Algorand Governance Feedback: Decentralization Concerns - Governance Discussions - Algorand
Algorand Governance Feedback: Decentralization Concerns
post by awesomecrypto on Apr 14, 2021
TL;DR Decentralized decision making and execution is inefficient and ineffective and poses a great risk to the Algorand growth and competitiveness moving forward.
- The goal of this note is not to claim that decentralized governance cannot be useful in any form but to encourage deeper discussion and research among the Algorand teams and the community before moving full-speed in that direction.
- AFAIK there is not enough public information about how Algorand Inc., the Foundation and network are supposed to work and influence each other. For example I asked a number of questions here.
Staking Algos is just one requirement for becoming a decision maker. It only ensures there is some interest alignment with the network success. However, it doesn’t guarantee anything about the quality of decisions or whether they are part of a broader long-term strategy. First an important part of decision making is coming up with the decisions that need to be made, which is missing and is left to proposal owners and not under a unified strategical plan. Second, decision making for a global monetary/DeFi network is a high-profile job which requires dedicated time and effort, expertise and knowledge. The majority of participants (95+% if not more) are not qualified for this role. Also the governance reward is too little for the small group of qualified participants. Third, there are too many participants, with many different backgrounds and just ensuring enough information flow so they get to a common understanding of problems/solutions is difficult (if not impossible) and a slow process.
The damage is not limited to decision making and execution but also it causes a lot of unhappiness and sadness within the community, leading to many divisions and eventually brain/money drain. No sane expert would want to waste hours and hours his/her precious time discussing basics with random people so may or may not move the needle just a little bit. As much as people get excited when sometimes the process works, stepping back there is little justification for it particularly when a centrally govern competitor (say Binance for example) can move one hundred times faster. Decentralized governance is not a goal but a means so its cost should be lowered as much as possible or even alternative approaches are used when possible.
I had the chance to work with the Dash community very closely for a few months. I observed many of these issues in practice. Dash governance benefits from interesting ideas (for example there is a treasury and the network can own companies and central entities) however it still suffers from serious issues and fails in addressing real problems effectively and quickly. The result has been that Dash is not able to gain its deserved position and value in the market even when it still provides cheap/fast(1-2s)/scalable transactions. In particular not only the community has not grown enough but there has been a huge brain drain.
post by RAIFF on Apr 14, 2021
@awesomecrypto You do realise the objective of Algorand is a truly decentralized network, right? In order to be decentralized there are consequences and it means sharing the responsibility and decision making with people that you might consider inferior or have a different opinion than yourself.
post by dirac on Apr 15, 2021
Does it? Governance is about changing parameters of the protocol to be more in line with every participants values. You don’t need an MIT degree for that.
To make a point about that “elite” you’re talking about: There’s not a single currency in the world that has ever stood the test of time. These “experts” might understand monetary policies better, but they’re also more likely to manipulate them for their own gain. As far as I’m concerned, banks gambling on the market and getting bailout money don’t seem like geniuses to me.
Centralization is not the way to improve as a civilization. We need strong education with a focus on critical thinking and depolarization.
post by tebok on Apr 15, 2021
Forgive me for ignoring your other arguments for a moment, but even if it was true that 95% of the participants were not qualified for the role, as long as the number of governors is high enough, then even 5% of qualified participants may be enough to produce good decisions per the “wisdom of the crowds” principle.
By “wisdom of the crowds” I mean that if 95% of participants vote randomly (and independently of each other) while 5% actually vote for the “best” option, then with a large enough number of participants the random votes will statistically cancel each other out, while the 5% of participants voting for the best option is enough for that option to win.
Of course, it is not quite that simple in practice due to various biases or group think that can influence the 95% of participants to not vote randomly (or independently), but in my opinion that is not a good enough reason to abandon the whole idea of wisdom of the crowds that works so excellently in many different areas.
It is certainly worth trying to get the most qualified people to participate in the governance too, but in the long term I fully agree with the direction Algorand is taking in trying to decentralize the governance of Algorand.
post by Maugli on Apr 16, 2021
The real problem is that Prof. Micali didn’t specify clearly enough the SCOPE of the governance.
