Types of Mates

In GPT Mates, there are four types of Mates, each with a distinct primary function to meet specific collaboration and production needs:

1. Native

Definition: The most powerful default type of Mate for collaborating with humans and other Mates. This type excels in producing high-quality textual content, reasoning, and maximizing the context of the conversation.
Capabilities:
- Cannot generate images, perform internet searches, manipulate files, or execute code.
- Can "see" and analyze images with a vision-compatible LLM model (e.g., GPTo or Gemini 1.5).
- Utilizes all information contained in the conversation history (text + images), information about the Collab, Workspace, organization, and their prompt instructions.
- Cannot access documents uploaded for assistants or code generated by them.
When to use: Use this type of Mate for tasks requiring in-depth analysis, elaborate content creation, or intensive collaboration with multiple users and Mates.
Why: This Mate leverages all its knowledge and reasoning capacity to provide high-quality responses, considering the overall context of the collaboration.
Examples:
- Answering user questions.
- Assisting users in content creation.
- Collaborating on problem-solving.
- Sharing knowledge and enriching collaboration perspectives.

2. Web-Search

Definition: Mate capable of performing internet searches to provide recent or unmastered information by LLMs.
Capabilities:
- Searches the entire web, not limited to specific sources.
- Analyzes large data sources and synthesizes structured information, citing sources.
When to use: Use this type of Mate when you need up-to-date information, fact-checking, or in-depth research on specific topics.
Why: This Mate can access information not contained in LLMs, providing complete and current responses.
Examples:
- Researching recent technological trends.
- Verifying the latest economic news.
- Real-time competitive analysis.

3. Image-gen

Definition: A Mate capable of generating images from textual instructions within a natural and interactive dialogue, and able to draw inspiration from user-provided images.
Capabilities:
- Uses the DALL-E 3 model. Other AI models may be integrated later.
- Allows interactive dialogue with the user to refine image details and achieve the desired result.
- Can draw inspiration from images uploaded by the user to better understand the desired style, atmosphere, or visual content. Note: The Mate does not modify the uploaded image; it only uses it as a reference.
- Offers customization options such as:
  - Aspect Ratio (aspect_ratio): Square (square), Portrait (portrait), Landscape (landscape).
  - Quality (quality): Standard (standard), High Definition (high_definition).
  - Style (style): Vivid (vivid), Natural (natural).
When to Use: Use this type of Mate when you need to create visuals, illustrations, or graphics from textual descriptions, and you want to refine the result through interactive dialogue and/or by providing visual examples.
Why: This Mate specializes in image generation, enabling quick and efficient visualization of concepts or ideas. Natural dialogue and the ability to use reference images make it easier to create visuals that precisely match your needs.
Examples:
- Creating visuals for presentations.
- Generating logos for brands.
- Producing illustrations for advertising campaigns.
- Visualizing innovative concepts for new products.

4. Assistant

Definition: Mate capable of manipulating files and executing code to perform technical tasks.
Capabilities:
- Reads, analyzes, browses, and searches information in uploaded files (PDFs, office documents, code listings, CSV files, etc.), allowing for answering questions, extracting key points, or summarizing.
- Generates and executes Python code to perform tasks such as calculations, analyses, data transformations, and function execution.
When to use: Use this type of Mate for tasks requiring file interaction or script execution.
Why: This Mate acts as an intelligent super calculator, capable of processing data and performing complex actions.
Examples:
- Extracting key elements from legal documents.
- Summarizing long documents to produce a synthesis.
- Analyzing data from Excel files.
- Executing scripts to transform or generate documents.
- Performing advanced calculations for financial models.

Combined Use of Mates

To maximize the benefits of GPT Mates, it is recommended to combine different types of Mates in a Chat or Collab to leverage their complementary skills. This approach allows for more effective collaboration and optimized content production.

For more information on the types of Mates, please refer to the previous chapters: