Auto-publish Appstorm Agents as SNET Services

chevron-icon
Back
project-presentation-img
Presentation
Soubir Acharya
Project Owner

Auto-publish Appstorm Agents as SNET Services

Funding Requested

$60,000 USD

Expert Review
Star Filled Image Star Filled Image Star Filled Image Star Filled Image Star Filled Image 0
Community
Star Filled Image Star Filled Image Star Filled Image Star Filled Image Star Filled Image 4.1 (7)

Overview

Appstorm is SaaS that creates generative AI Apps and Agents than can reason and perform simple or sophisticated tasks for fun or business use. Appstorm AI apps and AI agents can compose existing foundational models, purpose-built AI models and SNET services. We intend to extend this self-service application to create pure API apps and agents, that can be auto published as an SNET service.

Proposal Description

How Our Project Will Contribute To The Growth Of The Decentralized AI Platform

Appstorm as described here - https://www.appstorm.ai/ - is a simple smartphone acessible AI service that is intended to bring the power of AI to a non-technical mass audience. Appstorm already orchestrates existing SNET services to simply AI app and AI agent creation for a global audience with SNET inside. Adding the power of pushing a button to make the created Appstorm apps themselves a SNET service will make a thousand flowers bloom.

Our Team

We have been in the trenches since 2020 making AI accessible. We are part of SNET community since 2022. The team comprises over 50 years of ditributed computing, product launch and startup experience.  Startups with exits across DevOps, AI, Data Protection and Storage.

Our US patent: US11494171B1 - Decentralized platform for deploying AI models - Google Patents

Appstorm.ai just crossed 10,000 users. We just demonstrated composing SNET services as part of Appstorm App creation, 

View Team

AI services (New or Existing)

SNET Tech Support AI Agent

Type

New AI service

Purpose

To help troubleshoot SNET problems - simple and seemingly intractable ones. This is just a reference implementation. We expect AppStorm users will be publishing their own!

AI inputs

Problem description stack trace or screenshot

AI outputs

Diagnosis and step-by-step troubleshooting instructions.

Language Detection

How it will be used

This is just an example - Appstorm is capable of leveraging and creating pipelines application with one more SNET services as configured subject to availability and licensing.

Company Name (if applicable)

Zero2AI

The core problem we are aiming to solve

The power of generative AI and its applications are unfolding before our eyes. Appstorm is bringing this power to a mass audience with a open-source, multi-LLM playground that has decentralization, trust and the ability to "own" AI apps as its core tenets. Our current project "Composable AI" makes it to easy to create AI apps which can utilize existing SNET services. We intend to bring the same push button simplicity to SNET Service publishing. 

Today SNET service creation involves a myriad of low-level competencies from blockchain, to docker to cloud-ops. The process is especially challenging for AI developers and is an impediment to innovation.

We intend to bring the simplicity "Even a baby can do it" to the service creation and publishing.

 

Our specific solution to this problem

Our specific solution:

  1. Will add a publish button in our UX to allow publishing of the newly created app as a SNET service
  2. Sign-in with Metamask will allow a wallet to be supplied as a payee for the service, as well as supply gas fees for publishing.
  3. After a few minutes the service can be tested vis snet-cli commands (auto generated) or via the Appstorm UI itself 

