Tokens and the Open Knowledge Network
Tokens are being deployed in the OKN program as an experiment. We will probe their utility for sustaining ongoing operations. Lessons learned may be applied towards future initiatives.
What tokens will support
OKN has developed a large multi-subject knowledge graph, but it has also developed a variety of user-facing functionalities such as natural language interaction, visualizations, and much more. These AI-enabled functionalities require GPU servers. These are expensive, and are prohibitively so for supporting public access.
In the first stage, tokens will be utilized to support computational needs of the program, with a specific focus on the above workloads involving GPUs. User queries will be handled using a distributed network, facilitated by tokens.
Token issuance and redemption, getting started
Volunteering participants will install a software package made available on the OKN website. The Fabric team will provide online support via a software-integrated chat/messaging system.
There are two acceptable machine configurations initially:
-
Configuration S(erver)
- Linux or Macintosh
- 16 GB RAM
- 1 Gbps
- 16 cores ideal, less possible with lower query throughput
-
Configuration L(LM)
- Linux
- 1 Gbps
- Nvidia GPU with 4x 24-32GB vRAM
The volunteer runs the software, which will make itself known to the network and respond to queries. The overall system will ensure that queries are gated by token validity checks.