Review of Optimism retro funding round 1
Public.icon
$1 million was allocated to 58 projects.
22 people were badge holders (verified) and acted as experts.
This is how the grants were ultimately distributed.
https://gyazo.com/f7253fcf83f73a825b54a03fe493b25a
https://gyazo.com/e8a4b3bca87b92291d26b65d001ebc8d https://gyazo.com/f7253fcf83f73a825b54a03fe493b25a
Probably the most obvious property of the Optimism retro results that can be seen without any comparisons is the category of the winners: every major Optimism retro winner was a technology project .
There was nothing in the badge holder instructions that specified this; non-tech projects (say, the translations at ethereum.cn ) were absolutely eligible
And yet, due to some combination of choice of badge holders and subconscious biases, the round seems to have been understood as being tech-oriented.
Hence, I restricted the Gitcoin results in the table above to technology ("DApp Tech" + "Infra Tech") to focus on the remaining differences.
One of the most obvious characteristics of the results of the Optimism retro is the category of the winners, which can be seen without comparison.
>There was nothing in the badge holder instructions that specified this; non-tech projects (say, the translations at ethereum.cn ) were absolutely eligible.
>And yet, due to some combination of choice of badge holders and subconscious biases, the round seems to have been understood as being tech-oriented.
>Hence, I restricted the Gitcoin results in the table above to technology ("DApp Tech" + "Infra Tech") to focus on the remaining differences.
The distribution of funds using RetroPGF had (smaller) income disparities compared to Gitcoin many already well-known projects
As originally intended, many of Gitcoin's votes were predicting the future.
Infrastructure was popular.
Subjectively, it was of high quality according to Vitalik. The number of badge holders was small, but they made more appropriate judgments than the crowd. Judging this requires a Chinese background and technical understanding.
The badge holders were already familiar with the English-speaking world, so how to evaluate this?
The country selects various experts.
Several solutions are needed.
Splitting votes
When going back in time to pay money, is it optimal to add more? How much money should be paid?
https://gyazo.com/a5d97104f687a29fa43d40882294f44d
One possible solution is to impose a deposit of 0.02 ETH on proposals
If a proposal receives 0 votes, it is considered spam and the deposit is lost
Proof of Personhood (PoP) ID is required to prove that you are human
Chat thread
yoyotkgshn.icon*2
The part I'm translating right now is the Notion article at the top with the indentation
yoyou7693.icon*2
I'm going to bed soon. Thank you for staying up late to discuss Takagiu7693.icon
+1 My skin is ruined even though I have a presentation tomorrow wtkgshn.icon*2
Good luckｗu7693.icon
I want to do this occasionally and asynchronously like thisPrevious companies, interviews and appearances related to oneself.icon
+100 u7693.icon
I'll give it a read
The idea is to evaluate and distribute rewards for projects that have provided value as public goods in the past Is this what it's all about? u7693.icon
Ah, I kind of get the image now wPrevious companies, interviews and appearances related to oneself.icon
It seems to be evaluated by experts? (I don't know much about the Nobel Prize)
There's a vague similarity in the atmosphereu7693.icon
It's still not favorable for basic research, right? What field is appropriate for evaluating the past like this? Is there anyone who can evaluate basic research? Previous companies, interviews and appearances related to oneself.icon
It seems like later on, you might think "this was important." u7693.icon
I think Bitcoin is the same way.
But, when it becomes important, can't you recommend it then? Previous companies, interviews and appearances related to oneself.icon
That's true. u7693.icon
It's difficult to raise funds until you're evaluated. u7693.icon
He is familiar with the ecology of the desert locust, which causes large-scale locust plagues in Africa. Based on field surveys in the Islamic Republic of Mauritania and the Kingdom of Morocco in the western Sahara, he is engaged in elucidating the phase polymorphism of the desert locust (which changes color and becomes ferocious when it forms a group, and devours plants and crops) and developing control techniques.
Assuming that this won't scale, maybe some central organizations will have to create something like RetroDAO and talk among experts. Previous companies, interviews and appearances related to oneself.icon The RetroDAO should make funding decisions solely based on the value that a project, or a contributor to a project, has already provided, and should maximally avoid attempting to make predictions of value to be provided in the future.
If we should avoid making predictions about the future, I don't really understand the meaning of human evaluation. u7693.icon
Wouldn't it be better to evaluate mechanically?
Like citation counts? Previous companies, interviews and appearances related to oneself.icon
We were just talking about this earlier.
Oh, where? u7693.icon
I was talking to Sota Ogawa in person, but he had a similar reaction. "Then should we write code to evaluate embedded systems?" or something like that. Previous companies, interviews and appearances related to oneself.icon When humans do it, the evaluation axis shifts and emotions come into play. u7693.icon
Citation counts are the same way, so it's better to algorithmically determine the evaluation method in advance. u7693.icon
Regarding things like "If you define it, it will be hacked," do you have any thoughts? Previous companies, interviews and appearances related to oneself.icon
It's good to think about this, but it's also good to consider the possibility of the evaluators being hacked. u7693.icon
nisshi yui79 Version 1 visuals, lol, brings back memories tkgshn.icon*4 Hmm, maybe being conscious of the structure is one way. Like visualizing the interests?
Thorough transparency could work too.
It's not just about interests, but also appealing to more pure emotions (hobbies, preferences, etc.) u7693.icon
Nishio-san summarized it. u7693.icon
Ishii-san's opinion is also interesting. u7693.icon
I think as the number of OSS subject to evaluation increases, emotional filtering will become more prevalent. u7693.icon
Well, maybe a feature like being able to change the jury? Previous companies, interviews and appearances related to oneself.icon
This issue can be solved at the software layer, but we should also divide deliberation democratically. It's difficult to draw the line though. Court-like functions don't scale, after all. Previous companies, interviews and appearances related to oneself.icon
Using machines as the basis for evaluation and adding human evaluation as a plus alpha seems like a good idea, doesn't it? u7693.icon