Project details

It should be evident to anyone following the AI hype cycle powered by LLMs in the last year, and for good reason, that AI "agents" - which are just a short-form for an LLM with tools, where the tools are other task-specific models or an interface to any digital information, either stored in archives or appearing in real-time - will soon be awash in our digital landscape. Appstorm was built to allow the creation of these "agents" by end-users, a vision that was stamped in history by the release of CustomGPTs and the forthcoming GPT store. While the OpenAI release has simultaneous competitive and reinforcing connotations, we now firmly think of Appstorm as the agent playground. These agents, individually, are able to demonstrate reasoning capabilities that surpass most humans, and collectively they have the potential to encompass all "capitalizable" forms of cognition. Which is to say that a group of agents can exhibit a higher level of intelligence than a single agent. This cognitive architecture would take the form of myriad topologies, emergent, designed (human and AI) and both (through a lockstep of design-emergence or guided growth). We believe that the future is a playground where these agents can (a) solve most existing arbitrarily complex tasks through the construction of these cognitive architectures (b) combine to form cognitive architectures which will manifest truly novel applications (c) allow a human to guide the evolution of these cognitive architectures.
The evolution of model ecosystems adds to this thinking, what started out as HuggingFace will soon spill over to more open-source model ecosystems like Replicate, TogetherAI and de-centralized ones like SingularityNet. SingularityNET has the most compelling end-story of a benign Singularity but before we are ready for the Matrix we need composability.
The reason these model ecosystems are brewing is that AI has been gradually seeping into social consciousness, perhaps since 2012 and exponentially since Dec 3rd, 2022 when OpenAI released ChatGPT. More and more talented people from all walks of life - science, art, literature, music, video, games, technology will now pool their collective energy into making AI the "everything interface". If it were possible to interact with the universe of digital information, public or private, in natural language, computing would blend into reality.
This future goal, unconscious in the collective, will push more people to become model developers, push existing model platforms to make the model development process even more simple over time, and the universe of publicly available open-source models to explode. We understood this back in 2020, and it's beyond obvious in early 2024. While OpenAI retains the title of the king with GPT-4, a model that is hard to beat even by behemoths like Google, proprietary gatekeeping of a fundamentally transformative technology like AI is bound to fail. They can get rich, sure, but the dreams of AGI will be achieved in the crucible of chaos, deep within the annals of open-source weights.
The future of Appstorm lies in this chaos. We aim to provide a playground that can emerge higher cognition through multi-agent dynamics. Appstorm absorbs the open-source universe, providing the atoms-to-cognition plumbing, providing a playground where highly complex, and seemingly impossible to automate tasks are mapped to a multi-agent workflow. The user simply interfaces with an agent, the executive so to speak, which will deploy and manage other agents to accomplish the task. This workflow can be guided, or inferred or arrived at through an interleaving of both.
The value is in the topologies (which includes not just shape, but the constituents of the topology, for e.g., a multi-agent system composed of GPT-4 is different from one that has a tier of capabilities, for merely economic reasons) - that are enabled by Appstorm. Any arbitrary LLM in the world can connect to an arbitrary collection of tools, or other narrower models and play the role of a character. In the classical example, a Software development firm can envision a Project Manager, a suite of developers and a graphic artist - each an agent with their very specific capabilities. Appstorm would allow the modelling of this collaborative human enterprise in the multi-agent workflow using natural language. The LLM selections, the tools selection, the narrow model selection would be the "orchestration". On the surface Appstorm "extracts" user intent and provides a cognitive architecture (in the form of the freshly constructed multi-agent edifice) to achieve that task.
In that sense it is just the next generation of infrastructure. Instead of compute and memory, we allocate LLMs, models and their tools. The program is now natural human commands, like "Book a flight for 3pm tomorrow, but find me a ticket < $500 with an aisle seat not too far from the middle of the plane. Send a cab to my place at a time that allows me to reach the airport an hour before the flight. Allow for time to pick up food from MyDeli, an order of pasta in white sauce served for 2. Add my arrival time time to my wife's calendar so she can pick me up at the destination." This fictious request captures a suite of tasks where agents would have to navigate both human and digital interfaces and get stuff done. The result is many magnitudes faster and more efficient than a personal assistant one has trained for decades. Now available to anyone.
With Appstorm, the LLMs, models and their tools are constantly being discovered - as they are published to the community, mapping user tasks to an every-changing state-of-the-art cognitive substrate, tuned by price and/or performance to a standard (and potentially bespoke) metric. The agents could be passive, consuming information from any internet reachable endpoint, ping-ponging it through a multi-agent architecture and spitting out the results to another custom defined, hardware and/or software interface. Or it could be an active agent, guiding human thought, interleaving large chunks of artificial cognition with bits of human imagination. At scale this interleaving emergers architectures that are inconceivable now.
In the business world these assistants would ramp up efficiencies, make humans redundant for all but the most hard-and-costly-to-model cognition. 

Competition and USPs

There are other teams that are working on this problem with impressive progress but we expect out efforts to be complementary and lower the barrier to entry for an audience who would otherwise not be consuming or creating SNET services. New API updates, techniques etc will also be incorporated as they become available.over time. We expect the ease of use (and the ease of building) will make accelerated adoption of a new class of AI Apps and Agents along with SNET services possible.

Needed resources

We would urge the community to test and give us feedback even if our efforts are at an ealry stage. We believe our efforts will accelerate most projects, especially the on-boarding phase. We are already working with community members in this regard.

Existing resources

We will be leveraging our past work and existing software efforts without which this would be an impossible task. This is also reflected (and perhaps explains) our modest budget. We have already early adopters who are using our cli sandbox and our AI landscape guidance as they build.

Open Source Licensing

Custom

The project is partially open-source today but will be fully open-source (either GPLv3 or MIT) over time.

Revenue Sharing Model

Token Allocation

We are planning for a token launch in:

2024-Q4

Token Description (type, value, utility):

Incentive AI token. Reward Appstorm Creators and Consumers. Stay2Earn - rewards for engaging with the network. Token appreciation tied to community size and transaction volume.

