Table of Contents
Winners’ Spotlight: MoveSpace
The BNB Chain Hackathons serve as an opportunity for innovation, collaboration, and technological advancement. This event not only fosters creativity and problem-solving but is a platform for ambitious projects that are shaping the future of Web3 to create real world impact. BNB Chain provides opportunities and guides projects through the complete journey of launching their project on BNB Chain while supporting projects’ long-term development.
In this blog, we spotlight one of the winning projects in the recent BNB Chain Hackathon in Istanbul – MoveSpace.
MoveSpace is an on-chain data labeling and marketplace platform, aimed at enhancing transparency in AI training and data usage to incentivise human interactions with data with BNB Greenfield.
Tell us about MoveSpace
We are an on-chain data labeling and marketplace powered by a vector database. We are addressing the data transparency issue in AI training and aligning the incentives between data providers and data users. We primarily serve data consumers such as LLM teams, dapp teams, and corporate consumers with processed and curated datasets.
*Vector data is data that is stored as a collection of points, lines, or polygons along with attribute data.
What is the background of the core team at MoveSpace and how did you meet each other?
We currently have 9 people on the team.
Our team consists of PhDs [with expertise] in AI and vector database architecture, [and experience] from Huawei, Tencent, Didi, and Microsoft & Google Cloud APAC.
We currently have an AI and blockchain lab set up together with faculty from University of Alberta and the University of Tokyo. We have been a part of a developer DAO since late 2022, where we met during online hackathons.
How did you come up with your idea for your project?
We have friends in other ecosystems that have talked to the Microsoft Web3 team [about the problems they have identified]. We’ve also talked to a few Web 3 founders and a few VC friends. They all liked the idea of a data marketplace and bringing more transparency to the AI training blackbox.
Our own engineers that have worked in Web 2 companies before confirmed on the feasibility of bringing data labels and vector data on-chain. Our partners like Chainbase and Galxe are also interested in the [opportunity within the] vector and AI data vertical. [As such] it seems like a project we can get support from the industry.
The problem MoveSpace is solving and impact on users
How is your project related to opBNB or BNB Greenfield?
We are a data labeling protocol and a data marketplace utilizing BNB Greenfield’s decentralized storage and its relayer network to BNBChain to power data trading.
What is the problem and how are you solving it?
We are bringing data labeling on-chain to incentivize human interactions with data. Most people are already helping label data in cases like reCaptcha but never get paid for it. We are allowing everyone, anyone to earn money by paying them to label text, image, mapping, and even 3D data.
We are also making data tradable as a regular commodity or an asset and making data ownership meaningful. Data acquisition and processing is a huge market with diverse application settings. Currently it’s hard for users to monetize data without 3rd party platforms. In our on-chain data marketplace, users would be able to sell their datasets directly to projects and developers for diverse use cases like LLM training, dApp development, or business analytics.
What is a memorable story about your project’s journey?
We were looking for partnerships in data providers. We knew there were already a lot of data-related projects that have indexers built for data acquisition of on-chain data from L1 and L2. So we talked to various social and data projects that have their own databases.
It was good for us because projects like Galxe and Chainbase all were happy to allow us to use their APIs to get those data to label on our side and to store [the labels] on-chain. This means that we should be able to make data labels and datasets available from centralized databases or decentralized databases composably for developers.
How is your project impacting your industry and what impact have you made so far?
We have completed a vector data storage MVP on BNB Greenfield and data labeling contract on BNBChain. This MVP allows developers to add vector datasets to the Greenfield storage network and look up data availability. Once datasets are live on Greenfield, users can add labels to the individual data items in each dataset on-chain instantly.
MoveSpace’s Hackathon Journey
What made you want to apply to the BNB Chain hackathon?
We have been looking for a storage solution that has on-chain programmability with it because we wanted to realize the vision of making data ownership meaningful and tradable. BNB Greenfield offers that advantage over many other decentralized storage solutions because of its native relayer network and BNBChain.
Plus we also think showing it at the BNB Chain hackathon is a great opportunity to show our work to the community.
What did you learn during the hackathon and how did you apply it to your project development?
We’ve learned about the relayer network between Greenfield and BNBChain and the bucket structure from the BNB Chain developer during the hackathon. Our dataset was hence stored in buckets on Greenfield and the relayer network is being used to mirror dataset ownership to NFT on BNB Chain.
How did the Hackathon differ from other hackathons you have participated in?
We liked that there’s a specific wish-list of items that the BNB Chain foundation would like to see from the community so we know what the ecosystem needs. Once we saw that in the dApp and infrastructure track under Greenfield, our data labeling dApp and data infrastructure matched with the goals of BNB Chain, it was very easy for us to make a decision to build here.
BNBChain hackathon was a great experience to have to get feedback and support both from the community and from the official BNBChain team, especially if you are building something innovative in the data infra or op stack.
We have been communicating errors and troubleshooting needs with the tech crew from BNB Chain and also have been invited to a Twitter space from the BNB Chain team.
What are your project’s future plans?
We just finished deployment on mainnet and will release a new UI in about a month.
We will release a full-feature MVP of the data labeling Web App in late Q1/early Q2 2024 and then a data marketplace by mid-2024. We are currently fundraising for our first round.