Claude Pro
Score: 9.6
Claude Pro (Anthropic) is the **LLM subscription we use daily** for writing, research, code review, and long-document analysis.
Most "best AI tools" lists are SEO bait. This is what we actually pay for and use weekly. Tested 6+ months minimum each. For non-devs and power users.
Score: 9.6
Claude Pro (Anthropic) is the **LLM subscription we use daily** for writing, research, code review, and long-document analysis.
Score: 8.7
Notion AI is **AI baked into your notes**: if you already use Notion as primary knowledge base, this is the highest-ROI AI upgrade.
Score: 9.2
Perplexity Pro is the **AI search engine** we use instead of Google for any question that needs current information + multiple sources.
Score: 9
Cursor is the **AI-native code editor** for everyone from professional developers to non-devs who occasionally write scripts or modify code.
| Use case | Pick |
|---|---|
| Daily LLM (writing + reasoning) | Claude Pro |
| Web research with sources | Perplexity Pro |
| Coding (non-dev to ship) | Cursor |
| AI inside Notion workspace | Notion AI |
Power user stack ~$60-90/month: Claude + Perplexity + Cursor + Notion AI / Raycast.
Either works. Claude (Anthropic) is better for long-context reasoning, writing, code review, less hallucination. ChatGPT (OpenAI) has native image generation, GPTs marketplace, more recent web data. Many power users subscribe to both ($20+$20=$40/month) for complementary strengths.
Default: ChatGPT / Claude / Perplexity train models on your data unless you opt-out. Opt-out in privacy settings. For truly sensitive: use Claude Workspaces (enterprise), Anthropic Workbench (no training), or local models (Ollama, LM Studio). Never paste API keys, passwords, SSNs into AI chats.
Not deeply. 80% of value: be specific about what you want, provide context (paste relevant docs), iterate. Don't waste time on prompt engineering courses. Learn by doing.
Cursor (AI-native editor) is strictly better for non-devs and beginners: composer mode writes/edits multi-file changes, codebase context awareness. Pros sometimes prefer Copilot's lightweight integration in VS Code. Try Cursor first.
Claude. Writing voice less formulaic than GPT, lower hallucination on facts, longer context for editing book chapters or full articles. ChatGPT a close second. Avoid Gemini for writing as of 2026 (output quality varies).