Introduction
With the round having ended, I would like to write a post where I will outline my voting rationale and methodology for this round. Also what I learned and how I think I can improve as a voter. Finally have some feedback for the process.
My voting
Below I will try to explain my voting rationale and what I have been hoping to achieve by participating in the 3rd round of retroactive public goods funding.
Rationale
My goal is to divide the 30M OP based on what I perceive the impact/value of each project’s public good has been. I want the result of my funding to have as broad an impact in as many projects/individuals as possible. I believe the RPGF process to have the potential to be extremely impactful on a wide variety of projects and individuals.
The project has to be creating some kind of public good. The definition of a public good is that of something that is open, creates value and can’t be captured.
In some cases I also considered future potential of someone’s work though this mostly to help decide amounts towards the end as it was quite hard.
I took into account not only donations/grants funding but also VC/angel investments and any other form of revenue the project has had. The idea follows the impact = profit guidline, just makes sure to not only subtract the value of the grants/revenue but also investments (with a smaller multiplier as they are not income)
That means VC funded projects are not automatically excluded. It just means amounts and revenue matters in my methodology.
Methodology
First I started by using a spreadsheet and populating it with the projects of the opensource impactful list I had created (minus rotki of course). I slowly started adding more opensource projects consulting with other badgeholders, and then also checking what people sent me in DMs or other social media mentions.
To those projects I assigned various signal values from 1 to 10 which I later intended to scale according to the number of projects.
Then I started checking various lists and had many discussions with fellow badgeholders on what to include and with what kind of signal values from other categories on which I am not as well knowledgeable on such as education.
In the end I came up with a big list of ~210 projects and various signal token amounts assigned to each depending on their impact.
The final amount was determined by the signal multiplier and the following “equation”: amount_of_impact_generated - public_good_classification - grants_donations_revenue - investment_funding
It’s all quite subjective but I will try to explain how I determined the amounts for each variable.
- amount_of_impact_generated: How well has it affected my life as a crypto native, how has it affected other people I talked to. What has it given to the ecosystem? This is the impact.
- public_good_classification: Public goods is a very hard to define term. And in the ecosystem we use it for many different things. I assume every grant application is 100% completely public good. Then look into details and add some minus points in this variable (it can only be negative or zero). Things such as token gated, paywalled, not opensource, SaaS, centralized etc. would add minus points here
- grants_donations_revenue: This is trying to see if the project has already been compensated by enough grants/donations/revenue for the impact they have had. If yes, depending on the amounts this would deduct points.
- investment_funding: For better or for worse we have had many well funded projects in the round. Depending on the funding raised and time since the raise I subtract points from here. I tried to take into account the foundation’s guidance and argument that investment != revenue and as such the multiplier for the points here was smaller. What this means is that the value affected only extremely well funded big projects very negatively.
It’s possible that the project got so many negative points that the final amount was negative. In that case I voted with a 0 allocation. What’s more I used the 0 allocation when I saw projects from the same org having multiple very similar grants or something that’s a scam.
Once doing this and finishing allocation for most of the projects that I know, had done research on and had heard from friends, colleagues and fellow badgeholders I was surprised to see that I had more than half the $OP left to allocate.
It was at that point that I really realized how much money it is we are giving out and how much impact this can have.
At that point instead of just spreading the rest among the existing projects I set out to find more projects that could use this funding, had some impact and just did not have the brand awareness / marketing that others have.
Spent a lot of days on this up until the last hours of the submissions where I had to slow down due to technical difficulties. My goal was to try and help as many projects that had helped our ecosystem in some way as possible.
I hope that I achieved this in some degree with this methodology and if I participare in future rounds I would like to refine it and reapply it.
Learnings
Here I want to write a few things about what I learned as I went through the RPGF3 process during the last month. Some are a bit personal so please bear with me.
Background
Why am I doing this? I have been in this field working for the EF since 2014 even before the launch of Ethereum and contributed in a vast array of projects. People call me an OG. Unlike many of the people from back there who have already retired I am still here. Why? I am an idealist and really believe in this field, in opensource and am a big proponent of decentralization. Making money is not what I care about. I care about making the world a better place … yes I know kitchy as heck but I need to explain it in order for the rest to make sense.