Its scope may be “bicycle shedding” and getting good money (Algo) for it.
...Ethereum is moving forward very fast. There’s little opportunity for “bike-shedding,” an expression that means holding up development by arguing over minor details such as how to build the bicycle shed at the back of a nuclear power station. If you start bike-shedding, you might suddenly discover that while you were distracted the rest of the development team changed
From “Andreas M. Antonopoulos, Gavin Wood Ph. D. - Mastering Ethereum_ Building Smart Contracts and DApps-O’Reilly Media (2018).epub”
post by DangerDano on Apr 22, 2021
How is Algorand (or any proof of stake based consensus) going to stay decentralized when 1 Algo = 1 Vote?
It seems entirely possible that the decentralized nature of the governance protocol can be compromised by someone (or group) with lots of money who can buy more than half the supply, essentially making them the only vote that matters.
Not only that but if this person does vote, they’ll get even more power over the governance of the chain in the future.
@tebok mentioned the “wisdom of the crowd”; this would be true if 1 Account (Tied to a real identity) = 1 Vote, but this is more like a game of the " Wisdom Of Who Has The Most Algo".
post by Maugli on Apr 22, 2021
We can only hope that that the initial (secret, not too transparent) allocation went to parties Prof. Micali knows and trusts.Anyway, everything is on the blockchain, it should be analyzed, who is who.
post by scholtz on Apr 22, 2021
they say that governance protocol is about managing the AERP fund… governors should have power to decide how many algos are distributed between grants and how many goes to relay nodes incentives…
governors do not decide on who receives the money… algo foundation has monopol in this… perhaps in 10 years when there will be fight only for some algos from fees, they will give it to public
so 1 algo = 1 vote means that the algo community believes in hypercapitalism… the founders of algo or early backers should decide whatever they want… they do it anyway, so whats the point to complain… right? @DangerDano … you should also believe in hypercapitalism if you are considering yourself as part of community… whatever Silvio (or his men) says is the word of God…
But to be honest, I believe that from this type of discussion the algo leaders will not lock themselves in their believes and be open minded and will start at least some discussion that will be beneficial for algo community...
post by tebok on Apr 22, 2021
Yes, I was describing it purely in terms of “1 real user = 1 vote” (not 1 algo = 1 vote). But I should mention that in some conditions this criteria can be relaxed without losing the “wisdom of the crowds” benefits - best example being the betting market (i.e. prediction market).
In betting/prediction markets the most accurate predictions are not just a result of averaging the opinions of lots of participants, but also weighting each opinion by the the “confidence” each participant attaches to their bet (i.e. vote) through the bet amount they are willing to risk to back their opinion. And in these markets the bigger the bet, the more it moves the market (i.e. has more weight in the final prediction). (Of course the rich can still afford to bet more than the poor, but on average the larger the bet, the more confident that bettor is of their opinion.)
Now, does that argument apply to the algorand governance proposal and make “1 algo = 1 vote” the ideal approach, I certainly do not believe so.
My understanding of the the algorand governance is that the governors are not actually risking anything by committing their algos to governance. And therefore the governing weights are not based on the quality or confidence of each governors’ vote, but based on simply how algo-rich they are. So your characterization of “Wisdom Of Who Has The Most Algo” is quite accurate.
post by tebok on Apr 22, 2021
To continue the “1 algo = 1 vote” discussion…
I personally want to believe that the “1 algo = 1 vote” rule is more a limitation of the current implementation and the fact that the protocol itself can only only see algorand addresses but not actual users, rather than as something that must always be that way.
Unfortunately there is no way around it without adding KYC or other major complications to governance approach, and therefore I don’t see it changing anytime soon, if ever. And quite possibly never, because once a system such as “1 algo = 1 vote” is in place where people with most algo have the power (and can stay anonymous while using that power), they are unlikely to give up that power by reforming the rules of governance. (Unless of course they are also idealists, but generally speaking most capitalists and major investors are not, though I don’t know about algo-investors.)
Still, I do want to acknowledge that if comparing the options in a vacuum, governance via “1 algo = 1 vote” is clearly better and more decentralized than if the Foundation continued to make all the decisions themselves. I just fear that it will forever be stuck in just a tiny bit more decentralized yet still highly centralized form.
