A new Deep Funding Pool: ‘SingularityNET RFPs’

Author: jan Published: July 31, 2023

We are excited to present a new Pool aimed at ‘SNET RFPs’ or, written in full, ‘SingularityNET Requests For Proposal’ centered around requests created by the Deep Funding Staff and SingularityNET tech team. With this new format, the SNET tech team is inviting community teams to collaborate on specific features required for the platform and support of Deep Funding. Where the current ‘Pool E – tooling’ is -and remains- open for any proposal suggestion by the community, the new ‘Pool S – SNET RFPs’, will give direction to specific features requested by the internal team. 

Importantly, the launch of this new pool is organized against the backdrop of an emerging renewed platform strategy, where we are preparing for the coming AGI revolution powered by emerging Neuro Symbolic LLMs leading to a full-fledged OpenCog Hyperon! 

A renewed platform strategy

We have always wanted to collaborate more closely with the community on our core platform technology. Recent in-depth conversations on preparing the platform for our swiftly emerging AGI tech stack have given this idea a new push forward. We decided to delay the new DFR3 for a few weeks so we could incorporate this new pool and include the community in our platform efforts in the shortest possible term!

While cannot describe all details of our new strategy in an article that is dedicated to the new Deep Funding pool and the related RFPs, we are very pleased to share these high-level outlines with you:

Besides some known features that we will speed up, such as onboarding and pricing options, we will prepare our platform for external Memory and Neuro Symbolic Enhanced Large Language Models. We will be adding tools for the creation and editing of diverse Knowledge Graphs. Next to this, we will be adding at least one dedicated LLM API, as well as tools for the community to add different foundational and fine-tuned LLMs. On top of these 2 pillars (Knowledge Graphs and LLMs) we are providing tools that enable developers to create solutions that will enhance the native strengths of LLMs.

This approach has a number of important benefits and synergies:

  • Most of the tools and technologies planned for this platform-upgrade are being developed already or will be perfectly synergized with work on our AGI roadmap, such as MeTTa, DAS (Distributed Atom Space), and architectures for Zarqa or other Neuro Symbolic LLMs
  • We want to offer the broader community of developers the opportunity to start building on top of our tech stack and getting used to some technologies such as MeTTa, DAS, and Knowledge Graphs in general. 
  • We will be creating a fertile ground of tools, strategies, and developers for the launch of our own Neuro Symbolic LLMs, which will be super-optimized to make use of the offered tools.
  • Last but not least, with your help, we will be building:

The Blockchain Layer 3: The Internet of Knowledge, a Key Ingredient of Beneficial Decentralized AGI

How will this new pool work?

The process is well aligned with the current pools, but there are a few key differences:

  • The pool will not have a maximum size, but each RFP will have its own maximum awarded amount.
    • Some RFPs will require a single outcome. In those cases, only the highest-scoring proposal will be awarded. 
  • Other RFPs could be the source of multiple solutions. In those cases, the RFP will mention the maximum per proposal as well as the maximum for all proposals awarded for this RFP. In case there are more community-approved proposals than funding, the highest ones will be awarded as far as the maximum budget allows. 
  • When a team submits a proposal they will clearly state to which RFP the project is related.
  • During the ‘Stable Period’, the SingularityNET tech team will review the submitted proposals and recommend the ones they have sufficient confidence in. 
  • Finally, as always, the community decides which proposals are awarded. For this round, we will keep the same thresholds as we are used to: Each project should be voted upon by at least 1% of the voters and the average grade should be 6,5 or higher.
    The community can vote on all proposals, both the ones recommended and not recommended by the SNET tech team. 


  1. SingularityNET will publish one or multiple ‘RFPs’. (Requests For Proposal)
  2. When the new round opens, anyone can submit their proposal(s) to one or more of the RFPs. 
  3. The community can give their feedback and the submitting teams can still make improvements based on this feedback.
  4. During the Stable Period, the proposals cannot be changed anymore and the SNET Tech team will review the proposal and make their recommendations.
  5. The community will cast their votes.
  6. We will publish the results, including the Proposals awarded for an RFP and the development can start.

Overview of the RFPs available for DFR3:

Community Engagement Score – Part 2

Max. award = $80,000
Only 1 proposal will be awarded
Summary: In Round 2, this project was Initiated by the Deep Funding staff, but picked up by an independent team, as an SNET RFP ’avant la lettre’. We believe that a good and unbiased engagement score will be a crucial component for future decentralized governance processes, in Deep Funding, in SNET, and beyond. This, and the intermediate results so far, convince us that this initiative warrants further development. Both the current team and others are invited to submit their proposal for part 2 of the development. 
View RFP1 – Community Engagement Score


Unique knowledge graphs or knowledge bases coupled with clear user stories

Max award per proposal: $40,000
Max amount for all proposals: $100,000
Summary: A proposal request for Knowledge Graphs and related Use Cases that can later be integrated with LLM’s on the Decentralized AI platform.
View RFP2 – Unique knowledge graphs 


Memory-augmented LLMs

Max award per proposal: $150,000
Max amount for all proposals: $250,000
Summary: This Proposal calls for very innovative LLM-based approaches. We challenge you to try training a memory model augmenting an LLM, which will not be trained to reproduce information from training sets directly but will be trained to memorize and recall any new information. To avoid training from scratch, we propose to take a pre-trained SOTA foundation LLM.
View RFP3 – Memory-augmented LLMs 


Tools for Knowledge Graphs and LLMs Integration 

Max award per proposal: $40,000
Max amount for all proposals: $100,000

Summary: We request different tools to work with knowledge graphs which can be useful in the future research and development of Knowledge graph powered LLM. The main requirement is that the proposed tool should be an original work, not a simple wrapper for existing solutions.
View RFP4 – Tools for Knowledge Graphs and LLMs Integration


We hope you are just as excited by this new pool and these specific RFP’s as we are. We are looking forward to all your proposals!
The SingularityNET Deep Funding team AND Platform Team 

Join the Discussion (1)

Sort by

1 Comment

Related News Updates