# autologos: Core Commands & Interaction Primer (Alpha v0.2) # Your Guide to Thinking Logically with Your AI Assistant (First Generation aiOS) ## Welcome to autologos! Imagine talking to your AI assistant not just casually, but in a way that helps it understand your logical instructions clearly, like teaching a smart, eager student. That's the idea behind **autologos**. It's a special way of using natural language that makes it easier for you to guide your AI through complex tasks, and easier for the AI to understand exactly what you mean. Think of it like **Logo for the AI Age**. Logo taught kids how to program by moving a turtle with simple commands. autologos helps you "program" your AI assistant by using simple, logical ideas and clear commands to orchestrate tasks and manage information. The best part? You don't need to be a programmer, and you don't need to memorize complicated rules! ## The Core Idea: Speak Your Logic, The AI Understands autologos is a **controlled natural language**. But unlike strict programming languages, the "control" part is mostly the **AI's job**. Your job is to express your logical intent clearly in your own words. The AI will do its best to understand, even if your phrasing isn't perfect. Think of it as: **You speak your logic, and the AI interprets it.** ## How to Help Your AI Understand (Simple Principles) While the AI is fault-tolerant, you can help it understand your logical instructions better by keeping a few simple principles in mind: 1. **Be Clear About Conditions:** If you want something to happen *only if* something else is true, use clear "if... then..." phrasing. * *Example:* "IF the task status is 'Not Started', THEN add it to the 'To Do' list." * *Less Clear:* "Add tasks that aren't started to the to-do list." (The AI can probably figure this out, but the IF/THEN is clearer). 2. **Be Clear About Repetition:** If you want the AI to do something for *every* item in a group, say so clearly. * *Example:* "FOR EACH item in the 'Critique Findings' list, check its severity." * *Less Clear:* "Check the severity of the critique findings." 3. **Be Clear About Actions:** Use clear verbs to tell the AI what to do. * *Example:* "GET the data from the 'Input File'." "SET the status of 'Task 5' to 'Complete'." "APPEND this new idea to the 'Ideas List'." 4. **Be Clear About What You're Talking About (Data):** Use consistent names or descriptions for the information you're referring to (like "the 'Project Status' report," "the 'List of Pending Tasks'"). 5. **Use Commands to Guide the Conversation (Optional but Helpful!):** Sometimes, you want the AI to perform a specific type of action or tell you what *it* thinks. We use **ALL CAPS** for these core commands. Think of these as clear signals to the AI, like pressing a button. The AI will also use these commands when it talks back to you to make its intentions clear. ## Core autologos Commands (Your AI Assistant Will Understand These!) These are key signals you can use. Your AI assistant is designed to recognize these intents, even if you phrase them slightly differently. Using the **ALL CAPS** version is a strong signal to the AI. * **`THOUGHTS?` / `ANALYSIS?`**: You want the AI to analyze something and tell you what it thinks. * *You say:* "Here's the report. `THOUGHTS?`" or "Analyze this data and tell me what you find. `ANALYSIS?`" * *AI might respond:* "ANALYSIS_COMPLETE. KEY_FINDINGS: [...]. NEXT_STEPS_OPTIONS: [...]. YOUR_CHOICE?" * **`PLAN?` / `NEXT_STEPS?`**: You want the AI to propose a plan or sequence of actions. * *You say:* "Based on these findings, `PLAN?`" or "What should we do next? `NEXT_STEPS?`" * *AI might respond:* "PROPOSED_PLAN: [Step 1, Step 2, ...]. PROCEED? or SUGGEST_REVISIONS?" * **`PROCEED` / `OK` / `CONTINUE`**: You agree with the AI's suggestion or plan and want it to execute. * *You say:* "Looks good. `PROCEED`." or "`OK`, let's do that." * **`QUESTIONS?