Specialised Intelligence Update

[Shivansh779]

.BETTER

Solaris v1.1.0

Detail

.BETTER is the first major update to Solaris, introducing a dedicated system for getting a better response to a specific question through specialised AI models.

The command is especially useful when the response provided by the primary model, the models from the second provider (OpenRouter), or the local Ollama model is not satisfactory and requires a better review, explanation, implementation, or strategic approach.

.BETTER provides specialised modes for Writing, Coding, and Strategy, with two AI models available in each mode. This allows the user to compare responses from two different models specialising in the selected domain.

Except for the Strategist workflow, specialist models are used for one question only, rather than remaining active for the rest of the session. This helps prevent unnecessary model usage and reduces the possibility of hitting provider rate limits.

Because Solaris retains in-session context, the primary models can also use the specialist's response to compare, refine, and improve their own answer.

---

Requirements

This feature requires:

• An active Internet connection
• Nothing else

---

Program Flow

When .BETTER is entered, Solaris presents three options:

1. Writer
2. Coder
3. Strategist

---

Writer

The Writer mode provides two AI models specialised in writing and literary tasks.

The models can either:

• Be provided by default by the developer, or
• Be selected by the user

Their primary purpose is to understand the context and requirements of the user's request and produce an appropriate piece of writing.

Typical use cases include:

• Essays
• Letters
• Emails
• Other forms of written content

The two-model system allows the user to compare different approaches and choose the response they prefer.

---

Coder

The Coder mode provides two AI models specialised in programming logic.

Like Writer, the models can either be:

• Provided by default, or
• Selected by the user

The Coder is not a Coding Agent.

Its purpose is to assist with a specific programming problem by:

• Generating code snippets
• Reviewing code snippets
• Debugging code
• Suggesting improvements
• Reviewing programming logic
• Stress-testing the architecture of a project

It is intended to provide another perspective on a programming problem rather than independently operate the user's entire codebase.

---

Strategist

The Strategist is the most involved .BETTER mode.

It provides two AI models, similarly to Writer and Coder, but uses them through a structured planning process.

Step 1 — Define the Goal

The user first defines their goal.

The goal should preferably be clearly structured and described in a single response.

Step 2 — Planning Questions

The primary planning model receives the goal and asks the user a series of important questions.

A maximum of 10 questions may be asked.

These questions are intended to clarify the requirements, constraints, preferences, and intended outcome before the strategy is created.

Step 3 — Strategy Generation

The user's answers are then provided to the primary Strategist model.

The model uses:

• The original goal
• The user's answers
• The gathered requirements

to produce a detailed:

• PRD, or
• Design Draft

The resulting document provides a structured strategy for achieving the user's goal.

Step 4 — Secondary Model

If the user is not satisfied with the primary model's result, the same task is passed to the secondary Strategist model.

The secondary model receives the primary model's response as additional context.

This allows it to:

• Compare its approach with the primary model's approach
• Identify weaknesses
• Propose alternatives
• Avoid simply repeating the same solution

This creates a practical two-model comparison rather than two completely independent answers.

---

Core Philosophy

.BETTER is designed around a simple idea:

When the normal response isn't good enough, give the user specialised perspectives instead of simply asking the same model to try again.

The two-model approach gives the user control over which response, implementation, writing style, or strategy is ultimately more useful.

At the same time, specialist models are kept scoped to the relevant question instead of being unnecessarily used throughout the entire session.

---

Version

Solaris v1.1.0

First major feature update

.BETTER marks the first major expansion of Solaris beyond its existing primary-response pipeline, introducing specialised AI workflows for Writing, Coding, and Strategic Planning.

---

This discussion was created from the release Specialised Intelligence Update.