Curated a few workflows myself but want to see what everyone else is spinning up. Drop what you're building below, let's swap some ideas.⬇️

New model today, and it's not a small bump. Fable 5 is the strongest model Anthropic has released to the public so far. It beats Opus 4.8 on basically every benchmark, and the gap gets bigger the longer and more complex the task is. The numbers that stood out to me: - Stripe ran a codebase-wide migration on a 50 million line Ruby codebase in a day. Same job by hand would've taken a team two months. - It's the new state-of-the-art for vision. It can rebuild a web app's source code from screenshots alone. - Way better with long context and memory, so it stays sharp across millions of tokens on long-running tasks. Pricing is $10 per million input / $50 per million output, less than half what Mythos Preview cost. Here's the thing for us subscription folks: it's free on Pro, Max, and Team through June 22. After that it moves to usage credits until they get enough capacity to bring it back as standard. So if you want to test it on real work, the next two weeks are the window. One catch worth knowing. Fable has new safeguards, and on cyber, bio/chem, or distillation queries it quietly falls back to Opus 4.8 instead. They say it hits under 5% of sessions, so most of you won't notice. I'm going to run it through Claude Code today and see how it holds up on a live build. What are you testing it on first? Read more here: https://www.anthropic.com/news/claude-fable-5-mythos-5

I've been experimenting with Claude for automation lately and wanted to share what actually works without the fluff. Most people completely overcomplicate this. The only stack you need to start: - Claude: The brain that processes your data. - Claude Code: The hands that execute the actions in your terminal. If you want a solid first project, connect a new form submission to Claude, have it draft a personalized response based on the user's answers, and auto-send it as an email. It takes 30 minutes to set up but saves hours every week. The real progression timeline: - Month 1: Focus deeply on prompting and nail one simple workflow. - Months 2–3: Start chaining prompts together and connecting APIs. - Month 4+: Move into full agents, multi-step logic, and actual client work. 📢 Can you spot the traps? I left a couple of massive traps in that timeline that beginners drop into blindly. Let’s see who can catch them in the comments: 1. The Context Paradox: Why do beginners always give Claude less background context when trying to automate, and why does it completely ruin the output? 2. The Manual Mistake: What is the absolute danger of trying to automate a workflow that you haven’t explicitly done manually first? Drop your thoughts below, or let me know if you're stuck building something specific right now!

Let's get to know each other! Comment below sharing where you are in the world, a career goal you have, and something you like to do for fun. 😊

Got my biggest payout for one project so far. This is basically a chatbot built in n8n and connected to Slack, which can control the project management system 'Asana' of my client. With this the client is able to chat with the bot inside a Slack channel and the bot can create/edit/delete tasks/assignes/dependencies etc. Basically almost everything he can do in Asana, he can do now via his Slack. This system is important for him, because my client manages different company leaders, who all have their own slack channels in his workspace and he wants to have a centralised way for him and his cliens to manage the Asana Boards. His clients can use this bot as well, only for their own Asana workspace. My client can use this bot for the whole Asana workspace. I got paid 650€ for this + 60€ for launching 3 Email campaigns for him.