AI monitoring for projects using Dapp Staking - Proposal

General
1

Hello, everyone!

The purpose of this post is to propose AI-based project monitoring in light of the current situation where projects utilizing Dapp Staking are being delisted one after another due to community dissatisfaction.
This aims to minimize the financial damage to the community.

Three years ago, when Astar was released, this would have been unimaginable, but now AI-driven development has become mainstream in programming.
If AI development is possible, I believe it should also be feasible to monitor project progress rates using AI.

I understand that some projects cannot be open-sourced due to their confidentiality. However, by granting AI systems like Deep Seek or Claude project read permissions within a closed on-premises network, I believe AI can report on project development progress while maintaining confidentiality.

That’s right. There is no need to develop AI on your own. If this can solve the issue of the community not being able to track project progress rates, I see no reason not to do it.

I propose building an AI that is restricted by a firewall to only read GitHub and report to the community.

I look forward to hearing the community’s opinions on this proposal.
Thank you for your consideration!

2 Likes
2

Hey hello @AstarPunks, I think the idea is quite nice but not necessarily a priority at the moment.
From what I understand you mean the creation of a tool that allows you to monitor various dApp metrics and now we are in a delicate phase where we aim to delist several inactive dApps, so spending time and resources seems unnecessary to me.
We already have the Astar Community Council where various members are monitoring the statistics (AstarHood and Mounthmouth).
I believe that a similar tool could be helpful in the subsequent phases and I would like to ask you how you imagine it, in the sense will it be a Twitter AI agent? A web page with all the updated data? I would like to delve deeper into this idea of ​​yours :slight_smile:

2 Likes
3

Hello @VasaKing san

Thank you for your response.

I apologize for my lack of understanding.
It seems the Astar Community Council already has a tool for monitoring statistics.
If the existing monitoring tool allows dapp stakers to know the development status in real-time, this proposal may be unnecessary.

The fundamental issue I’m concerned about is that dapp stakers lack a way to know the development status, which results in insufficient information to decide whether to change the dapps they stake.

This issue seems to stem from a combination of the following problems:

  • Dapp stakers may not have access to GitHub, making monitoring impossible in the first place.
  • Many dapp stakers are not developers, so they cannot understand the development status even if they see the source code.

The tool I envisioned would:

  • Have AI read whitepapers or forum posts aimed at dapp staking listings.
  • Require projects to grant AI access to GitHub as a mandatory rule.
  • Have the AI analyze GitHub commit history and the overall source code, compare it with the whitepaper content, generate a progress report, and post it on forums or Discord.
  • Post alerts on forums or Discord if there’s no development progress for a certain period.

combining reports on X post activity could make this even better!

1 Like
4

I find your proposal fantastic, but I believe it may be somewhat limited in terms of effectively measuring development.

From what I understand, it would function as a GitHub and code tracker. While this approach can be useful, it only goes so far — development extends beyond lines of code. Without community building, marketing efforts, or the generation of real value, code alone can quickly become irrelevant.

That said, I truly appreciate your initiative and would love to hear more about how you envision this tool, especially considering that it will rely on teams to supply the data. Perhaps there’s a way to incorporate a more holistic perspective of project growth?

1 Like
5

@pitcoin777 san
Thank you for your response.

I’ve always been impressed by your achievements, so I’m truly delighted to have this opportunity to talk with you.

I agree that development goes beyond just lines of code.

Some dishonest developers seem to commit unnecessary data to make their commit history appear more substantial.

Therefore, I propose using AI to analyze whitepapers and compare entire codebases.
This serves as a metric to determine whether a project is truly progressing toward its goals, rather than relying solely on code trackers.

Regarding how to provide data to the team,
I envision using a scripting language like Python to generate prompts for AI, read the results, and post them to Discord or forums.

I plan to build a lightweight AI, such as Deep Seek, that doesn’t require high-performance GPUs in a closed on-premises network environment.
Within the same environment, a Python-based program would create an interface with the community.

As a developer myself, I’m reluctant to suggest measures that involve monitoring other developers.
Ideally, I’d prefer to propose ways to increase the number of projects that achieve their goals.
However, the only idea I have for this is to build and energize a local developer community.

If possible, could you share details about the functionality of the monitoring tools currently held by the community council?
Additionally, I’d like to know if there’s a way to share the information from those tools with Dapp Staker in real-time.

2 Likes
6

Great idea, but we need more precise and complex solution to make this work.

Let’s think more!

7

Hello, @Matt san

Thank you for your interest in this idea.
Please forgive me, as I was unaware of the council members’ monitoring tool when I posted this.

I don’t expect the current proposal to work perfectly right away, but I believe it could provide some rough metrics. My concern is that without even these rough metrics, Dapp Stakers might face sudden delisting, leading to community confusion and frustration, which I find regrettable.

Therefore, I’d like to know the details of the functions of the monitoring tool held by council members and whether the metrics it provides can be shared with Dapp Stakers in real-time.

8

Hey @AstarPunks , as far as I know the community council members don’t use AI tools to check metrics but there is still a manual check method. At the moment there is not a lot of data or dApps to analyze daily and I think it’s fine now. In the future with the expansion of the Astar ecosystem and more interest I think a similar tool would be very useful. The various possibilities should be explored with AI agents, I can’t wait to see the winners of the recent hackathon promoted by Soneium :star:

1 Like
9

Hi @VasaKing san

Thank you for providing the information.
I somehow misunderstood, thinking there was a monitoring tool.
My apologies.

