Today a friend who’s also a former entrepreneur showed me something cool. He’s a technologist and loves to tinker. We talked about AI agents and how much running them continuously via Claude will cost, and he showed me his setup. He bought inexpensive (maybe $100) Raspberry Pi computers and set them up in his house on his local network. He installed the Claude app on them. He has Claude Cowork build software that solves a specific problem for him and then install the software to run locally on a Raspberry Pi.

My friend then uses Raspberry Pi Connect to monitor his agents and access the software that’s running locally on each Raspberry Pi machine. He likes this setup because it doesn’t require continuous work and token usage with Claude. Instead, Claude builds the software and the agent’s job is done. The software runs to do whatever task my friend wants. Because it’s running on a Raspberry Pi at his house, it doesn’t cost him anything to keep the software running 24/7. And he can access the software from anywhere, as long as he has internet access.

I thought this was an interesting setup to get around the increasing costs associated with using Claude and its agents.

The speed at which AI is moving and new products are being released is wild. People are still getting used to Claude Cowork, and yesterday Claude released its Managed Agents platform (see here). Like everyone, I find it hard to stay on top of it all. Actually getting around to playing with all these tools is even harder.

A friend told me about a group of AI engineers who work at companies like Meta during the day and do evening sessions on Wednesdays where they share the latest and explain it to the participants. On top of that, they do hands-on sessions where everyone practices setting up their own versions of the new tool(s) and building with them. The AI engineers help everyone, but participants also help each other.

I’ve decided I’m going to attend these sessions. Other people who want to stay on top of AI and learn the latest sound like a good group to be a part of. Plus, the idea of learning by doing suits me well, especially when I can do it with a group of like-minded people.

AI is real and will have a profound impact. The speed at which it’s moving is forcing me to take a different approach to learning about it.

Over the last few weeks, I’ve chatted independently with two entrepreneurs who sell physical products online. Both of them mentioned that TikTok is a core piece of their growth strategy. I wasn’t expecting that, so I asked a few questions. I knew that TikTok had launched TikTok Shop, where users could purchase products directly on TikTok. But what I didn’t know was how popular that service is. And I was unaware that TikTok has also launched Fulfilled by TikTok, a service that allows merchants to send items to one of 14 TikTok warehouses so that when an order is received, TikTok handles processing and shipping to the customer. It sounds just like Amazon’s Fulfilled by Amazon (FBA) service.

I’m curious to learn more about this. The idea of combining social media with commerce and fulfillment in a single platform intrigues me. Capturing the attention of a user and letting them complete a purchase in a few clicks without leaving the platform is a combination with huge potential.

Last week I listened to an interview with Annie Duke. She’s a former professional poker player who wrote two books on decision-making that I enjoyed (see here and here). During the interview, she explained why moving from implicit intuition to explicit decision-making, including expected-value calculations, leads to better outcomes in high-stakes situations. The gist of it was that instead of making decisions in our heads, we should write down our thinking and share the written document. The exercise of writing down your decision logic will help you spot errors (and avoid or fix them) and make it easier for others to spot flaws in your thinking.

I believe this to be true, but I don’t do it consistently enough with high-stakes decisions. I wanted to test it. Last week, I wrote, with the help of AI, an investment thesis for a publicly traded company. I then shared that thesis with two friends and asked them to discuss it with me after they’d reviewed it. Here are a few observations from the exercise:

• Writing down my logic and reading back through it helped me identify two points that contradicted a timeline that I’d proposed. When I did more digging to get more context around each point, I realized that my thesis was weaker than I’d thought, partly because of an update to my timeline.
• AI helped me create parts of this report, so going back and reading through it helped me catch errors in AI’s logic, too. AI made some irrational assumptions. When I pointed them out, it agreed and fixed them.
• Giving my decision logic to someone in writing, having them read it, and then discussing it afterward was a game changer. When we talked, they were totally up to speed on my thinking and were able to go straight to the areas where they thought my argument was strongest and weakest. The conversation was much deeper than if I’d tried to explain it to them orally.

This experiment showed me that Duke is spot on. When I’m making high-stakes decisions, it’s worth the effort to be explicit in my decision-making process by writing it down, reviewing it, and sharing it with others. The act of writing forces me to face holes in my thinking and allows others to easily see holes and point them out. All of this should reduce the chances of my making a bad decision.

If you want to listen to the section of the interview where Duke talks about explicit decision-making and calculating expected value, take a listen here.

I’ve been playing with the Claude for Mac app and using Cowork, its AI assistant. Overall, I like Cowork, but I can use it on my computer only when the Claude for Mac app is open. The issue with this is that it’s another app, one that isn’t in my normal workflow. Since I don’t naturally use it, I have to put extra effort into remembering to go to it.

The apps I use most are for email and text messages. So, I think it would be best to interact with Claude Cowork, or an equivalent product, via one of these; ideally, text messages. I need to do some digging on this, but I think that’s what I’m looking for: an autonomous AI task executor that I can interact with via text messages.

Continuing what I started last week, I’m going to share content I consumed and learned from.

