bullying, focusing on one person or small group and essentially downvoting them into the ground in perpetuity regardless of the "quality" of their content. The only defense we have against this is asking good actors to dilute their stake to reverse the damage.

And while I agree that downvotes are drastically underused compared to upvotes, and downvote abuse is actually quite rare, but we do have a hole on paper. Just because something is small does not mean it cannot fester and grow larger. It can still wreak havoc, and we have been "fortunate" enough in the past to have very large stakeholders act like complete maniacs, downvoting everything in site good or bad. This gave us real-time experience in handling downvote wars on a real-time multimillion-dollar monetary network. We have had the most epic downvote wars you can imagine; most older Hivers here know them by name. And while the names of the epic downvote war list are small in number, their impacts can still be felt to this day. The abusive downvotes are akin to a turd in a punchbowl; there is still much more punch than a turd.

The main issue is to counter downvote abuse has an opportunity cost to the upvoter trying to help. Under the new flat ruleset of curation rewards, the only thing that can lower your known rate of return is downvotes. Again, asking someone to do good, and in return, they get diluted is not a good business deal, and no one will take it up constantly, nor should they. Good people acting good for the sake of good at the cost of dilution end up becoming irrelevant power-wise; thus, their acts of good are useless in terms of having an effect. We saw what happened when you removed the opportunity cost of downvoting; people used them to help the platform.

So I propose a few options.

One free upvote per day that can only be used on a post that is already downvoted and can't surpass the downvote amount. The post cannot be voted on twice, meaning everyone who voted before the downvote can't revote the post with the free upvote.

For example, say a post is nuked to zero, the exploiter cant use the free upvote because the exploiter already self-voted and cant self-vote the same post twice, so all the bad actors trying to exploit a post cant re-exploit it.

However, if the post is nuked to zero by bad actors, outside good actors can come in and use their free upvote to counter.

A bad actor has no good outside votes, only self-votes; therefore, the only thing that can be downvoted is the attacker's own vote; therefore, they cannot use the free vote to counter. Since the free upvote can only be used on a downvoted post, you can't use it to earn anything, IE it's not like a free spam post upvote per day for attackers.

This is doing the same thing we did with free downvotes except for upvotes; it removes the opportunity cost to reverse downvote abuse. In practice, we know if you let good people do good without being penalized, they will do good.

EDIT* 8/11/21 - thanks to @smooth who rightfully pointed out an obvious attack vector here that I somehow overlooked. "You can split your stake into two accounts, post with both, upvote one post with each account, and then use the other account to counter downvotes." - This makes this defense actually more of an attack vector by giving abusers the ability to recoup losses. - But there may be a way to do it in a different way and accomplish the same goal. Further from smooth: "I think it would be possible to take the downvote curation reward penalty only from those upvotes chronologically before the downvote, so upvotes to counter the downvote wouldn't be penalized."

Another idea is an off-chain solution that can be paired with the above idea or done separately.

Downvotes are human-run and, therefore, a human issue to solve both inside and outside of the code. We can always have anti-abuse proposals for those that are being downvoted maliciously.

A DAO proposal where a powered-up, ownerless account (null the private keys) that is set to work when thresholds are met - IE, this account only upvotes when this proposal passes x amount of votes, etc.

The idea is you have a DAO-funded bot that powers up Hive and is only activated via a trigger. The trigger happens when the "victim" sets a proposal to the community with their information and evidence, and enough of the stakeholders agree that there was, in fact, abuse—the community votes to reverse the damage. The victim would receive "anti-abuse tokens" (AAT) where they burn them to use the anti-abuse bot, input the URL of the link they want an upvote on it, and the bot votes the post based on the number of tokens burned. AAT user beware; having the anti-abuse bot vote your post does not guarantee it will not be downvoted, so do not think of them as a "free immutable vote" that can be cast on "spammy" content. The tokens would need to be tied to the account, the account can burn the tokens, when burned the bot can check that they were burned, then be activated. Basically, a way a token is "stuck" in an account and an event can be triggered telling the bot to vote a post.

A neat feature here, when one is deemed an "abusive" downvoter, one could put them on a "list" pending community approval, whereby every time the attacker downvotes their victim, the victim automatically receives AAT proportionately.

I've yet to find a way to punish abusive downvoters in the same way you can punish abusive upvoters. See, when someone upvotes a spam post, downvoters come and remove rewards; ouch, there is your punishment as the upvoter. The downvoter does not have any recourse to deal with; so far in this post, we've only managed to waste said attackers' time, which is a sort of punishment but not a direct one that results in monetary loss.

PoB is a growing organism, and I believe we have gotten very far on this idea. However, if something's not broken, but it can be improved, it's as good as broken. Let us always strive for the best we can achieve.

I have other ideas; as I was writing this, I actually had to rewrite it as I came up with the 1st idea I thought worked better in practice. In theory, you want to give the "good people" as much ammo as possible while giving "bad people" the least. In the 1st situation, both good and bad have equal footing at 2.5 free downvotes, but the 1 free upvote vs. downvotes gives the good an edge both ways, for positive and negative curation.

H2
H3
H4
3 columns
2 columns
1 column
90 Comments