Newsletter #66 – Building AI Agents, Financial Independence as a Software Developer, and Rethinking Success

In other news:

• 💰 Recently, I've (re)become very interested in the personal finance / FI movement. Many of the stories are from the US, but I also found this great example from Europe: Michael Houghton, a software developer from Ireland. He is very transparent with his story and numbers on how he built a €700k investment portfolio from scratch. Wounderfull stuff!

• 📝 This article has one of the best ideas I've read recently. Moral of the story: don't compete with AI on marginal improvements — instead, do things it cannot do. Knowledge work is dying — welcome to the age of wisdom work!

• 🎙 I didn't get a chance to do a new podcast episode, but I recommend one of my favorite episodes of the JS Craft podcast with Hrishi Olickel on AI engineering, building with LLMs, AI agents, and more.
  

• 💼 Last month, Pragmatic Engineer released an in-depth analysis of the tech job market, highlighting a gradual but consistent increase in hiring across both Big Tech and startups. The data also pointed to surging demand for AI engineers, a decline in remote opportunities, and the growing importance of location.

• 🎓 Most devs don't understand how LLM tokens work is a really good explanatory video made by Matt Pocock on what a token is and how token size affects AI behavior and required memory. Highly recommended!

• 👨‍💻 Paolo Ricciuti replied to my previous newsletter edition and told me about TMCP - TypeScript MCP, a project he started. It's a lightweight alternative to the MCP JavaScript SDK, which sometimes feels kind of bloated and overcomplicated. By the way, I will have Paolo as a guest on the podcast next week.

• 👾 In the same MCP thread, Nicholas Zakas — the maker of ESLint — built this Crosspost tool that I use very often now to post on social media. It also comes with an MCP server, and you can see its code on GitHub. By the way, I've interviewed Nicholas some weeks ago on AI, JavaScript, and more.

• ♟ This Game Theory Problem Will Change How You See the World. The Prisoner's Dilemma is a classic game theory scenario that asks whether it's better to cooperate or compete — and when each strategy makes sense. In this video, Veritasium dives into fascinating simulations that reveal what truly drives cooperation and trust.

• 🎧 I'm a big fan of the books of Nassim Taleb. Andrei, a reader of this newsletter, just recommended The Incerto Podcast to me. It's quite good, especially the episode with Rory Sutherland.

• 🏛 Success isn't just about hard work! It's also about timing, luck, and the people you choose to share the journey with. I wrote this small rant on why the daily process, values, and companions matter more than outcomes: On Success, Luck, and Values

• 🏡 Speaking about taking home office infrastructure to the next level, I found this older article by Stephen Wolfram on his home office setup. In the same vein, I am increasingly tempted to install a fully local LLM at home and use only that. I'm becoming less and less comfortable sharing too much with OpenAI and others.

• 💪 To say that the weather is wonderful here in Andalusia is an understatement. This opens up many other possibilities. It's the first time in many years that I don't have a gym membership, and I do all my workouts outdoors.

• 🇪🇸 Also, this weekend we made today's trip to Ronda which has now become my favorite non-coastal town. Mind-blowing views. Thank you, Spain!
  

• 📘 When it comes to AI, I believe orchestrating LLMs through code is far more powerful (and fun) than relying solely on prompt engineering. That's one of the reasons I wrote Build AI Agents with LangGraph.js. As Harrison Chase, founder of LangChain, aptly puts it: "It's all about communication!" Judging by LangChain's rapid growth on NPM, it seems poised to become the next big star in the JavaScript world.

And with that, we finish this edition of the JS-Craft newsletter! Be kind, keep coding, and keep learning! See you the next time!