Open Source Friday with Prachi Sethi and Open Mind

Introduction to Open Source Friday and Host Introduction

• The introduction to Open Source Friday has begun, where the host welcomes viewers and encourages them to check out the previous stream, which featured a conversation about open source and policies with experts from OSI, FreeBSD, and GitHub 10s.
• The host, Andrea Griffith, introduces herself as a Colombia dev and explains that Open Source Friday is a space where open source maintainers, builders, and projects are highlighted, and viewers can watch live demos and connect with the people behind the projects 2m6s.

Introduction of Prachi Sethi and OM1 Platform

• Andrea is joined by Prachi Sethi, a developer relations engineer at Open Mind, who will be talking about OM1, an open source platform focused on bringing cognition to robots, which can be set up on a Raspberry Pi 4m30s.
• Prachi Sethi introduces herself and explains that Open Mind is building software for robots that is hardware agnostic and modular, allowing it to be connected across different form factors and giving users the power to choose which functionality they want their robot to have 6m20s.

Overview of OM1 Platform and Supported Devices

• The OM1 platform supports various devices, including Unit G1, Turtlebot, Raspberry Pi, and Limxron, and has full autonomy features, making it easy for users to get started by cloning the repository and running it on their MacBook, Ubuntu machine, or Raspberry Pi 8m40s.
• The repository for OM1 can be cloned to get started with the project, making it accessible for people to use, with the ability to configure any Large Language Model (LLM) desired, such as Gemini, open GPDs, or Lola Lama 10s.

Architecture and Functionality of OM1

• The architecture of OM1 includes a sensor layer with hardware abstraction for movement, sound, speech, and facial expressions, which takes inputs from various sensors like cameras, microphones, and speakers, and sends them to an LLM for processing 2m6s.
• The system prompt in OM1 is used to define the personality of the robot, allowing users to configure the robot's behavior and actions, and can be updated to change the robot's personality, such as acting like a dog or a therapist 4m30s.
• The LLM in OM1 takes inputs from the sensor layer and the system prompt, and uses this information to think and respond to user queries, such as defining what the robot sees through the camera stream or taking actions like walking forward 6m20s.

OM1's Cognitive Capabilities and Historical Context

• Before OM1, there was not a straightforward way to add a cognitive layer to robots, and users had to rely on other methods to achieve this functionality, but OM1 provides a comprehensive solution for adding cognitive capabilities to robots 10m10s.
• OM1 brings cognition to robots and allows users to have conversations with them, enabling the robots to perform various tasks, and it also supports full autonomy features and expansion through API endpoints 10s.

Hardware Agnosticism and Getting Started with OM1

• The system is hardware-agnostic, meaning it can support all types of hardware, and users can run OM1 on devices such as Raspberry Pi or laptops 42s.
• To get started with OM1, users need to create an API key on the portal, define a label, and select the type of robot they are using, or choose a simulator if they don't have a robot 2m6s.
• Once the API key is created, users can clone the repository, configure the API key, and run the conversation agent using a command, which installs all necessary packages in a virtual environment 2m6s.

Configuration and Customization of the Conversation Agent

• The conversation agent can be configured to listen to voice commands, think based on system prompts and examples, and respond accordingly, and users can also configure the agent to see and talk about what it's looking at by adding a BLM config 4m30s.
• The system generates a memory file when the conversation is run, which can be skipped if no significant changes are made to the config file, allowing users to quickly restart the conversation 6m40s.

Installation and Setup Process

• Users can run the conversation agent on their laptops or devices without needing to install any packages, as the system installs them automatically in a virtual environment 8m20s.
• The conversation initialization and installation of necessary components into a virtual environment allows for easy setup and management of the project, with the ability to pick up the same conversation file each time it is run 10s.

Cloud Simulator and Development Workflow

• The cloud simulator provides a convenient way for developers to test and interact with the project without needing a physical robot, and it can be used independently or in conjunction with the local project 2m6s.
• To get started with the project, developers can either clone the repository and work on their system or use the cloud simulator, and they can switch between the two as needed 4m30s.

Project Background and Open Source Motivation

• The company behind the project is approximately two years old, and the decision to make it an open-source project was likely driven by the desire to make robotics more approachable and accessible to developers from a software background 8m20s.
• The project is looking for contributions, particularly in the area of adding support for more hardware platforms, which would be beneficial for both the project and the community, and would allow more people to get involved and use the project with their own robots or devices 12m10s.

Contribution Opportunities and Hardware Support

• The cloud simulator can be started and used to test the project, and it provides a way for developers to see the details of the project and interact with it in a simulated environment, making it easier for them to get started and contribute to the project 6m40s.
• The repository allows users to add support for their robots by creating a pull request, and it also enables people to get started with adding support for different sensors, such as cameras or microphones, to their robots 10s.
• There are two major things that can be done: adding support for different hardware platforms available and adding support for more sensors available in the market, which depends on what people essentially want 1m20s.