I was thinking that sharing the method for checking those metrics with the community might help reduce some confusion.
(As for the Japanese community, the confusion and anxiety still doesn’t seem to have subsided.)

I’m aware that a hackathon for AI agents is being held on Soneium, but personally, as Gavin Wood predicted, I have doubts about Ethereum’s future potential and have been hopeful about the development of Astar L1’s ecosystem using Polkadot.
I also wanted to do something to help Astar L1.

However, as you mentioned, it might be better to focus on events like the Soneium hackathon.
(Soneium’s Discord has a member limit, and I can’t join, plus I can’t listen to or speak English, so I’m quite anxious…)

1 Like
10

Hi @AstarPunks

I find your idea/initiative very appealing as a way to improve our project review flow and create efficient summaries that allow us to monitor dApp progress based on their GitHub contributions, I really like the concept!

However, I do have a few questions. We should keep in mind that a project doesn’t progress solely through technical development. User acquisition, marketing campaigns, social media metrics, and other aspects are also vital when evaluating the “real progress” of a project. Do you have something in mind that could gather all of that data?

Secondly, how do you envision it? A dashboard showing each dApp and their metrics? A report-style document with all this data? What kind of structure would it have?

Although my teammates have mentioned that this isn’t a current priority, there’s no reason why you can’t keep working on this idea you’ve brought up. Go ahead! Build a PoC and share it with the community here.

Thanks for contributing to Astar!

4 Likes
11

Hi @Juminstock san

Thank you for appreciating my idea.

I deeply respect your efforts in cultivating the developer community, and I dream of cultivating a developer community in Japan and building an Astar developer online community in Japan.

I look forward to your continued guidance.

I will answer your questions.

As you say, I recognize that the true progress of a project lies in development progress, user acquisition, marketing campaigns, and social media metrics.

I have considered how to obtain these metrics.

As for development progress, as mentioned above, I propose to analyze white papers and forum posts to list them and compare them with the entire code base.
Whether this is feasible requires building and verifying AI, but I believe it is logically possible, so I would like to try it.

As for user acquisition, I propose to identify the project’s contract address and the number of associated wallets.
I currently work as an engineer in the Web2 world, so my knowledge in this field is limited. However, I believe that there is a high possibility that these data can be obtained by running an Astar node and accessing it via API. Please let me know your thoughts.

As for social media metrics, I think it will not be too difficult to obtain them since various program modules already exist.

As for the interface, since I have limited knowledge of front-end development, I proposed a report format that is easy to build. However, a dashboard format is preferable.

Ideally, Dapp Stakers should be able to check real-time progress via an API from the Astar Portal’s dashboard.

As for building the PoC, since I work as a full-time engineer at a Japanese company, I think I can gradually verify the feasibility using my free time.

I’m encouraged to have comrades who can work together to improve Astar.
Looking forward to our continued collaboration.

1 Like
12

Will projects have to constantly send their data to the AI system or will it be received regularly by an autonomous system?

Obviously this is a very sensitive issue.

As a data source, you could perhaps use publicly available information or regular reports of projects (which are also publicly available). Based on this data an AI agent (or web-based AI) could be created.

This does not mean that I am currently positive about this topic.

There is an interesting proposal and I am thinking out loud what can be done with it.

1 Like
13

Hi @MrKarahanli-Emre san
Thank you for being interested in my proposal.

Projects do not need to send any data to the AI system.

In the first place, I’m making this proposal based on the dissatisfaction within the Japanese community, which stems from numerous experiences where projects turned out to be untrustworthy.

Therefore, I cannot trust any data sent directly from the project.

What we should require from projects are only things like read-only access to their GitHub, social media account information related to their activities, and the contract addresses of their projects.

14

I like the idea you’ve presented. Currently, I think AI is quite beneficial for both personal and work matters. For this idea, I’d like to ask a bit more. I don’t have much technical knowledge. In terms of cost, I understand that developing AI will have a certain expense, which may require revenue to support the break-even point. So I’m not sure what your thoughts are on this issue. =)

1 Like
15

Hi @BoomBLB san,

Thank you for raising your concerns about the cost.
I share your concerns as well, but to be honest, I won’t fully understand the actual costs until we build a PoC.
While I love working with technology, I’m unfortunately not very good at cost estimation.
(Building a PoC will cost money, but since I’m not sure if it will succeed, I’m thinking of covering the expenses myself.)

That said, I don’t intend to build the AI from scratch.
I plan to select an existing lightweight AI (such as DeepSeek, Claude, etc.) that can achieve our objectives.

The main development will likely focus on collecting data such as development progress, user acquisition, and social media metrics.
These pieces of information will then be passed to the AI, which will generate simple reports.

If we decide to integrate this into the Astar Portal, API development and other related work will also be necessary.

Once it’s built, the only ongoing cost would likely be server fees for something like a VPS.
However, I haven’t completed a detailed analysis of the required server specifications, so I can’t give a definitive answer at this point.

I’m just a Dapp staker, not an agent, so it’s indeed questionable that I receive no benefit for contributing to Astar…
It seems to go against Astar’s vision of improving incentive structures for developers.

@Juminstock san,
I’m considering applying for certification as a technical agent in Japan.
Do you think I might be qualified?

I am eager to develop on Astar and contribute to building the online development community in Japan. I am also building connections with talented Japanese engineers who are interested in developing on Astar (although the number is still small).