FutureMap

A fun app for open source collaboration on maps of problems through crowdsourcing and crowdfunding.

Comment 2

Inspiration

Our inspiration stems from our team’s participation in the 1729.com Network State course, participation in the MapsMap office hours & deep dive session, and shared desire to build tools to aid human-machine information-exchange models.

What it does

FutureMap is a new way to build and discover knowledge. Leveraging the art of cartography, FutureMap creates harmony from the chaos of complicated problems with inclusive crowdfunding tokenomics to ensure great ideas are sustainably funded into fruition. Try out our working prototype, and be sure to check out the "trees" view in the dropdown!

How we built it

Languages

Frameworks

Platforms

Cloud services

Databases

Licenses

APIs

Referential Resources

  • Design Data-Intensive Applications by Martin Kleppmann
  • The Practioner’s Guide to Graph Data by Matthias Broecheler
  • Effective Data Storytelling by Brent Dykes
  • Introduction to Algorithms by Cormen, Leiserson, Rivest, & Stein
  • Transformers for Natural Language Processing by Dennis Rothman
  • The Art of Doing Science and Engineering by Richard W. Hamming
  • The Secret Life of Programs by Jonathan E. Steinhart
  • Yossi Gilad, Rotem Hemo, Silvio Micali, Georgios Vlachos, and Nickolai Zeldovich. 2017. Algorand: Scaling Byzantine Agreements for Cryptocurrencies. In Proceedings of the 26th Symposium on Operating Systems Principles (SOSP '17). Association for Computing Machinery, New York, NY, USA, 51–68. DOI:https://doi.org/10.1145/3132747.3132757

Challenges we ran into

  • Strategizing instead of doing the obvious, thinking from first principles
  • ipld integration

Accomplishments that we're proud of

  • Forming an ad-hoc international team and working asynchronously across time-zones
  • Actively participating in office hours and deep dives with the community for this hackathon
  • Learning as much as we have about graph databases and how they will revolutionize how much value can be extracted from digital data dynamics and human language -Spending hundreds of hours collectively to develop this application and prototype in under 2 months

What we learned

  • Making a mapping application can be approached from many different perspectives and it’s incredibly difficult to define what will ultimately suit users best without developing sophisticated feedback loops to guide continuous integration and development.
  • FutureMap requires the use of advanced reinforcement learning in order to extract value from the natural language sentiments stored into immutable states of a maps ongoing development. This feature is far more difficult to program than to conceptualize. We need to structure mapping constraints to mathematically advance the efficacy of feedback suggestions back to users for streamlining their map development. The goal of leveraging AI is to provide timely suggestions to users when they have reached a “mapping fog” and need additional options suggested to them for continuing development. Just as life is cyclical, knowledge is as well. Knowledge goes through cycles of being chaotic, complex, complicated, and obvious. The state of the knowledge is determined by the weight of the surrounding content and information. Mapping out new information poses a risk of breeding further chaos in a knowledge map. That said, it’s imperative for FutureMap to not only be a platform for maps, but also needs to be an intelligence engine for maps. Nodes and vertices can be effectively represented in a graph database wherein natural language processing algorithms and transformers can extract unforeseen relationships between all the information stored throughout a maps state lifecycles. This will require more work to develop and test. We believe that information within each map and sub-node components must be structured in mathematically optimal ways to maintain reliability of the application at scale.
  • We also learned that physics must govern the dynamics of collaborative map-making: akin to how gravity governs the orbit of planets in a solar system, crowdsourced knowledge that’s piled up under a given topic should govern whether a super-majority is formed to foster its own crowdfunding appeal.
  • We learned why choice of consensus protocol is important for ensuring an application scales to meet the needs of its many users. We chose to pursue development on Algorand after learning about its Pure Proof-of-Stake (PPOS) protocol that leverages cryptographic sortition and verifiable random functions (VRFs) to keep fees negligible, achieve consensus in seconds, defend against quantum computers, and scale without sacrificing decentralization or security.
  • We also learned how to be agile and adapt our development to what was necessary. Due to time constraints and difficulties in troubleshooting the use of IPLD, our team opted to utilize IPFS as a more readily deployable mechanism for censorship & tamper-resistant file storage. In the near future, we hope to integrate IPLD into this application as it will be a more seamless bridge for passing the cryptographic hashes of files directly into a trust-less, immutable blockchain ledger while remaining protocol agnostic. IPLD and IPFS will enable a truly interoperable web3 MapsMap application.