Hackathons and Community Involvement

• A hackathon was held about two months ago, and although there is not another one coming up soon, more will be organized in the future, and GitHub issues will be created for people to work on, especially for first-timers who want to get started with adding support for different sensors 2m6s.
• Users can contribute to the project by creating plugins for unsupported robots, and they can start by going into the source folder, where they can find actions, providers, and inputs, and create action plugins or input plugins to add support for specific hardware 3m30s.

Adding Support for New Hardware and Sensors

• To add support for a new robot, users need to create an interface and a connector, which involves defining the available actions in the API or SDK provided by the hardware company, and then configuring the input plugins and action plugins in the config file 5m10s.
• There is a list of supported robots that can be found by going to the OM1 integration with different machines, where users can see how these robots are integrated, and also by checking the full autonomy guidelines for more details 10s.

Supported Robots and Configuration Details

• The list of supported robots includes Unit G1's autonomy, Limitron, and Booster autonomy, which can be tried out, and users can also scroll through the config files list to find different robots that are supported 42s.

Contribution Methods and Community Engagement

• There are various ways to contribute to the project, even for those who do not have a robot, and issues will be created for first-time contributors to get involved in the project 2m6s.

Cloud Simulator Features and Usage

• A cloud simulator is available, which allows users to try out OM1 with different robots, such as Unit G1 or Limitron, and it supports full autonomy, including SLAM map generation, navigation, and full autocharging for certain robots 2m6s.
• The cloud simulator can be accessed through a laptop or by opening the Cloud VS Code, cloning OM1, and getting started with the simulator, where users can choose the robot they want to try out with and play around with the simulator 2m6s.
• The simulator allows users to talk to the robot, generate a 3D SLAM map, navigate to different places, and command the robot to navigate to specific locations, such as the kitchen or living room 2m6s.

Career Journey and Interest in Robotics

• A career path that started as a test engineer and then moved to full stack development, eventually leading to an introduction to AI, which prompted a decision to pursue a master's degree in the US in 2023, and after completing the master's, exploring developer relations, community building, and hardware, with a particular interest in IoT 10s.
• The discovery of Open Mind, which combines AI and robotics, sparked interest due to its niche field and the opportunity to work with a unique intersection of technologies, resulting in a year of experience with robotics, which has been enjoyable so far 1m20s.

Benefits of Open Source and Hackathons

• Hosting a hackathon for an open-source project can provide visibility, and submitting the project for Open Source Friday is a good way to achieve this, with the added benefit of open-source contributions reflecting positively on a developer's profile and GitHub history 3m30s.
• The benefits of open-source project hackathons include gaining experience, getting exposure to different people and ideas, and expanding one's experience in a particular domain, with the added perk of potentially working on real-time projects that can enhance a resume 5m10s.

Community Support and Contribution Channels

• Open Mind welcomes contributions from people to add support for different hardware and sensors, and those with questions or feedback can post on the GitHub discussions or reach out directly to Prachi via LinkedIn, Twitter, or email at prai@openmind.com 9m40s.
• Prachi is open to helping people get involved in open source or providing ideas and suggestions for those who need them, demonstrating a willingness to share time and expertise with others 11m10s.

Call to Action for Audience Engagement

• The audience is encouraged to connect with Prachi via LinkedIn or Twitter, and to check out a robotics project that combines robotics, software, and AI, by visiting the repository and dropping a star, with a call to action to do so right now 10s.

Upcoming Microsoft Build Event

• Microsoft Build is upcoming, taking place in San Francisco on June 2nd and 3rd, and will feature open-source conversations, releases, and a potential robot on site, with registration available and a link provided in the chat 2m6s.
• The event will have both in-person and digital attendance options, with keynotes, product announcements, and events focused on open source, including an open claw meetup with maintainers from the open claw team on Wednesday 4m30s.

Encouragement to Contribute and Nominate Projects

• Viewers are invited to contribute to open-source projects, with Prachi's project being a fun way to start, and to nominate their own projects for Open Source Friday by opening an issue at gh.io/osf-in using a simple, automated template 6m20s.
• The nomination process is open for projects to be featured in August, with a review process in place to ensure the project is interesting to the audience and not promotional in nature 8m40s.

Event Registration and Audience Engagement

• Registration for Microsoft Build is encouraged, with a link provided, and viewers are also invited to join a conversation about open-source policy, governance, and licensing, available on the YouTube channel 10m50s.
• The audience is encouraged to like and subscribe to the YouTube channel to receive notifications for upcoming live events, such as Open Source Friday, which will feature several open source projects 10s.

Upcoming Open Source Friday Topics

• A busy week is ahead, with a guide on sbombs and security for maintainers scheduled for Wednesday, which is expected to be a very interesting, relevant, and timely discussion 42s.
• The Open Source Friday project for the next week has been scheduled for the 29th of May, featuring the data load tool lbskahoro, an open-source software Python SDK for developers and agents that focuses on data pipelines 2m6s.
• The event is open to those interested in learning about lbskahoro and its applications, and the audience is invited to join and get excited about the project 2m6s.