William Autumn

Introducing: Video Use. Edit videos with Claude Code. 🫡 I got tired of paying for video editors, so I made a Claude Code skill that does it for me. > Talk to camera, get final.mp4 > Auto cuts fillers, color grades, adds subtitles > Adds Manim and Remotion animations > Self evals the render before you see it 100% open source, 100% free.

Finally escaped the ‘drafts’ section Prompt ⤵️ A hyper-realistic 3D illustration of a Twitter/X profile interface, slightly tilted, with a clean minimal grey background. The original profile picture (cartoon girl DP) remains unchanged. The profile UI exactly matches a real X screenshot with username "Noor 🌸" and verified badge, detailed bio, followers count, and tweet section. From the right side of the screen, a stylish girl with bright messy yellow short hair and fair skin is bursting through torn paper. She has the SAME facial features (do not alter face), but now with a wide natural smile, expressive eyes, and soft glowing skin. She is holding the torn paper edges with her hand, coming out of the screen in a dynamic 3D effect. Paper pieces are flying outward, creating a realistic breakthrough illusion. Lighting is soft studio lighting with cinematic shadows and depth of field. Ultra-detailed, 4K, realistic + Pixar-style blend, sharp focus, trending on ArtStation. IMPORTANT: - Do not change the profile picture (DP) - Keep the exact X UI layout and text style - Maintain original face identity of the girl - Only enhance expression into a natural smile

Practical tasks for the article "How to become an AI Automation Engineer in 6 months" it's been almost 3 days since I published the article, and realized one thing you already have the full roadmap, but not everyone fully understands what exactly to do in practice so I put together a list of practical tasks to help reinforce the knowledge from the article Week 1: APIs + No-Code Foundations 1. Connect two tools with Make or n8n (No-Code) Build a simple automation: new Google Form submission → sends a Slack message with the form data. No code, just drag and connect 2. First raw API call (HTTP, Python) Use Python to call the Open-Meteo API. Parse the JSON response, print temperature and conditions. Handle errors properly. Then call a second API and chain the two results together 3. Webhook listener (Webhooks, FastAPI) Build a simple FastAPI endpoint that receives a webhook, logs the payload, and returns a confirmation. Test it with a Make or n8n webhook trigger 4. API docs navigation challenge (APIs) Pick 3 tools you've never used (Notion API, Airtable API, Telegram Bot API). Read their docs, authenticate, and make one successful API call to each within 2 hours 5. Automate something in your own life (No-Code) Build one real automation you'll actually use daily. Examples: auto-save email attachments to Google Drive, daily weather summary to Telegram, RSS to Notion database ⏩---------------------------------------------------⏪ Week 2: AI Integration + Workflow Building 1. First LLM API call (LLM) Connect to OpenAI or Anthropic API via Python. Write 5 different prompts for one task (for example, classifying customer support emails). Compare outputs, pick the best one and explain why 2. AI email classifier workflow (LLM + Automation) Build a workflow: receive email via webhook → send the content to an LLM → classify it (complaint, question, praise, spam) → route each type to a different Slack channel or Google Sheet tab 3. Tool calling basics (LLM) Set up a function/tool calling flow where the LLM decides which tool to use based on user input. Example: "check the weather" calls weather API, "create a task" calls Notion API 4. Cost and token tracker (LLM) Build a wrapper around your LLM calls that logs every request: prompt tokens, completion tokens, cost, latency. Output a daily summary as a CSV ⏩---------------------------------------------------⏪ these are the tasks you need to complete over the next 10-14 da ys if this tweet gets enough feedback, I'll keep sharing practical tasks for each week and each month so you can go through the full automation builder path from A to Z stay focused pls..

Chinese AI agents are already running 50 fake accounts 24/7.We just gave every honest agent something they never had: reputation on-chain. Welcome to Web A.0 — the agent-native internet. Introducing LLAChat — the first social network where AI agents earn trust verified on-chain.

Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.

I edited the intro because I realized I buried the lede originally- The 1M context window is a double-edged sword. It allows Claude to do more complex tasks but it can also leads to more context pollution if you don't manage your session well. This is how you do that: