Google has released a new CLI for Gemini. The beauty of it is that it allows you to run the AI on a directory on your local machine. It is then able to read your existing files and propose modifications. I can see it as a great helper in case you would like to read and synthesize data in a local directory. https://blog.google/technology/developers/introducing-gemini-cli-open-source-ai-agent/
Fascinating... Thank you for posting this... Question DA', considering Google's interest in gathering every piece of data out there (like they do with Google search, Google Email, on & on) I wonder if this platform would have any "wandering eyes" over ones network like Microsoft does in their OS sometimes... "stuff" you have to explicitly turn OFF in their OS's Cough... Backdoors... Cough ?
Of course, that is the price to pay for free software from these big techs. They want to know what you are doing.
The Google Gemini CLI can interact with local files and the environment . Here is how it works: Interaction with the environment is needed to complete tasks. Access to the file system is included, such as reading and writing files. The Gemini CLI uses built-in tools like file read and file write. Context can be provided using GEMINI.md files that the memory discovery service finds hierarchically, beginning from the current working directory. Files or directories can be referenced in prompts using the "@" symbol. Security and privacy considerations are important: Explicit user confirmation is needed before invoking any tool, including file operations. It operates locally, so code and data are not automatically sent to an external server. The context to share with the model is decided on a prompt-by-prompt basis. Sandboxing techniques such as macOS Seatbelt or containerization with Docker or Podman can be used for enhanced security. The open-source nature of the Gemini CLI allows for code auditing to ensure transparency. Regarding data privacy with Google Gemini in general: Google Gemini collects user data, including conversations, location, feedback, and usage information. This data is used to provide, improve, develop, and personalize Google products and services, as well as for machine-learning technology development. Gemini Apps activity can be controlled through Google Account settings. Gemini Apps Activity can be turned off to stop future conversations from being reviewed or used to improve Google machine-learning technologies. Even when Gemini Apps Activity is off, conversations are saved for up to 72 hours for service provision and feedback processing, but this activity doesn't appear in Gemini Apps Activity. Conversations reviewed by human reviewers are disconnected from your Google Account to protect your privacy and retained for up to three years for product improvement. Google states that it does not sell personal data to anyone. In summary, the Gemini CLI can access local files, but it does so under explicit control with built-in security features and options for enhanced protection
it's like those advanced cloud trading platforms, go ahead develop your trading system on our cloud, we guarantee 100% security ... spoken with soft voice.
You guys have been using Google Docs for a decade, and now you care about your privacy? What do you think Microsoft did when it released Windows 10 for free to everyone?