This was a full two-day build. The main reason being that I spent 3+ hours on Day 1 trying to fix a Claude Code issue with my Fire Crawl MCP server not working. I used ChatGPT for help and eventually decided to have it write Claude a full summary of the problem, and then Claude Code finally figured it out. Instead of fully copying Nate’s Day 2 build, I used Fire Crawl to scrape information from a real website and gather data on 50+ multifamily properties in Florida with a maximum value of $5 million. Unfortunately, I did not fully complete the csv because I somehow managed to use all 1,000 Fire Crawl credits. Nate only used 6 credits in his video, so I have no idea how I somehow managed to use nearly 200x more 🤣 One thing that surprised me was Claude Code did a great job of self-healing throughout the process. Has anyone here used Fire Crawl for web scraping for a client or in their own business?

Created my first newspaper today. I have always been interested in AI, but never got my feet wet for one reason or another. I finally decided to do something about it. Although it is relatively basic, I am still really proud of creating my first project. I am looking forward to creating more. One problem I kept running into was not knowing the fundamentals of coding projects, like Nate mentions in his videos. Does anyone have any advice of where to learn or if it is really necessary?

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. 😊

Our cold email reply rates have been strong lately, averaging around 2 to 4% consistently which doesn't sound like a lot but when you're sending at scale that translates to a serious volume of qualified conversations every month. The way we got here was by building a full automation using Claude Code, Apify, Million Verifier, and Instantly. The part that made the biggest difference was the personalisation layer. Instead of the usual mail merge tokens that everyone can see through, we built custom Claude skills that actually research each prospect and write something that feels like it was typed specifically for them. That one change alone moved reply rates significantly. Most of the cold email problems we see come down to three things. Deliverability issues that mean emails never even reach the inbox. List quality that looks good on paper but is full of unverified or irrelevant contacts. And personalisation that is either too generic to get a reply or too time consuming to do at scale. Curious where people here are running into the most friction with their cold email. Is it getting emails delivered, building the right lists, or getting people to actually reply?

Today I completed the third challenge of learning and working with skills. I had a blast with this one, and the concepts were definitely reinforced. I referred to the list in the module to build the following: Name - linkedin-post Description - Generates 10 LinkedIn posts on any given topic, written in the user's established voice and style. Each post is scored across five criteria, and a ranked summary table follows all posts. Trigger - "Write me a LinkedIn post on/about [Topic] / [Context] The one change I made was updating the formatting and the instructions to conform with the LinkedIn best practices for posting written content. Even though it was the first on the list in the module and it seemed pretty simple, the ideas just kept flowing on how I could expand things further. I am going to build upon this further by adding a LinkedIn image generator skill to generate a single image and a carousel images for the posts produced so that the post can get more engagement with visual content.