Zcash AI drafting tool

Quick question before we commit: self-hostable or accessible — which matters more for a Zcash AI drafting tool?

We’re building a read-only AI agent to help Zcash users draft better grant proposals and navigate the ZCG process. No wallet access, no trading, no automatic posting — just drafting and research assistance.

Before we commit to a runtime, we have a genuine choice to make and want community input before deciding:

Option A — Developer-facing runtime. More auditable, self-hostable, open-source. Steeper setup for non-technical users.

Option B — Consumer-facing runtime. Easier to use out of the box. Less control over where prompts and data go.

For a tool aimed at Zcash users, which tradeoff matters more to you?

And the harder question: would you actually use or recommend a Zcash AI tool that isn’t fully self-hostable, even if it’s useful?

Cheers,

Groggs

1 Like

Have you consulted @ZCG if they actually want to have more AI content in the applications they have to review?

A project should stand on its own merits of what it’s trying to accomplish for the zcash community, not based on how well an applicant can draft their proposal.

2 Likes

+1

In my opinion, the use of AI for proposals is a red flag and when I see things like that, it’s usually a good (but not basic) reason candidate for my rejection.

Thanks for the feedback.

A bit of background on myself. Im groggs from Ireland. Near protocol has been my “home” chain since January 2022 as the first place that I opened a wallet onchain.

On a professional level im a qualified accountant and since moving back to Ireland in January ive worked for a large tech multinational in Dublin.

The advancement of AI within the company has gone exponential such that if I didnt use AI in my day to day operations I would be inefficient. We all had openclaws but now they are replaced by a central AI system that literally knows everything. I drag and drop and I get the answer. My professional qualification really doesmt matter anymore.

And yet I am absolutely confident in walking in to a position when my contract ends in 6 mths knowing that the AI hype will engulf the rest of the finance industry over the next few mths.

Im outlining all this to give you some context around how important I now value my time with regards to learning everything I can around AI.

And that leads on to zcash community and grants in particular and exploring creating an AI agent that can be of value to the zcash community and ofc, in so doing, return value to myself.

An AI agent is only as good as the corpus of data upon which it can access aswell as the rules you give it to follow. With good information and training it should be able to reduce the time taken to undertake certain processes.

The zcash community has alot of open source available information including this forum and previous and current grant proposals that could form the knowledge base for an agent to produce insights around the grant process that ultimately could assist either the committee in reviewing proposals or applicants in creating them, of course without reducing the expected quality.

Ultimately I want to research as much as possible where this agent could be of most value and would be grateful if the committee could direct me to further information around how much time it spends on reviewing proposals or any ideas on how such an agent could be of benefit to them in assisting to reduce the overall time spent on such.

I welcome the community’s input on ideas around the above

Thanks

(Im on the train home after work so excuse the grammatical errors)

Today, we all use AI in one way or another as a useful tool in areas where we have expertise. But proving that someone is truly an expert in a particular field is another matter entirely. And AI certainly isn’t much help with that.

Im not trying to convince you one way or the other.

Im just interested to hear ways the community think AI could reduce the time spent on certain processes.

For example, a committee members remuneration appears to be, at current Zec prices, 7kusd a mth × 5 × 12mth = 420k usd a year.

Is there somewhere where we can see the breakdown of that figure?

How much time is spent on the different responsibilities?

If you’re thinking of using AI to decide on grant applications, that’s a bad idea. I, for example, use ChatGPT Pro for my day-to-day tasks. My main and test Zebra nodes update automatically, I receive real-time updates on on-chain activity, I update the Zcash website using AI, and I write various scripts and do other things. But when it comes to ZCG’s work specifically, the AI would approve everything I’ve rejected. We don’t see eye to eye on everything. Although it does help me find information about applicants’ involvement in other grant schemes. Yes, AI is excellent at identifying grant hunters.

1 Like

Of course. An AI “assistant” assists, the human manager decides.

Its good to hear you are a pro user as im sure you well know, the time efficiency gains that this allows you.

This is a valuable conversation to have for the community:

In what ways can responsible AI use benefit the zcash community?