Seroter's Daily Reading — #765 (April 16, 2026)
Seroter's Daily Reading· 
Listen: https://blossom.nostr.xyz/cd7947e683d1cd6f5a0f46828a907ebf2f37501072416d4a7159a30733836258.mpga
Source: Seroter's Original Post
Seroter's Daily Reading, Episode 765, recorded April 16, 2026.
No time to build personal projects today, but it was rewarding to check in on next week's announcements, launches, and the very cool stuff my own team is shipping at Google Cloud Next. Let's get into it.
Our first piece today is a deep profile from First Round Review titled Reluctantly Influential: Inside Lenny Rachitsky's Demandingly Chill Life. Seroter called this a really good piece about someone he admires, someone he describes as a more successful version of himself.
Lenny Rachitsky is the name behind Lenny's Newsletter, which has grown to 1.2 million subscribers, making it the top business newsletter on Substack and top four publications in the entire United States. His podcast has over 500,000 YouTube subscribers with each episode pulling 100 to 200,000 downloads. On top of that, he's running a 40,000 person community, launching two additional podcasts, maintaining product partnerships infrastructure, and even put on a 1,200 person conference called the Lenny and Friends Summit. Twenty-five pieces of content per month, across all channels. And the kicker is that he started all of this while trying to live what he describes as a chill life.
The article does a beautiful job of tracing the contradiction at the center of his story. He set out to build something on his own terms, with a rule of no meetings before 3pm, yet he's built an empire of content and community. The piece goes back to his origins. Born in Ukraine to Jewish parents labeled refuseniks, his family eventually emigrated when he was six, landing in Los Angeles via Italy. His mother, an economist in Ukraine, became a CPA in America despite barely speaking English. His father drove taxis and limos while breaking back into mechanical engineering. The article doesn't frame this as trauma, but rather where some of his drive comes from. Rachitsky himself says he always felt like he could achieve things, always had a chip on his shoulder, a desire to show people what he could do.
He went through nine years at a web monitoring company, spent time at Airbnb for seven years becoming one of their first product managers, and then took a sabbatical that became a kind of existential reckoning. About 45 days into the sabbatical he checked in at work and realized the stuff felt boring and didn't actually matter anymore. That was the Kool-Aid leaving his bloodstream, as he puts it. He quit.
What followed was a period he calls Project Avoid Getting a Real Job. He created his own accountability structure, a board of directors made up of friends who would receive his progress emails every two weeks. He wrote forty free newsletters before ever charging a dollar. When COVID hit, he made the decision to introduce a paid subscription, fifteen dollars a month, which he jokes was more than Netflix. The private Slack community he built was hand-picked, the first thirty members being the most engaged readers. He pinged each one personally to answer questions he knew they had expertise on. He was setting the tone and rhythm of the community himself.
Hamish McKenzie, co-founder of Substack, said Rachitsky has become the poster child for what's possible on that platform. The piece is really about the tension between wanting flexibility and autonomy while building something that demands more and more of you. It's a great read about how sustainable success often looks nothing like the relaxed life you imagined when you started.
Our second piece today is on a more technical note, but no less interesting if you're following the multi-agent AI space. This one is on the Google Cloud blog and covers Multi-Agent A2A with the Agent Development Kit(ADK), Cloud Run, Agent Skills, and Gemini CLI. Seroter flagged this as a good combination to get familiar with, noting that even if you swap out components, they represent smart categories to invest in.
The piece walks through building a multi-agent application using Google's Agent Development Kit, a Python-based framework designed to streamline the creation and orchestration of multi-agent AI systems. The author acknowledges that there are already plenty of Python agent demos out there, but says this one is different because it leverages some advanced tooling in Gemini CLI, Google's command line interface for interacting with source files and providing real-time coding assistance.
The walkthrough starts with environment setup, getting Python 3.13 configured with pyenv for consistent version management, then moving into Gemini CLI installation via npm. Once authenticated, you can see the tooling comes alive, running with Gemini 3 and about 240 megabytes of model data.
The Agent Development Kit itself treats agent development like software engineering, offering modularity, state management, and built-in tools. The piece then shows how Gemini CLI integrates with ADK through a set of agent skills, including cheat sheets for API quick reference, deployment guides, development lifecycle guidance, evaluation methodologies, observability setup, and project scaffolding. These skills are context-aware and activate automatically when you're working in an ADK project.
The strategy outlined is incremental. First, get a basic development environment running with the right system variables and a working Gemini CLI configuration. Second, build, debug, and test the multi-agent system locally. Third, deploy the entire solution to Google Cloud Run, the serverless containerized platform that scales automatically and bills by the hundred-millisecond. The article walks through cloning the sample repository, running setup scripts to configure environment variables like PROJECT_ID, and installing dependencies.
The real insight here is architectural. The A2A protocol, which stands for Agent to Agent communication, allows these multiple agents to coordinate and delegate tasks across distributed systems. Pair that with Cloud Run for scalable hosting and Gemini CLI with its ADK integration for local development and debugging, and you've got a coherent stack for building production multi-agent applications. Even if you prefer different underlying pieces, understanding these categories, the agent framework, the deployment platform, and the tooling integration, is worth the investment.
That's the reading for today, April 16, 2026. Two pieces that couldn't be more different in subject matter, one about the psychology and arc of a creator building an audience at massive scale, the other about the technical stack for building multi-agent AI systems. But there is a thread connecting them. Both are about building systems that scale, whether it's a content ecosystem or a distributed agent network. And both require deliberate structure, accountability, and iteration over time. Sometimes the chill life and the impressive output aren't opposites after all.
That's episode 765. I'll be back tomorrow with more.