Why do I participate in RPGF?
For the “public goods” part. If this program ever became just a retroactive funding program I would not be interested to be a badgeholder and devote all this time to it. I believe public goods (which are hard to define) need to be funded and that RPGF is in a unique position to achieve this.
Learning 1: Nuance is hard
When discussing things in social media, it’s very difficult to discuss nuanced things. No matter how much you try due to the “drama = clicks” algorithm any argument gets simplified to a black or white kind of argument.
And this is really bad because nothing is only black or white.
For example people were saying that I claim all VC funding is bad. That is not what I was going for, but it may very well be what prevailed as the message in social media due to the above.
Learning 2: People seem to trust me
For lists I noticed that many people were mentioning both in the TG, Twitter and the badgeholder calls that they respect me and would follow my opensource list.
To be quite honest this terrified me and made me feel a very big burden of responsibility. My original intention with the list was to try the feature and highlight some impactful opensource projects I know of and use.
Then as more and more people starting sending me projects I did not know of but satisfied the requirements I had set for the list I made a V2 including more (also including some I had missed from v1).
But then I noticed that the same people later were using the fact that they made it into the list as a huge endorsement and advertising it to all other badgeholders. That’s when I stopped making any more lists and tried to make it clear in the list descriptions that this is not an endorsement.
My biggest learning from this is that probably lists can be quite harmful and that I should be really carefu with any endorsements while the round is active.
Also would like to ask for people to trust me less. I am just another dev and I am flawed as hell.
Learning 3: VCs are not bad per se
Okay this may not be a personal learning. I knew that but I think that the narrative got twisted somehow to the point that I had friends who had raised small amounts, were building opensource and had to fire their people due to lack of funding, DM me.
So let me make this absolutely crystal clear. (VC or otherwise) funded projects can and are building public goods
That can be opensource software(e.g. paradigm), can be an open ethereum RPC that they keep open for all to use (e.g. ANKR), can be a team that builds the most impactful QF mechanism we all use (e.g. gitcoin).
The argument I tried to make during this round is that size of funding matters. Much as we were asked to take into account any grants/donations/revenue I argue we should very well take into account investments also. With a different multiplier than revenue/grants/donation indeed as it’s not income ofcourse.
This is why I specifically called out only 3-4 projects that are literal giants and have raised multiple times the amount of the entire round with the most impressive case being more than half a billion.
I believe that at that point they don’t need the RPGF and that it’s unfair and greedy of such projects to leverage their huge marketing and brand recognition to get many votes in RPGF.
One approach to a solution for this would be to better formailze and apply the “equation” I mentioned in the rationale section taking the amount of funding into account.
But I still don’t know how to fight branding recognition and popularity contest problems. As this is not limited to funded projects. Many bootstrapped projects also have very good branding recognition.
Now all the above said, I have a lot of other issues with the VC model which I won’t go into here (already analyzed many times in Twitter and other writing) but they are offopic to RPGF or public goods.
Learning 4: Sometimes I can be too intense
The reason I gave the background section above was to explain why I am here and why I am doing this.
For better or for worse I can sometimes be too passionate and intense and my tone can make some people feel uncomfortable. I even got a Code Of Conduct violation warning from the foundation for my criticism this round (it was not specificied why exactly, nor was any other action required on my part – just a warning). So I assume it was for my criticism in Twitter.
I stand by everything I said and I don’t believe calling out the truth, criticizing what’s wrong and speaking truth to power is wrong. I will always call a spade a spade. In that respect I do not consider criticism as a CoC violation.
But the tone matters and some times I can be intense. I am a hot-blooded southerner guy. To that end, if I made any of my fellow badgeholders or other people uncomfortable during the round due to my tone I would like to apologize. That is never my intention.
I will keep doing what I do and if I sometimes become too intense please bear with me. Will try to not overdo it.
Feedback for RPGF rounds
In this section I would like to give as much feedback as I can on the round process now that I have it fresh.
Include investment funding
As I asked many times during the round and also tried to explain why we need it above, I strongly believe asking for all funding details including investment is paramount for us badgeholders to make informed decisions.
Individuals versus teams
I am a bit confused when I see individuals having their personal grant and then also teams where that individual works in having another grant. I don’t know what to do in that case. Perhaps we could have different subrounds for individuals and different for teams?
Collections of projects versus projects
In a similar tone there is grants like Protocol guild which is a curated collection of core projects and researchers. Some of those have their own grants. How should we handle those? Voting for both of them makes sense since core development is important but then some projects end up double dipping.
It would make sense to handle such grants in a totally different way but not sure how yet.
Lists seem to be harmful
I am not sure if lists are good. As I wrote in my “learnings” I got terrified of how people used my list. And also definitely saw some people abuse lists, either by creating a single list with the Protocol guild (sorry that’s wrong – no matter how much I like them), or by being a “lazy” badgeholder and just copying lists.
There was also a lot of criticism on lists in the TG with other arguments mentioned.
One many people were afraid of is people blindly copying list amounts
Allow lists without amounts, just recommendations
Not sure if we should have lists as I said but having lists without amounts seem like something pretty obviously missing. Giving amounts to a list preconditions the list reader in a pretty ugly way. It’s much better to leave amounts to each individual badgeholder.
Create local app/browser local storage webapp
We have had too much trouble with Cloudflare and other technical problems. What this entire thing is, is just a spreadsheet that can and should be configured locally and submitted in one RPC call to whatever backend you guys want.
But working on the ballot should not be a roundtrip to the server. Keep it all local. I wasted 4 hours simply modifying the amount and scaling some amounts up/down in the last day.
Allow CSV import
Since many of us worked on a spreadsheet please allow us to just export this into a CSV and import/submit the results via CSV.
Better project application process
I don’t think FFA works. We had so many projects that we needed to have some people do reviews and still we ended up with many projects.
I think the old way of having a nomination based system is good enough to filter lots of spam out. Perhaps there is better ways but free/open to all is not a good idea.
Longer round duration
Even with better application process I guess we will see more, perhaps double the amount of projects next RPGF round. This is something that @kaereste suggested at least once and I think I agree. We need a lot more time and rounds should be many months long to give us ample time to work on these ballots and do this research.
Badgeholder compensation
In the last month I spent 2-4 hours per day, with some days(like today) being a lot more, working on RPGF and either researching the applicant projects, or giving feedback and/or thinking on the process.
Just like the other badgeholders all this is pro bono and we all do it cause we believe in the impact this process can have.
But as the workload increases and the time commitment gets even crazier (which is why the “longer round duration feedback”) we should think of scaling the process up by thinking of some scheme for compensation of all this work.
Brand recognition/popularity contest
This is not an RPGF only problem. It’s in every such voting system. We have seen it very strongly in gitcoin too.
It’s very hard to fight some project’s grant recognition and the rounds often devolve into a popularity contest. People just vote for what they know or have heard of and not necessarily what has had the most impact.
I don’t know a good solution here. If we could find ways to solve it … that would be an amazing achievement.
Using the median of votes
I am curious to see how this round plays out. I understand the median was used so that extreme votes and collusion is less likely. Extreme amounts in either way should not reflect on the final result too much.
But the weird thing with the median is it’s counter intuitive. As I posted here
- Project A has 17 votes and the median is 150k OP
- Project B has 90 votes and the median is 145k OP
Even though project B was deemed impactful by 90 badgeholders as opposed to 17 it takes less funding. This is super counter-intruitive.
And for the people who were celebrating their project being in a big number of ballots they don’t seem to understand after 17 it’s not important as the more there is the more possible it is that they include voters who just wanted to give a small amount to a project since they know it pull the median down.
It’s a natural reaction to search for projects you know and without judging the impact give a smaller amount and without wanting to actually penalize them.
Also as abcoathup mentioned in TG, “Projects not making (or not likely to make) quorum are incentivized to promote to every badgeholder so they can get any RPGF.” And this is what happened with the DM/mentions spam in social media.
Still curious to see how this round plays out and all these are things to consider about the methodology.
