Nine weeks ago, I decided that even though I’ve got a 5-month-old keeping me up at night, a 2-year-old keeping me on my toes, and a PR agency keeping me up in the day… going to start building things with Claude. Come what may. I wouldn’t say I’ve got AI psychosis (no grand, delusional ambitions of winning any Nobel prizes), but I will say, without exaggeration, the following… First, I’ll describe how I’m doing it, and next, I’ll tell you what I’ve been building. HOW The beauty of AI is that it’s a kind of meta software. I remember teaching myself HTML 20 years ago, poring over badly-written manuals, scouring forums, just to understand how to make silly little websites for fun. With AI, you can literally say: “Explain in simple terms how to (whatever it is you’re trying to do.” Not only that, but I went a step further, had it ingest my emails/correspondences, and asked it to do it in MY own language 😎 Software that teaches you how to software. How meta-sick is that?! WHAT 1. Fed Claude everything I've ever written (emails, Slack, articles, podcasts) and built a custom skill that writes in my voice automatically 2. Connected Granola to Claude so that action items from every call go straight to my Gmail tasks 3. Set up real-time Qwoted monitoring so that the second a journalist query matches one of my clients, I get an alert (result: roughly 10 extra media hits per month) 4. During the Apple event, I had Claude check a live transcript every five minutes against my client roster while doing other work. It flagged a Siri announcement as relevant to my voice-AI client and I had pitches out presto 5. Building a pitching game that pulls real news stories and challenges PR pros to match the best subject line to a pitch brief (more on this soon!) AND?! I honestly believe that if you’re not tinkering with AI and building things, then you're kind of screwed. Plus… I can’t remember the last time I was so fired up and enjoying myself this much…