Proposal Video

DF Spotlight Day - DFR4 - Soubir Acharya - Auto Publish Appstorm Agents As SNET Services

4 June 2024
  • Total Milestones

    6

  • Total Budget

    $60,000 USD

  • Last Updated

    4 Jun 2024

Milestone 1 - API Calls & Hostings

Description

This milestone represents the required reservation of 25% of your total requested budget for API calls or hosting costs. Because it is required we have prefilled it for you and it cannot be removed or adapted.

Deliverables

You can use this amount for payment of API calls on our platform. Use it to call other services or use it as a marketing instrument to have other parties try out your service. Alternatively you can use it to pay for hosting and computing costs.

Budget

$15,000 USD

Milestone 2 - Design and User documentation

Description

Technical design and user documentation This will describe the technical architecture of the solution as well as the user experience using the service.

Deliverables

New GitHub repo with documentation and a how-to guide.

Budget

$10,000 USD

Milestone 3 - TESTNET implementation of Tech support AI Agent

Description

Tech support AI Agent - reference implementation on TESTNET

Deliverables

GitHub update Demo video.

Budget

$10,000 USD

Milestone 4 - Metamask payment integration - TESTNET

Description

Metamask integration payment flow from payer to payee on TESTNET

Deliverables

GitHub update Demo video with detailed payment flow and MetaMask interactions.

Budget

$10,000 USD

Milestone 5 - Metamask payment integration - MAINNET

Description

Metamask integration payment flow from payer to payee on MAINNET

Deliverables

GitHub update Demo video with detailed payment flow and MetaMask interactions.

Budget

$8,000 USD

Milestone 6 - MAINNET implementation of Tech support AI Agent

Description

Tech support AI Agent - reference implementation on MAINNET

Deliverables

GitHub update Public release. Final documentation

Budget

$7,000 USD

Join the Discussion (4)

Sort by

4 Comments
  • 0
    commentator-avatar
    Gombilla
    Jun 2, 2024 | 3:30 PM

    Hey there. Kudos to your team.  I think this project has the potential to foster collaboration, creativity, and experimentation within the SNET community, ultimately advancing the development and adoption of AI technologies for the benefit of society.

  • 0
    commentator-avatar
    CLEMENT
    Jun 2, 2024 | 3:21 PM

    Great Job putting this up. I am positive this will bring much benefit to the SNET AI community.  Kudos to you and your team ! Also, you are also welcomed to make comments on our team proposal as well https://deepfunding.ai/proposal/4757/  - AI4M (Enhancing Malaria Predictability using AI) https://deepfunding.ai/proposal/biotek-nexus-next-gen-biodiversity-conservation/  - BIOTEK NEXUS (Blockchain Biodiversity Conservation)

  • 0
    commentator-avatar
    HenriqC
    May 20, 2024 | 9:23 AM

    Does it mean that 10k users have created an AI service by using the Appstorm? Can they let others use them through your app?  Anyway, this might indeed be a potential way to get the number of the services up on the SNET platform.  

    • 0
      commentator-avatar
      Soubir Acharya
      May 26, 2024 | 2:30 PM

      Yes.  These are not SNET services yet. With the feature add the creator (if they so chose) could publish them as a SNET service. Thanks.

Reviews & Rating

Sort by

7 ratings
  • 0
    user-icon
    CLEMENT
    Jun 2, 2024 | 3:27 PM

    Overall

    5

    • Feasibility 5
    • Viability 5
    • Desirabilty 5
    • Usefulness 5
    Will enable users publish their AI applications

    Hi Soubir. I must commnd this great innovation. For me, this project represents a significant advancement in democratizing access to AI services within the SingularityNET ecosystem. I like your approach of integrating with Appstorm, a SaaS platform for creating generative AI Apps and Agents, and I affirm this project will enable users to easily publish their AI applications as SingularityNET services.

    As a contribution to the SingularityNET AI community, I also believe your Auto-Publish Appstorm Agents will enrich the ecosystem by bringing in a diverse range of AI applications and services created by users from various backgrounds and expertise levels.

    Kind regards !

  • 0
    user-icon
    Nicolad2008
    Jun 8, 2024 | 11:19 AM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 4
    • Usefulness 4
    quality and reliability of AI applications

    I think the project has great potential in simplifying the application process and AI agents for amateur people. Automation ability will help many people approach and create AI applications with just one click, promoting the creativity and sustainable development of AI technology. However, I am concerned about the quality and reliability of AI applications created, as well as a strong user support system to solve technical problems. Although this project promises many prospects, quality assurance and support for users is very important to achieve long -term success.

