Welcome back to the Launch Key ☘

You may not be interested but the speed of change is incredible right now (see today’s Visual Crapshoot real life story).

Over the past 2 months I’m sure I have bored some Launch Key subscribers with my talk of autonomous AI agents. I watched friends eyes glaze over as I droned on at an event that was supposed to be fun.

But I’ve also gotten a fair amount of inquiry from some of you. Have I actually used OpenClaw? Have I deployed other agents? What am I doing personally? And the truth was … not much.

So I set out to change that by building my own small local sandbox - the kind that any of you can do starting today.

Let's get into it.

I’ve had one of my email addresses for decades.

And I wasn’t about to let some AI agent clean up 17,000 unread messages from a half-baked solution I cobbled together. I understand the mess of my digital desktop, and am not sure this old dog wants to learn a different file storage system. Plus, I’m not wrangling enough meetings or projects that a Chief of Staff arranging my daily calendar makes sense.

Still, the twitter feeds on my timeline were intoxicating.

People like Alex Finn preaching how one-man companies running multiple autonomous agents are delivering bespoke products, content creation and social media growth overnight. Or Greg Isenberg giving away free startup ideas and playbooks.

You’ve all read as my AI opinion has changed over the past year.

From Mary Meeker’s summer predictions to The Power of One to How to Win When Everyone has AI. I now embrace this as an advantage - especially for late-career folks.

But many of these ideas just involved building way more tech unknowns than I wanted to deal with. I wanted to experiment with non-critical information.

I needed to eat my own dog food. Where to start?

I trusted Claude more than OpenClaw and decided to build a small local test that I could manage from a headless Mac Mini in our bonus room. It’s an older machine (Intel i5, 64GB RAM), clean OS install, created an Admin and now multiple users (agents), none connected to my email or calendars or personal data. (I do have another agent spun up that will create, edit and leverage my Notion instance - more on that later.)

I decided on my first use case: create a morning briefing waiting in my inbox from the disparate sources I review most days. Nothing crazy, just a content aggregation play. Markets, weather, crypto, local and national news, opinions, technology, fishing, trends, gear, gadgets and interesting commentary.

Like I said - disparate. But I imagine most of you are the same.

I told Claude I wasn’t a programmer, and it designed and then walked me through building the following open source stack on the Mac Mini : Python, SQLite, Ollama, Claude API, Resend (email delivery), RSSHub (Twitter feeds), launchd (scheduling). It literally took a few hours of copy/paste into Terminal and I had a working model. The whole thing costs $20 a month because I pay for Claude Pro (highly recommend).

I quickly realized that all my various sources would not make 1 brief but instead a big mess. So I organized my thinking into 3 topic buckets:

Naturally, I chose the fishing content to build out my first agent.

I wired up NOAA tide data from my nearest intracoastal station. I use the API from OpenWeather to create daily forecasts. Every morning at 5:45am, the briefing agent creates a localized daily chart like the one seen below complete with best fishing windows.

What hyper-local sources would make the Sportsman really useful? It would be ideal to have locations if local fishing guides shared that info (spoiler alert: they don’t 😉 ).

Another shocker: the outdoor sportsmen and women are not huge twitter posters. Salt Strong creates very good YouTube content, North Carolina Wildlife and the South Carolina Dept of Natural Resources folks post less frequently. There’s a local magazine that puts out monthly fishing reports. I added 8 other content feeds. Once the framework was built it has been easy to add or remove feeds as I find them.

After getting the same story links for a few days, I decided to make a weekly (Saturday morning) full briefing and simply deliver daily tides and fishing overview (example below).

Finally, I created and iterated a very detailed ‘voice’ for our local guide who knows 3 counties of public game lands (I put actual links in the prompt), as well as all the inland waterways (Myrtle Beach to Wilmington) like the back of his hand. Taught by his father and grandfather, the old boy has forgotten more fishing spots than we’ll ever know. Part of the prompt was that our guide share knowledge in each briefing. The Saturday Sportsman is especially good. My friends thought I was writing all of this commentary.

I have not written one word.

Is this a great use of AI agents? Absolutely not. Do my local fishing buddies find it useful? 100%. And they send me resources and ideas which will make it even better over time. It’s an easy win that I hope puts more fish in the boat.

I’ve also finished creating my daily news and opinion Dispatch. It includes a 3-day local weather forecast, markets update, North Carolina and national news, independent journalists and Twitter feeds. It’s not perfect but its a pretty good start to the morning and points me to which stories deserve more of my time. Plus, the Futurist technology briefing is days away from making it’s debut.

So yes – I have built a personal AI-powered morning briefing system that collects data from RSS feeds, APIs, and Twitter, synthesizes it through Claude, and delivers 3 (soon to be 4) formatted email briefings on a schedule.

Guess what? I bet that you have much better ideas.

Now go launch something 🚀

A labor shortage in tax is driving the need for a new skill set: one that blends technical tax knowledge with digital fluency.

Automation, AI and data-driven insights now define the role of tax professionals.

This new era of tax is not simply about adopting new tools, it’s about reshaping the skill set and mindset required to thrive in this field. Check out this guide for actionable insights into how to cultivate these skills with your team. See how advanced technologies can help bridge the tax tech gap to increase efficiency, ensure compliance, and drive better decision-making.

Download the guide

Face the facts: before you start building an agent, you should be planning like a Business Analyst. This Emrah Yayici booklet includes tips and best practices to help you start your project on solid footing.

Corey Ganim posted this very good Masterclass on AI Agents for Beginners.

Most people couldn't explain what makes an agent different from ChatGPT if you asked them.

Here's the short version: A chatbot waits for you to ask a question. An agent goes and does the work.

That distinction changes everything about how you use AI in your business.

Here's a full breakdown of AI agents, what they are, when to use them, and how to get started.

💼 Thriving Freelance - Gigs and Growth : Hundreds of job opportunities sent directly to your inbox each week, plus timely resources to support you as you grow and evolve.

📕 MGMT Playbook : Practical management insights straight to your inbox every Wednesday.

📚 StoryBound : StoryBound is the first-ever educational storytelling series designed to bridge the gap in finance literacy, career development, and personal growth concepts that the classroom left out.

🗃 Dealroom Business Success Uncovered : Learn directly from billionaire entrepreneurs on how to grow a business. Join a community of 2,000+ innovators.