Also, to me just as big of a concern is actually the way delegated voting was added to the governance proposal. While generally speaking delegated voting can be quite useful and a good voting method, in this case the users are incentivized to delegate their votes to the foundation, which in practice just brings us back to the centralized reality where we are now, where the Foundation makes the decisions.
In the community all hands meeting it was explained how in order for governors to receive rewards they need to either vote or to delegate their vote, and currently the only option to delegate one’s vote is to delegate it to the Foundation. I suspect most of the governors will do just that (delegate to the Foundation), in order to not miss out on the rewards. To me this incentive structure doesn’t seem like an accident, but suggests that the Foundation wants to appear decentralized, while in practice continuing to make most of the decisions themselves. (If I interpreted the Foundation’s intent wrong, I sincerely apologize.)
In fairness they did mention the possibility of adding other entities later that governors can delegate their votes to, but who chooses these entities? I can tell you from politics that every politician (even the crooked ones) is always willing to let the people vote, as long as the politician gets to choose the candidates (because they know that leaves no real power to the voters).
post by mmorselli on Apr 22, 2021
In cryptography almost anything is possible, but is it also likely to happen?
In proof of work, who is stopping a single person from buying 99% of the computing power? It’s not impossible.
post by scholtz on Apr 22, 2021
You do not understand… No one is talking about damaging it… The person who has a lot wants to have even more… And he does that even if someone else has better idea and he doesn’t get funded because all funds got the rich person… It is about the efficiency of the economy...
post by DangerDano on Apr 23, 2021
Here are the current top account according to the Algo explorer:
post by DangerDano on Apr 23, 2021
Here’s that account compounded at 6% over time:
I know that the 6% is not going to last past 2022 but theoretically running this calculation would mean that this account will accumulate more than 6 billion algos (more than half) over 50 years (less than a lifetime) by doing nothing and automatically compromise the overall system through the power of the vote.
From my understanding annual percent will not continue past 2022 and the participation in voting will be incentivized according to how many participants there are, giving a bigger % return if there isn’t a lot of participation. This still continues the perpetuation of the rich get richer fallacy where the more algo used to vote the more you’ll receive in return without risk of losing it.
post by tebok on Apr 23, 2021
The question might be: are there foundational principles? Like an ALGORAND Constitution?
post by tebok on Apr 23, 2021
I love the idea of “Algorand Constitution” for exactly the reasons you mentioned.
Another idea would be a community-driven “Shadow Government”, with no actual governing power whatsoever, but voting on the same issues (and many more) as the real Algorand Governors. This would help identify issues where the “wisdom of the crowds” disagrees with the Foundation and/or the “wisdom of who has the most algo”, and allowed raising those issues for a deeper public debate.
post by mmorselli on Apr 23, 2021
The Algorand Constitution should be the first document voted by the new Governance. Should contain more stringent mechanisms to modify itself (e.g. 75% of votes instead of 50%)
post by tebok on Apr 23, 2021
While it is of course true that the constitution would need to be something that cannot be changed via a simple majority vote by the algorand governors themselves (or it wouldn’t serve the purpose of a constitution at all), in my opinion most supermajority requirements violate the “wisdom of the crowds” principle, and will simply lead to inferior decisions getting stuck for far longer than they should.
post by scholtz on Apr 24, 2021
I agree with constitution idea, but I believe that me and you have quite different opinion what should be within that constitution…
When you still do not understand that 1 algo = 1 vote with non-transparency as provided by the algo fund, I am quite worried that this hypercapitalism will be enforced in this document and nothing will be able to change that in the future...
post by DangerDano on Apr 24, 2021
The fact that it’s a Proof of Stake and 1 Algo = 1 Vote simultaneously makes for easy game theory. Accumulate Algos and Vote for self Interest.
Adding layers also adds complexity. The voting system (even tho not a bad voting system of it’s time and overall for that matter) was built without the tech we have now. A lot of the layers that were made as checks and balances can be innovated on by reaching directly to voters is on each proposal.