Who will use FutureMap?

We envision a variety of possible structures and relationships to incentivize users to develop nodes, fund bounties, collaborate, and quickly adopt the FutureMap platform. The potential users we see include: University departments and students, university applicants, corporations, independent researchers, and independent internet users incentivised to share their ideas and solve problems.

Individuals will be either campaigners (who are incentivised to spread awareness for a certain plan or map), fundraisers (who seek to invest in projects or new ideas), or people with “skills'' such as such as organizer, manager, engineer, scientist, builder, or artist. These individuals may be capable of either actually building a map, or implementing a plan laid out in a map of a given topic. These different groups will be derived from universities, corporations, and the general public.

Individuals will be incentivized to participate in bounties in order to win tokens. Winners of bounties automatically get their work published as an NFT (which can be a great way for users who cannot pay out-of-pocket to "patent" their work via NFT). When a graph is published as an NFT, if any part of that graph is forked into another graph that wins a bounty, a portion of that bounty prize will be allocated to the original graph creator. Individuals will therefore be incentivized to quickly produce high-quality maps and nodes that other users will want to build off of.

Forecasters will be incentivized to bet on bounty winners in order to win tokens. Other forecasters may be incentivized by forecasters’ reputation, which increases the volume of betting, and subsequently, the amount staked in the liquidity pool which is used to stabilize the token’s value.

Fundraisers will invest in projects and maps through the use of tokens. Tokens may be earned through certain structured tasks (especially to kickoff the FutureMap community), purchased, or won through either winning, or accurately forecasting the winners of, bounties. The number of tokens won through bounties (and there can be different bounty types, i.e. building maps, fact-checking maps, editing nodes, etc.) can be a reputation mechanism to assess the skill of a user.

Collaborators may be incentivized to work together to build maps for bounties because the system can manage the coordination of mappers via smart contracts. The work of collaborators in the system can be easily quantified, and depending on the nature of the contract, bounty winnings can be easily and accurately allocated. Collaborators can evaluate each other through users’ works (publicly available maps and nodes) or reputation scores built over time (related to users’ skills).

Universities may be incentivized to produce and publish maps as a prestige mechanism (similar to scholarly article publishing). Because maps may be published as NFTs, Universities may want to encourage departments, professors, or students to publish works on the map.

University departments and professors may be incentivized to use maps to supplement their educational curriculum, including referencing their own departments’ works.

Incoming University applicants may be incentivized to develop maps because the nature of how students apply to universities has changed (i.e. SATs are now optional at some Universities). So, universities may instead be interested in reviewing students’ works on FutureMap as a part of their application. Universities may even put up bounties for incoming university applicants exclusively, so students have the opportunity to distinguish their application or win scholarships.

Companies may be incentivized to use FutureMap by either funding bounties or, similarly to Universities, establish prestige in a domain by quickly mapping a subject. Companies may use bounties as a scouting mechanism, also.

The Foresight Institute may be incentivized to use FutureMap in order for its fellowship cohort to articulate the challenges, solutions, actors, research requests, funders, and opportunities in their fields, including building tech trees on the subjects of longevity, nanotech, biotech, space governance, intelligence cooperation, and existential hope. They would be able to create and kick off crowdfunding for their fellowship members to actually begin and build projects and organizations based upon these maps: working towards solutions and impacting quantifiable progress in any number of areas. FutureMap may also be used as the application process for future Foresight fellowship programs and/or Foresight fellows may be requested to make a contribution to Foresight tech trees as a part of their fellowship program.

What's next for FutureMap

  • Integrate NFT minting functionality on Algorand TestNet
  • Prototype and test optimal mathematical structures for the node and vertices of each maps data model.
  • Investigate use of fast Fourier transforms with 1,729 dimensions, wherein knowledge is pooled by 3 randomly chosen users/wallets including 9 core map properties across 10 different dimensional states, representing the nodes in the given map and vertices to nodes of other maps, respectively. Possibilities include advanced error-correction suggestions before staking, self-propagating maps based on consensus-driven insights of the global maps.store, simulated maps to investigate the collective value of a given research direction.
  • Meet with Algorand Business Development Lead on March 21st to discuss additional funding opportunities

Built With

Share this project:

Updates