jmikedupont2
commented
1 month ago It sounds like you have a complex and innovative project planned for the hackathon! Setting up AI ticketing with self-hosted GitHub runners is a great way to leverage GitHub's capabilities for your own infrastructure on the Huck server⁵. Utilizing the LLVMs (Low-Level Virtual Machines) from Agent Artificial will allow you to create a robust environment for running and managing your tasks.
Creating a Bitensor/substrate service to run QEMU traces as a service is an advanced approach. It aligns well with the Bitensor network's ability to manage interactions with the subtensor chain, which can be particularly useful for blockchain-related tasks¹⁵. This service will enable you to use QEMU traces for validating the Guix bootstrap process. Guix's bootstrapping process is indeed critical, as it defines the trustworthiness of the software from a security and user freedom viewpoint[^10^].
Your idea to create an application-specific subnet targeting a set of packages for AI inference tasks is quite intriguing. By running QEMU traces and analyzing the results with agents within the framework, you're looking at a powerful method to validate jobs against the Guix bootstrap. This could potentially enhance the reliability and security of the packages being deployed.
The ability for swarm agents to request traces as part of function calling adds a dynamic and flexible layer to your system. It seems you're planning to enable these agents to call any function within the network, whether it's a Rust function, Guile function, or any Guix or later Nix package. This level of interoperability and functionality could be very beneficial for a wide range of applications.
Overall, your project appears to be well-thought-out and ambitious, aiming to integrate various technologies into a cohesive system for AI inference and package validation. Best of luck with your hackathon project! If you need any assistance or have more questions as you progress, feel free to reach out.
Source: Conversation with Copilot, 5/31/2024 (1) Managing self-hosted runners - GitHub Docs. https://docs.github.com/en/actions/hosting-your-own-runners/managing-self-hosted-runners?platform=windows. (2) Bittensor API Reference | Bittensor. https://docs.bittensor.com/reference/bittensor-api-ref. (3) Bootstrapping (GNU Guix Reference Manual). https://guix.gnu.org/manual/en/html_node/Bootstrapping.html. (4) Tracing — QEMU documentation. https://www.qemu.org/docs/master/devel/tracing.html. (5) Features/Tracing - QEMU. https://wiki.qemu.org/Features/Tracing. (6) Tracing — QEMU documentation. https://virtio-mem.gitlab.io/qemu/devel/tracing.html. (7) qemu/docs/devel/tracing.rst at master · qemu/qemu · GitHub. https://github.com/qemu/qemu/blob/master/docs/devel/tracing.rst. (8) About self-hosted runners - GitHub Docs. https://docs.github.com/en/actions/hosting-your-own-runners/managing-self-hosted-runners/about-self-hosted-runners. (9) REST API endpoints for self-hosted runners - GitHub Docs. https://docs.github.com/en/rest/actions/self-hosted-runners. (10) configuring-the-self-hosted-runner-application-as-a-service.md - GitHub. https://github.com/github/docs/blob/main/content/actions/hosting-your-own-runners/managing-self-hosted-runners/configuring-the-self-hosted-runner-application-as-a-service.md. (11) Deploy self-hosted runners to EC2 (Spot/On-Demand) - GitHub. https://github.com/marketplace/actions/deploy-self-hosted-runners-to-ec2-spot-on-demand. (12) Full-Source Bootstrap (GNU Guix Reference Manual). https://guix.gnu.org/manual/en/manual/devel/en/html_node/Full_002dSource-Bootstrap.html. (13) Validation | React Bootstrap - GitHub Pages. https://react-bootstrap.github.io/docs/forms/validation/. (14) Validation · Bootstrap v5.3. https://getbootstrap.com/docs/5.3/forms/validation/. (15) Validation · Bootstrap v5.0. https://getbootstrap.com/docs/5.0/forms/validation/. (16) Introduction | Bittensor. https://docs.bittensor.com/learn/introduction. (17) Bittensor | Substrate_. https://substrate.io/ecosystem/projects/bittensor/. (18) Explore Bittensor - Learn, Build and Earn TAO. https://bittensorwiki.com/. (19) undefined. https://qemu-project.gitlab.io/qemu/devel/tracing.html.
Josephrp
as part of this hackathon i will get it running with or similar to ai ticket on self hosted github runners on the huck server using the llms from agent artificial. will create a bitensor / substrate service to running QEMU traces as a service. use those traces to validate the guix bootstrap. the guix bootstrap will be the task, so the jobs created will be validated against the boostrap. so basically an application specific subnet that targets one set of packages as tasks for ai inference. we can run qemu traces and analyse the results of them with agents in the framework. they can request traces as part of the function calling. so the swarm agents can call any functions in the network basically and those are any rust function or guile function, or guix package and later any nix package