  • 0
    user-icon
    TrucTrixie
    Jun 9, 2024 | 2:16 PM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 3
    • Usefulness 4
    The fun of Appstorm

    In my opinion, the ecosystem will be enriched the way Appstorm Agent publishes automatically - then many diverse AI applications will be created by users. It seems that there are quite a few highly usable proposals for users when the users themselves are directly involved in the product creation process. Interesting!

  • 0
    user-icon
    Max1524
    Jun 9, 2024 | 2:50 AM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 4
    • Usefulness 4
    The team knows how to simplify the complex

    The Appstorm construction plan is presented quite reasonably and I rate its feasibility as relatively high. Milestones have contributed to saying that. The key here is that the team knows how to present and simplify the complex so that readers can understand easily. I appreciate this.

  • 0
    user-icon
    Joseph Gastoni
    May 22, 2024 | 3:54 PM

    Overall

    4

    • Feasibility 4
    • Viability 3
    • Desirabilty 3
    • Usefulness 4
    a platform for creating AI apps and agents

    This proposal outlines "Appstorm," a platform for creating AI apps and agents using generative AI models and SingularityNET (SNET) services. Here's a breakdown of its strengths and weaknesses:

    Feasibility:

    • Moderate-High: The core concept is feasible, but technical complexity and user adoption pose challenges.
      • Strengths: The use of existing frameworks and focus on user-friendliness simplify development.
      • Weaknesses: Integrating various AI models and SNET services seamlessly requires significant technical expertise.

    Viability:

    • Moderate: Success depends on user adoption, the value proposition for developers, and integration with existing AI ecosystems.
      • Strengths: The proposal addresses a growing interest in user-friendly AI development.
      • Weaknesses: The proposal lacks details on the business model and how to ensure the quality and security of user-created apps.

    Desirability:

    • Moderate-High: For non-technical users interested in creating AI applications, this could be desirable.
      • Strengths: The proposal offers a low-barrier entry point for AI development.
      • Weaknesses: The proposal needs to clearly explain the limitations of user-created AI and potential privacy concerns.

    Usefulness:

    • Moderate-High: The project has the potential to democratize AI development and expand the use of SNET services, but its impact depends on user engagement and the quality of created apps.
      • Strengths: The proposal offers a platform for building and sharing AI applications.
      • Weaknesses: The proposal lacks details on evaluating app quality and ensuring responsible AI development practices.

    Overall, the Appstorm project has an interesting approach, but focus on:

    • Technical Integration: Clearly define the process for integrating various AI models and SNET services within Appstorm.
    • Business Model: Develop a clear plan for generating revenue and incentivizing developers to create valuable apps.
    • Quality Control and Security: Outline strategies to ensure the security of user data and the quality of AI apps created on the platform.
    • Responsible AI Development: Integrate features that encourage responsible AI development practices, such as bias mitigation and transparency.

    Strengths:

    • Lowers the barrier to entry for creating AI apps and agents.
    • Leverages existing AI models and SNET services.
    • Offers a user-friendly platform for non-technical users.

    Weaknesses:

    • Requires careful integration of various AI components.
    • Needs a clear business model and quality control measures.
    • Must address data security and responsible AI development.

    By addressing these considerations, Appstorm can become a valuable tool for promoting accessible AI development and fostering a vibrant SNET developer community.

  • 0
    user-icon
    Tu Nguyen
    May 23, 2024 | 3:38 AM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 3
    • Usefulness 4
    Auto-Publish Appstorm Agents As SNET Services

    This proposal addresses the challenge for AI developers, who find it difficult to create SNET services that involve a multitude of low-level capabilities from blockchain, to docker to cloud operations . This proposal also provided a suitable solution to solve the problem. 
    Other opinion: They should share more clearly the differences of their solution compared to other solutions on the market. They should also determine the start and end times of milestones.

  • 0
    user-icon
    BlackCoffee
    Jun 10, 2024 | 1:15 AM

    Overall

    4

    • Feasibility 4
    • Viability 3
    • Desirabilty 4
    • Usefulness 4
    What AI developers need

    This proposal is expected to bring many benefits to the community, I personally think so, especially helping AI developers when they are having difficulty creating SNET services related to blockchain. That is the usefulness of this proposal.

Summary

Overall Community

4.1

from 7 reviews
  • 5
    1
  • 4
    6
  • 3
    0
  • 2
    0
  • 1
    0

Feasibility

4.1

from 7 reviews

Viability

3.9

from 7 reviews

Desirabilty

3.7

from 7 reviews

Usefulness

4.1

from 7 reviews

Get Involved

Contribute your talents by joining your dream team and project. Visit the job board at Freelance DAO for opportunites today!

View Job Board