Generally, in this second round, it was harder to find groups to interview. Several DAOs failed (PragueDAO, Reward DAO, Unicorn DAO) but these projects continued under different models. RetailDAO also has seen many of the things they try have failed, and is continuing to innovate. Both PragueDAO and RetailDAO are physical DAOs, so anyone looking at creating a DAO in physical space would be well advise to look at the issues they faced. RewardDAO resulted in useful findings about funding, and Unicorn about reputation.
As in the previous round, we had several major successes (Ethfinex, Diversefi and dOrg) and several which are just too early to tell. Onboarding continues to be a major issue. Below is the feedback, both positive and negative, for developers of DAO technology.
Kleros was excellent, functioned as expected.
Process was perceived as fair.
High level of trust: There is no one person who holds the money. Trusting the majority is much better than trusting a few people who sign checks.
Unhackable multisig wallet functionality works well.
With the Vermont incorporation, ease of adding new members.
Ability to go into overdraft: Money can be tagged for distribution even if it’s not yet in the DAO. This allows a FIFO queue even when there is a delay in inflow of funds. Allows continuous operations.
Onboarding continues to be a common issue with the DAO technology, across the board. Everyone noted it was extremely difficult to onboard people and there were lots of complaints. One organization appointed a paid employee to help people with this.
Easier ways for people to launch their own DAO (dOrg is addressing this with DAO-Creator).
Integration of communication about proposals with the proposals. Discussions of proposals might take place on Telegram or Discord, but the proposal was on Google Docs or Git, and there was no connectivity of the documentation.
Ability to connect identities of discussion personas to the proposal personas is missing.
Politics and campaigning are increasingly becoming an issue within the DAOs. Although people recognize this issue, there is no cohesive addressing of the problem of politicking and the increasing competitive nature of some of the conversations in the DAO. Genesis and Retail DAOs have people who are designated to create community, but there are no scalable solutions to date. Retail DAO's https://www.collaboration.ai/ is the only current known attempt to address this in an automated fashion. (Similar to some of the work done by Satalia, which has been partially effective in reducing politics.)
Staking tokens in Kleros was not as simple as it could have been.
Integrated proposal documentation, such that once a proposal is passed, it is stored persistently and can’t be changed after approval.
Proposal lifecycle: <Staging → Voting → Production → Milestone-based grants> The complaint here was that the proposal lifecycle today is more like: throw something up, get feedback, and then try again, which seems sub-optimal.
Proposal searching, categorization, and filtering.
Ability to change the amount proposed without re-submitting the proposals. Often people agreed to a proposal but did not agree to the amount requested.
Ability to edit proposals within the platforms. In some cases, people had to re-start proposal process because of a spelling error or disagreement on a small part of the proposal.
Multiple-choice votes (design contests, for example).
Rules other than immediate allocation of funds in ETH/DAI. For example, allowing participants to earn interest or REP if they leave their money in the pool after it is approved, particularly at times when the organization has a cash flow issue.
Some participants would like to see task management on chain, some don’t.
Ability to be transparent within the DAO but not to the public at large. People don’t want their salaries public.
Transaction fees were universally an issue. (In October, we saw Aragon begin the move to a proprietary chain in order to handle this)
Ability to have more than one recipient per proposal. Right now all proposals are individual, and it’s a manual process that takes hours, particularly for the person with the administrative responsibility to manage the DAO.
Audit trail for financial reporting.
Ability to hold funds in escrow until work is performed, tracking of work, or other accountability capabilities.
Carrots-only, no sticks. When people do not deliver or do not work as much as they promised, they invoice less, but there is no way to reduce their reputation or make any indication that this person is consistently late on delivery. There is only the ability to increase reputation and pay people, but no other mechanism for collaboration or task management.
When there is an obvious funding source, people in the DAO instinctively try to align with the desires of that person/entity in their work, so even if there is supposedly a hands-off approach, the behavior of the group doesn't completely reflect that.
Identity solutions. Pseudonyms allow people to be different people in github, telegram, metamask, etc. Some people have different names on different channels. Long-term that could be an issue.
Ways to integrate existing reputation knowledge (LinkedIn, etc.) into someone’s reputation on chain.
Profile management and more information on member profiles.
Reputation doesn't reflect the person's ability to execute on proposals.