The user wants me to summarize an article with the title "Stop Babysitting Your AI". Here's the plan: 1. **Analyze the request:** * Role: News summarization expert. * Task: Summarize the provided article. * Constraints: * Concise: 2-4 sentences. * Focus on the most important information. * DO NOT copy/paste from the original text. * DO NOT simply restate the title. * Add new information beyond the title. * Answer in English. * **Problem:** The article text is empty. It only provides the title: "Stop Babysitting Your AI" and the source "hackernews". 2. **Deconstruct the title:** * "Stop Babysitting Your AI": This is a call to action. It implies that people are currently spending too much time and effort managing or guiding their AI systems. * "Babysitting": This metaphor suggests a need for constant supervision, intervention, and care, which is inefficient and not the intended use case for advanced AI. * The core idea is about autonomy and efficiency in AI usage. 3. **Brainstorm the core message (without an article):** * The article likely argues against micromanaging AI tools. * It probably promotes the idea of giving AI more autonomy to perform tasks independently. * The goal is to free up human time for more strategic or creative work. * It might suggest that AI should be trusted to handle routine tasks without constant human oversight. * This shift from "babysitting" to "empowering" AI leads to greater productivity and better utilization of both human and machine capabilities. 4. **Formulate the summary (2-4 sentences):** * **Sentence 1 (State the problem
본문
Your AI writes code fast but makes terrible decisions. Think Better injects structured decision frameworks directly into your AI prompts. Website · Documentation · Quick Reference · Contributing Works with Claude · GitHub Copilot · Antigravity You ask your AI "Should we migrate to microservices?" and get a generic pros/cons list. No framework. No bias detection. No structured analysis. Just vibes. Think Better fixes this. It gives your AI access to 10 decision frameworks, 15 decomposition methods, 12 cognitive bias detectors, and 160 knowledge records — turning surface-level responses into structured, rigorous analysis. # macOS / Linux curl -sSL https://raw.githubusercontent.com/HoangTheQuyen/think-better/main/install.sh | bash # Windows (PowerShell) irm https://raw.githubusercontent.com/HoangTheQuyen/think-better/main/install.ps1 | iex Then install skills for your AI: think-better init --ai claude # Claude Code think-better init --ai copilot # GitHub Copilot think-better init --ai antigravity # Antigravity Other install methods: go install github.com/HoangTheQuyen/think-better/cmd/make-decision@latest or clone &make build Just talk to your AI naturally. Think Better auto-activates when it detects a decision or problem: You: "Should we migrate from React to Next.js for our main app?" AI: → Detects: Binary Choice → Framework: Reversibility Filter → Warns: Overconfidence Bias, Status Quo Bias, Sunk Cost → Generates: Weighted comparison matrix + action plan You: "Revenue dropped 20% despite market growth" AI: → Detects: Opportunity Gap → Decomposition: Profitability Tree (MECE) → Analysis: Root Cause (5 Whys) + Fermi Estimation → Warns: Anchoring Bias, Confirmation Bias "choose", "compare", "should I", "pros and cons", "nên chọn cái nào", "phân vân" | 10 Frameworks | Reversibility Filter, Weighted Matrix, Hypothesis-Driven, Pre-Mortem, Pros-Cons-Fixes... | | 12 Bias Warnings | Overconfidence, Anchoring, Sunk Cost, Status Quo, Confirmation... | | Comparison Matrix | --matrix "A vs B vs C" with weighted scoring | | Decision Journal | Track → Review → Improve calibration | "solve", "debug", "root cause", "I'm stuck", "tại sao bị vậy", "không biết làm sao" | 7-Step Method | Define → Decompose → Prioritize → Analyze → Synthesize → Communicate | | 15 Decomposition Frameworks | Issue Tree, MECE, Hypothesis Tree, Profitability Tree, Systems Map... | | 12 Mental Models | First Principles, Inversion, Bayesian Updating, Pareto... | | 10 Communication Patterns | Pyramid Principle, BLUF, SCR, Action Titles... | Control analysis depth with slash commands: | Command | Depth | Records | Best For | |---|---|---|---| /solve.quick · /decide.quick | Quick | 0.5× | Fast scan, simple problems | /solve · /decide | Standard | 1.0× | Default for most situations | /solve.deep · /decide.deep | Deep | 1.7× | Complex, high-stakes decisions | /solve.exec · /decide.exec | Executive | 2.5× | Board reports, stakeholder briefings | Examples: /solve.quick API latency spiked after deploy /decide.deep AWS vs Azure vs GCP for cloud migration /solve.exec Revenue declined 20% quarter over quarter YOU ─── "Revenue dropped 20%" ──────────────────────────────────┐ │ ┌─────────────────────────────────────────────────────────────▼──┐ │ AI ASSISTANT (Claude / Copilot / Antigravity) │ │ │ │ ┌─ Auto-detect ──────────┐ ┌─ Slash Command ────────────┐ │ │ │ SKILL.md triggers │ OR │ /solve.deep → deep mode │ │ │ │ "solve" → problem-pro │ │ /decide.quick → quick mode │ │ │ └────────────┬───────────┘ └────────────┬───────────────┘ │ │ └───────────┬────────────────┘ │ │ ▼ │ │ ┌────────────────────────────────────────────────────────┐ │ │ │ 🐍 BM25 Search Engine │ │ │ │ 160 records × depth multiplier (0.5× → 2.5×) │ │ │ └────────────────────────┬───────────────────────────────┘ │ │ ▼ │ │ ┌────────────────────────────────────────────────────────┐ │ │ │ 📋 Advisor Engine │ │ │ │ Classify → Framework → Bias Detection → Plan │ │ │ └────────────────────────┬───────────────────────────────┘ │ │ ▼ │ │ 📄 Structured Output + Next-step suggestions │ └───────────────────────────────────────────────────────────────┘ Add "save step-by-step" to any prompt to generate a full markdown workspace: solving-plans/project/ decision-plans/project/ ├── 00-OVERVIEW.md ├── 00-OVERVIEW.md ├── 01-PROBLEM-DEFINITION.md ├── 01-DECISION-TYPE.md ├── 02-DECOMPOSITION.md ├── 02-FRAMEWORK.md ├── 03-PRIORITIZATION.md ├── 03-CRITERIA.md ├── 04-ANALYSIS-PLAN.md ├── 04-ANALYSIS.md ├── 05-FINDINGS.md ├── 05-OPTIONS.md ├── 06-SYNTHESIS.md ├── 06-DECISION.md ├── 07-RECOMMENDATION.md ├── BIAS-WARNINGS.md ├── BIAS-WARNINGS.md └── DECISION-LOG.md └── DECISION-LOG.md think-better init # Install skills for your AI think-better list # Show installed skills think-better check # Verify prerequisites (Python 3) think-better uninstall # Remove skills think-better version # Show version think-better/ ├── cmd/make-decision/ # CLI entry point (Go) ├── internal/ # Core logic │ ├── skills/ # Skill registry + embedded data