tamarandom
commented
2 years ago Breaking news!
YGG and Richard would like to expand the scope of this nomination. For the sake of transparency - and since voting has already started - it is being included as a thread instead of updated in the body.
The update expands further, more concretely, on YGG's nomination for an AI Commons.
Tell us about the community that you are nominating
Algovera is a community of independent AI teams creating useful AI applications. Our mission is to empower data scientists to work for themselves on their own ideas and keep ownership of their creations. Check out a recent podcast on our history and vision.
What public goods does this community support or will they support in the future?
-
Funding for independent AI teams: Algovera is a community of independent AI teams (called Squads). We are especially passionate about AI for Good including healthcare, science and climate. These teams are funded through the Algovera Micro-Grants Program. Proposals are submitted on Discourse. We set up a DAO for each of these teams (currently using DAOhaus, previously Aragon). You can check out our full list of Squads and links to the respective DAOs here. All of our code is open source.
-
Infrastructure for decentralized AI: We want to make it as easy as possible for individuals to get from idea to deployed AI app. To achieve this, we are integrating many different technologies such as decentralized storage, compute, marketplaces and integrating environments such as JupyterLab with MetaMask.
-
Decentralized AI community events: We run study groups, reading groups, discussions, hacking sessions, hackathons and IRL events.
Who are the people, DAOs, and other organizations already part of this community?
-
Algovera AI Squads:
-
-AI Drone Driver: Building a tool to collect synthetic data from virtual 3D worlds to train autonomous drones (e.g. using object detection for vision and natural language processing for decision making). This team has gone on to receive an OceanDAO grant.
-
-Smart NFT Search: Aiming to make anything in the NFT space searchable with text. Check out their proof of concept at https://smartnftsearch.xyz.
-
-Decentralized Healthcare: The team have created a white paper for HopeDAO, which aims to address problems in mental health clinical trials. They have also been creating smart contracts for decentralized clinical trials.
-
-Safe Brain Computer Interfaces: A team of neuroscientists and data scientists creating algorithms that classify EEG measurements while preserving data privacy and ownership. More information about the hackathon that led to the formation of the team here.
-
-Medical Image Classification: A team of independent orthopaedic surgeons and data scientists developing classification models for arthroscopic surgery images. More information about the project here. Check out the tutorial and video of the app that we created on HuggingFace Spaces.
-
-Face Anonymization: Improving fairness of face anonymization algorithms by crowdsourcing data using Web3 technologies. More information about the project here.
-
-ML Model Profit Share: Exploring the use of Arweave Profit Sharing Communities (PSCs) to distribute value generated by AI apps.
-
-DAOkit: Creating a scripting language that integrates crowdsourcing of human intelligence tasks with smart contracts. Check out the original proposal and a demo app.
-
-LoCoFast: Creating a toolkit which would allow one to effectively train fast efficient models for segmentation tasks.
-
Ocean Protocol: We are one of the largest community-based projects in the OceanDAO ecosystem (where we’ve been working for nearly one year, completing 9 grants successfully). We work very closely with many projects in the OceanDAO ecosystem including DataUnion, LYNX, Opscientia and nCight.
-
Kernel: We learned most of our community practices from our time in Kernel Block 4. We hosted many juntos and created a guild. Many of our oldest members joined the community during this time.
-
Protocol Labs: We are partially funded by the PL ecosystem. We have community members from many PL working groups including bacalhau/IPCS, Network Research (sub-team of Network Goods) and shared-zarr.
-
DeSci: Many of our community members are heavily involved in the decentralized science world. We are hosting a workshop on decentralized AI at DeSci Berlin.
-
TEC/TEA/Commons Stack: Some of our community are members of the Trusted Seed and Token Engineering Commons, and alumni of Token Engineering Academy. We have contributed to token engineering software such as TokenSPICE. ygg_anderson is heavily involved in many of these projects.
-
Long Tail Financial: A community that focuses on the long tail of DeFi
Why do you think this community needs a Commons?
AI is the most powerful technology that we collectively have access to. Currently, the power of AI is centralized in the hands of a few big companies and universities. This centralization includes many layers of the AI stack such as data, infrastructure (compute), intellectual property (code, patents, know how) and talent (researchers, engineers). Furthermore, the use of these resources is largely dictated by commercial interests in a top-down manner. Often, this results in the development of technologies that are harmful to society (e.g. surveillance, ad clicks). In our experience, generally well-meaning people can be persuaded to work on these technologies through top-down rigid hierarchies. Also, AI continues to automate many jobs and the issue of supporting an AI-dominated economy is little explored.
We believe that new types of organizations can take us off our current path. Given the freedom and support to decide on their pursuit, we believe that independent teams will generally work on use cases that are better for the world. Apps can be collectively owned and governed with proceeds flowing into a DAO treasury to be further distributed. Token engineering simulations can be performed to come up with new economic models for the redistribution of value generated by AI. You can read more about our vision for AI ownership here. There is a strong open source ethos, online presence and passion for good in the AI community (e.g. HuggingFace, EleutherAI). However, workers tend to contribute to these projects on evening and weekends, while sustaining themselves with salaries from tech companies. Can new mechanisms of funding free more people to work full time on open source and AI for Good?
What other resources do you have that will make your Commons deployment a success?
Our community is our strongest resource. We are well connected across much of the decentralized AI stack:
-
Infrastructure: Critical for providing an alternative to the centralized cloud. We have relationships with infrastructure providers such as Ocean Protocol, Protocol Labs, Gensyn and Filecoin miners (Filswan).
-
Data Unions: We have close relationships with data unions such as DataUnion and Unbanx.
-
Science: Critical for collaborating on large scientific datasets and performing AI research.
-
Token Engineering: Critical for designing future AI economies.
We are influenced by many leading thinkers such as Trent McConaghy (see a video of our discussion with Trent around decentralized AI).
Do you have an idea for the name of this Commons?
The AI Commons, Algovera
Submitted by [Discord handle or Twitter handle]
Discord: “richardblythman | Algovera.ai#3425”, ygg_anderson#4998\ Twitter: @richardblythman, @ygg_anderson
Jose5048
graial
Tell us about the community that you are nominating
The AI Commons
What public goods does this community support or will they support in the future?
AI Public Goods
Who are the people, DAOs, and other organizations already part of this community?
https://oceanprotocol.com/dao https://twitter.com/LongTailFinTech https://forum.tecommons.org/t/mega-working-group/334 https://openai.com/ https://www.deepmind.com/ https://numfocus.org/ https://github.com/notadamking/RLTrader https://www.chrisatmachine.com/ https://www.youtube.com/watch?v=OL_ZnBCJzMs
Why do you think this community needs a Commons?
AI is the most powerful technology that we collectively have access to. We should utilize that technology in service of the commons. We should collectively create and utilize AI tools to serve public goods. An example of how this could be done, is have a collective of deployed AI systems found and scale a DAO. Build in a mechanism such that a portion of this DAO's treasury automatically flows into a commons allocators.
What other resources do you have that will make your Commons deployment a success?
Here is Trent McConaghy discussing the AI commons and how we can get there in 2019 https://www.youtube.com/watch?v=rivHsOtcdOg
Do you have an idea for the name of this Commons?
AI For Good / AI Commons
Submitted by [Discord handle or Twitter handle]
Discord: ygg_anderson#4998
Twitter: @ygg_anderson