Opanuj 90% Codex'a w 43 Minuty (na żywych przykładach)

📣 Host your websites and apps: https://hostinger.com/mikolajai 📣 Extra 10% off with code: MIKOLAJ10 #płatnaspołucza 📌 Check out the newsletter: https://mikolajabramczuk.substack.com/ 🔗 Try Codex: https://openai.com/pl-PL/codex/ 🔗 Copy the prompt (below): 🟢 I'd like you to explain all of the following Codex functionalities to me: I need a contextual understanding of these functions. So, before you start translating them, ask 2-3 questions that will allow you to best tailor the translations to my needs. Prioritize a proactive approach. If I can see a feature in action right away, have me do it. I'm looking for practical implementation. 🔴 === DESCRIPTION === In this episode, we take a closer look at the OpenAI Codex application. ChatGPT "gives the answers," and Codex "rocks the boat." ;) We build a real-world project from scratch to show all the key features in practice. In the video, you'll learn: • How OpenAI Codex differs from classic ChatGPT and how to configure it. • How to manage permissions (auto-verification vs. full access) and limits. • How to use advanced chat options: Scheduling Mode, side chats, branches, and compacting long conversations. • How to launch multiple sub-agents simultaneously to discuss with each other (e.g., from the perspective of renowned experts). • How to integrate Codex with external tools like Notion or GitHub. • What is the Model Context Protocol (MCP) and how to connect the agent with external applications (e.g., the Paper design app). • What are "Skills" – repeatable skills for your agent – ​​and how to create them. • How local (within a project) and global storage work, and how to program automation (e.g., daily progress checks in Kanban). === TIMESTAMPS === 00:00 - Intro 00:19 - ChatGPT vs Codex: the most important difference 01:22 - Where to download the application and how to get started 01:56 - Interface overview (permissions, models, limits) 03:55 - Workflow: creating a dedicated folder 04:56 - Function 1: Chatting 06:01 - Planning mode 07:01 - Agent proactivity (independent research) 08:14 - Automatic folder structure creation 09:44 - Side chat for asking for details 12:35 - Branching conversations 13:35 - Parallel work of multiple agents 14:55 - Referring to specific files ("@") 17:11 - Persona simulation: sub-agents discussing with each other 20:45 - Compacting (summarizing) long threads 21:57 - Publishing a website online: GitHub and Hostinger 24:53 - Feature 2: Plugins and the MCP protocol 26:50 - Integration with Notion (automatic project Kanban) 29:32 - Custom MCP configuration (using Paper as an example) 34:10 - Feature 3: Skills (repeatable instructions) 36:34 - Feature 4: Built-in web browser 37:21 - Invitation to the Chłopski Rozum AI newsletter 38:38 - Feature 5: Computer Use 38:57 - Feature 6: Memory (global, local, and preferences) 41:27 - Automations (e.g., periodic work reports) 43:24 - Conclusion and summary