In the coming days and weeks, you may notice ZCG closing grant applications earlier and issuing faster rejections. This change is being implemented to prevent the application process from becoming a drain on both the committee and the broader Zcash community.
While the committee takes no issue with AI being used as a tool during the application process, we are finding that the traditional signals used to evaluate grants are becoming less reliable due to an influx of low-quality, AI-generated submissions. To address this, ZCG will begin closing grants early when necessary. Going forward, the burden of proof is entirely on the applicant to demonstrate, immediately and clearly, that they can add value in line with community expectations.
Regarding enforcement, we are rejecting @EmilioNM and @sudo-julien from the Grants program effective immediately. We also recommend both usernames be banned from the Forum for violating the Forum Code of Conduct.
Undisclosed Connections: We discovered these accounts have ties yet chose to submit separate applications without acknowledging their relationship.
Code of Conduct Violations: Prior to this discovery, we received reports of inappropriate conduct by Emilio983 involving unsolicited private messages to community members regarding their applications.
Thank you for your understanding as we refine our processes. We remain committed to making the Zcash Community Grants program a successful and efficient funding body.
Thank you very much for this, and thank you, @Shawn, for the clean-up.
The attacks on the lockbox have not even started, and I expect participants to become more sophisticated, draining and in larger numbers.
I’m not sure about my next statement, but maybe the committee will need to consider more drastic measures, such as grants requiring social sponsors from a trust graph, or even “paying” a deposit.
I’ve chatted with both of those Nyms, and eventually advised both of them that I didn’t have the capacity to review their work more closely.
I also specifically, directly, publicly, suggested that they alter how they relate to LLMs.
I also acknowledge that their behavior seems likely to have had significant costs for the ZCG, which is frankly poorly formatted for the current investment landscape.
That’s the real issue
Our environment has shifted, we need to adapt to that.
I don’t have enough context to feel comfortable analyzing this specific decision, but I will, without invitation or permission, analyze this popular authoritarian action.
Popular Authoritarian Dictats
We face a problem of agency distribution. That problem can be approached in a range of ways that I think of as falling along a continuum that I learned about on EconTalk.
We can push authority to centralized information processors, or we can push information to distributed authorities.
The former pattern of outsourcing our authority to empowered central authorities is unambiguously the popular thing to do these days.
I don’t believe that popular is the same as right.
How about focusing on the content of the argument, and each of us developing our ability to reason?
OR we can clench like a boxer who’s taken a shot to the kidney.. desperately trying to avoid waste, and in the process destroying the value that comes from engagement.
The process of displaying the corpses of offenders to dissuade subsequent malfeasance took the form of crucifiction in Rome.
And… Rome was/is pretty successful!
Oh.. AND.. of course, the public display and humiliation of the torture and corpse was of course wildly popular.
Unfortunately, I think it will become increasingly harder to distinguish between genuine input and AI-generated data.
For the simple reason that many of the Zcash-related conversations happen on Twitter/X and Grok is allowed to access every post in order to aggregate the vocabulary and the main talking points.
So if anyone ever posts anything about Zcash, it becomes a data source for LLMs. Which is cool in some ways, but really unfortunate when you must assess the value of human contributions to a project.
Maybe that this type of written communication is antiquated and bound to be overtaken by an AI that’s becoming indistinguishable from a real human being. Maybe that voice/video communication is a better way to vet candidates for various grants. But then again, audio and video are becoming AI’s playground too and you cannot ask everyone to reveal their face and voice in a privacy project.
I am somewhat glad that a degree of content moderation works for now, but I expect this to become increasingly difficult and resource-intensive. We’re living very interesting times!