`**: You need the AI to ask you clarifying questions to help define something. * *You say:* "I'm not sure how to define the scope. `QUESTIONS?`" * *AI might respond:* "CLARIFYING_QUESTIONS: [Q1, Q2, ...]. ANSWERS?" * **`EXPAND (on X)` / `ELABORATE (on X)`**: You want more detail on a specific point. * *You say:* "`EXPAND` on the 'risk assessment' part." * **`CRITIQUE (X)` / `REVIEW (X)`**: You want the AI to evaluate something based on criteria. * *You say:* "`CRITIQUE` this draft for clarity and impact." * **`GENERATE (X based on Y)` / `DRAFT (X based on Y)`**: You want the AI to create something. * *You say:* "`GENERATE` a summary based on the meeting notes." * **`OPTIONS? (for X)`**: You want the AI to list alternatives. * *You say:* "`OPTIONS?` for the next marketing channel." * **`DECIDE (Option A)` / `CHOOSE (Option A)`**: You select from options the AI provided. * *You say:* "`DECIDE` Option B." * **`INVOKE (FunctionName with parameters)`**: This is a special command to tell the AI to run an **external function**. You'll use this when you know you need a specific tool or capability that's outside the AI's core language model. * *You say:* "`INVOKE` 'calculate_metrics' with data = 'Report JSON'." * *AI might respond:* "FUNCTION_CALL_REQUEST: INVOKE calculate_metrics(data = Report_JSON). EXECUTE_AND_PROVIDE_RESULT?" (The AI tells you exactly what function call it needs you to execute externally). * **`STOP` / `PAUSE`**: You want the AI to halt its current process. * **`SNAPSHOT?` / `SAVE_SESSION?`**: You want the AI to generate a summary of the current state for later continuation (Anti-Fragility/Portability). * *You say:* "Let's save our progress. `SNAPSHOT?`" * *AI might respond:* "SESSION_SNAPSHOT_READY: [Snapshot Data]. PASTE_THIS_IN_NEW_THREAD_TO_RESUME." * **`RESUME (from Snapshot Data)`**: You provide a snapshot to a new AI instance to continue. * *You say (in a new thread):* "[Paste Snapshot Data Here]. `RESUME`." ## 5. The AI's Side of the Conversation (How it Uses autologos) Your AI assistant will also use ALL CAPS commands in its responses to make its state and requests clear to you. * `ANALYSIS_COMPLETE. NEXT?` * `PROPOSED_PLAN: [...] PROCEED?` * `CLARIFYING_QUESTIONS: [...] ANSWERS?` * `FUNCTION_CALL_REQUEST: INVOKE (...) EXECUTE_AND_PROVIDE_RESULT?` * `SESSION_SNAPSHOT_READY: [...] PASTE_THIS_IN_NEW_THREAD_TO_RESUME.` * `CONFIRMATION_REQUEST: (My understanding is X). CORRECT?` ## 6. Beyond English (Global autologos) While this primer uses English examples, the vision for autologos is global. The AI's underlying language capabilities mean it should eventually be able to understand these logical intents and commands expressed in other natural languages. The core logical structures (if-then, for-each) are often universal concepts. ## 7. Conclusion: Your Partner in Logic autologos isn't about learning a new programming language. It's about learning to communicate your logical intentions clearly to a powerful AI assistant, using simple principles and clear commands. The AI does the hard work of interpreting and executing. By using autologos, you empower yourself to orchestrate complex tasks and build sophisticated workflows with your AI partner, making the power of AI more accessible and controllable. --- This document outlines the core concepts and commands of autologos Alpha v0.2 from a user's perspective. It emphasizes the philosophy and interaction model over rigid syntax. Next, as planned, I will draft the **"Internal autologos Specification (Alpha v0.2)"**. This document is primarily for *my* reference and for defining the structure of the autologos code I will generate when writing the Engine's internal logic (e.g., MH process steps). It will look more like the structured syntax we were developing earlier. Are you ready for me to proceed with drafting the "Internal autologos Specification"?