Introducing Pinboard

by Paul Bricman, CEO

  • Tooling
  • DX
  • Access

At Noema Research, we're constantly exploring ways to enhance the development process using generative systems. Today, we're excited to share an internal tool we've been using to streamline our workflows, called Pinboard.

Pinboard is a command-line tool that we developed to improve developer velocity. It's designed to efficiently manage file and terminal references, request in-place updates, and boost productivity in codebase-level development tasks. By sharing this tool, we aim to help other developers and organizations facing similar challenges in AI-assisted development.

Key features of Pinboard include:

  1. Flexible Pinning: Easily grant models access to particular files, folders, or even terminal sessions by adding them to your pinboard.
  2. Contextual Updates: Request in-place updates to your files while maintaining the full context of your project.
  3. Clipboard Integration: Quickly copy the contents of your pinboard to the clipboard for querying language models through other apps.
  4. Interactive Shell: Engage in an extended dialogue about your pinned files, making complex development tasks more intuitive.
  5. Undo Functionality: Safely experiment with suggested changes, knowing you can easily revert to previous versions.

One of the primary advantages of Pinboard is its ability to minimize vendor lock-in. By providing a standardized way to interact with various models, the tool allows developers to switch between different LLM vendors with a single pin llm command, without altering their workflows. Moreover, the project is open source and licensed under the permissive Apache 2.0 license.

This release aligns closely with our broader vision for access control in generative deployments. While our core technology focuses on managing access to particular generative capabilities, Pinboard complements this by providing a structured way to manage a model's access to external resources.

We're committed to improving Pinboard based on user feedback. In the coming months, we plan to introduce several enhancements:

  • Support for more pinnable sources, including online documentation and remote files.
  • Quality-of-life improvements to streamline DX.
  • Integration with additional model providers, including local model serving.

By sharing Pinboard, we hope to contribute to the broader conversation on how to genuinely make use of generative systems in development workflows. To get started, visit our GitHub repository for installation instructions and documentation. We'd love to see what you build with it.

More resources

Deconfusing AI-based IAM & IAM for AI Capabilities

Exploring the distinctions between AI-based Identity and Access Management and IAM for AI capabilities. How do these concepts intersect, and what are their implications?

Read more

Misuse & The European Union

Policymakers around the world are rushing to catch up with the breakneck pace of AI development. In this article, we explore language dealing with misuse and loss of control in the European Union.

Read more

Become a Challenger.

Challengers are individuals who can push frontier models to their absolute limits. They're passionate about the integrity of digital, biological, and social systems, and are stress-testing our simulators across cybersecurity, biosecurity, and beyond — for fun and profit.