What I struggled with:

• No material struggles this week

What I learned:

• The science behind naming billion-dollar products – YouTube video from My First Million where they interview the guy big brands pay to name their products. The process of how he picks names is intriguing. There’s much more science to it than I realized.
• Should your portfolio be positioned for defense or offense? – YouTube video where Howard Marks shares his thinking about risk versus reward in today’s investing environment.
• Declining liquidity is negatively impacting stocks – YouTube video from Monetary Matters with an interview of global liquidity expert Michael Howell. Interesting to understand how a booming economy reduces liquidity in financial markets, which can lead to flat or lower markets. Very counterintuitive.
• Emerging trend: internal AI platform team – YouTube interview from The Biography Podcast with the CEO of Speak, Connor Zwick. He has an internal team dedicated to helping the entire company use AI. They go around the company and set up AI infrastructure for internal teams so they can make the most of AI, which helps them move faster. Interesting concept that makes a lot of sense.
• How Blockworks bootstrapped to $150 million – YouTube interview from The Biography Podcast with the CEO of Blockworks, Jason Yanowitz. Interesting strategy of going from events to media to a data company. Focusing on an overlooked niche helped them separate from the competition. Landing a busy CEO for an interview by creating a book about his company was a nice angle. Flattery works. I think I found a digital version of the book here.
• Building a $100 million AI company – YouTube interview from The Biography Podcast with Dylan Fox, founder of Assembly AI. One point stood out to me: Having clarity on how the customer hears about new products and how they buy new products is important for building successful marketing efforts. You should design processes like onboarding and demand generation by working backward from that understanding. Taking what worked at another company or space and replicating it isn’t a strategy that’s guaranteed to succeed at your company.
• $1.6 million business anyone can start – YouTube interview from Chris Koerner on The Koerner Office Podcast with Sariah Howell, founder of Memory Magnets. Howell went from broke to making $600,000 in profit in a year with a simple business idea. I loved how she turned sharing her secrets about building her business into an additional revenue stream by selling products to help others start businesses just like hers. Here, the use of social media is savvy and stood out to me.
• Zara founder getting $3.7 billion dividend – See my thoughts in my post earlier this week. Building to hold forever can lead to great wealth (he’s worth $126 billion). Building to sell will get you cash quicker but won’t create this level of generational wealth.

That’s what I consumed and learned from and struggled with last week.

This past week, I watched a video in which the creator of Claude Code, Boris Cherny, explains what Claude Cowork is and walks through a hands-on demo of how to use it. Cowork is basically agentic AI that can do things on your behalf through your computer. The demo was helpful, and I decided to download Claude Cowork. It was easy to install.

Before I started playing with it, I wrote down one problem that I wanted Cowork to help me with (I’m trying to be mindful of Kindlin’s law). I did this because using a tool to help me solve a specific problem I’m excited about helps me better understand the tool (and its limitations). Otherwise, I poke around and get only a surface-level understanding of its capabilities, which I likely won’t retain.

I then started to work with Cowork to get connected via MCP to the data sources it needed to help me solve the problem. Connecting to Google Sheets API and Google Drive API took some troubleshooting, but Claude helped me figure out the issues quickly. Then I started putting the tool to work in Google Sheets.

I haven’t leveraged all of Cowork’s capabilities, but so far I’m impressed. Playing with it got my wheels turning about what’s possible. I feel like Cowork is something that nontechnical people can use to gain leverage and become more resourceful. I think it’s especially useful for entrepreneurs who are actively building because it’s like a quasi-employee you don’t have to pay and who never gets tired.

This tool is new, but I think becoming familiar with it is worth the effort because of the potential gains. Cowork feels like a potential force multiplier of my efforts that will increase my output and improve my outcomes.

‍

Today I read a blog post (see here) that had caught my eye. It was supposed to be about how to create a personal customer relationship management (CRM) system. Like a lot of people, I struggle to maintain relationships, so I gave it a read. The article turned out to be more than I expected. It’s a great read for anyone trying to understand how to use AI to help with manual tasks and how AI can connect with various tools you already use.

It presents a high-level overview of how you can use AI to talk to the tools you commonly use (Slack, Calendar, Gmail, Google Sheets, Notion, etc.) using the model context protocol (MCP). It describes how MCP works, who created it (Anthorpic), and how the average person can easily download and use it. MCP is powerful because most manual tasks involve shifting between systems. MCP connects all those systems to an AI assistant like Claude, empowering the AI to see the same data you do and execute on it just as you would. You can have an AI assistant complete tasks that would have taken you tons of time or energy, or you can have it do 90% of the work so you have to do only 10%.

The author then takes it a step further and shows you, pretty much step by step, how to use MCP to connect to Gmail and Calendar. It then provides example prompts for instructing AI to do tasks that you’d normally have to manage manually in a CRM. For instance, the author has AI prepare a list of names, using specified criteria, of people he needs to follow up with. He also has AI analyze his calendar for the past week and summarize all his meetings so he can report to his cofounder on Monday what he’d done that week.

I found this post helpful and will be sharing it. Not everyone needs a personal CRM, but the post does a great job of showing how anyone can use MCP and AI assistants to become more efficient and also explaining how to set it up.

I’m in the middle of synthesizing The Little Book of Market Wizards. I like this book because it details how some of the top market participants (traders and investors) avoid psychological errors that lead to bad decisions. It’s essentially wisdom on how to execute an investing strategy without making errors.

One of the people quoted in the book is Ray Dalio, founder of hedge fund Bridgewater. What he said stuck with me:

[T]here is an incredible beauty in mistakes because embedded in each mistake is a puzzle and a gem that I could get if I solved it (i.e., a principle that I could use to reduce my mistakes in the future).

I really like this framing. Mistakes are a puzzle to be solved, and solving them yields a gem. I think that’s a great way to think about future and past mistakes. Each one is an opportunity to figure something out and find a gem that will help you going forward.

So, I’m adding one more thing to my Christmas and New Year holiday challenge. I haven’t been consistent with the Getting Things Done (GTD) method. I read David Allen’s Getting Things Done in 2024, and if I’m being honest, I loosely implemented it. I want to be more productive in 2026, and I think the GTD method could play a big role.

So, my goal is to read the GTD book again, create a clean, short-term (completed within a year) project list, and define next actions for each project.

That’s it. Wish me luck.