TWA-001July 16, 2026

Hello (again), World!

Hello again,

I had an unexpected staycation that had many downsides, but the one upside was a gift of space and slowness.

My pup had a serious health event that redirected our vacation plans. While staying close to home, I had time to play with Fable and Sol, follow a thread I’d been tugging for years, and build the first version of something that feels surprisingly foundational.

I feel inspired to write again.

This letter will mostly assemble from the digital plumage of work already happening elsewhere - a few things I’m doing, thinking, and taking in, with a human hello and goodbye. This first one assembled itself from a very literal pile of email.

In motion

Thriving Quest #001: No-stress email, with agents

Email has been a constant thorn in my side for more than a decade. I’ve tried auto-tagging, fancy clients, new workflows, and a human executive assistant. I always regressed to an uneasy truce: ignore most of it and accept a vague sensation that I had probably let someone down.

So I started a quest to increase my capacity for connection with agent-native email.

  • The foundation is real. My thrivinghenry.com mailbox now receives on a self-hosted OpenBSD server and sends through a relay I can replace. The stack scored 10/10 on mail-tester, and a full restore from backup took 5 minutes and 28 seconds.
  • The first agent is working. An EA triages the live inbox and drafts a morning brief after its golden-set eval caught two real defects before launch. Replies to this letter land in that same mailbox.
  • The build is public. The quest hub now holds the plans, milestones, what’s working, what’s not, and the real token costs. Slack is live as the orchestration channel, and the first agent seats are defined.

The deeper reason I started with email is simple: communication done well creates more and richer opportunities for connection. Over time, that compounds.

On my mind

Then an agent screwed up the launch of agent-native email.

I had asked it to prepare a re-permission email for my old Kit list so I could test every step before it reached another human. Instead, it sent the message to all 144 active subscribers before I approved the experience.

The link technically worked. The experience did not. It dropped people onto an awkward page and asked them to re-enter an email address I already knew. The irony arrived immediately and wearing a tiny party hat.

We turned the recovery into a better system. Each old subscriber received a secure one-click link containing an encrypted identity token. One click now creates or updates the right Resend contact, carries over the name when Kit has one, and records explicit consent for Thriving with AI. No form. No email re-entry. No ambiguity.

After an actual inbox seed, a button test, and a full read-back of all 142 personalized links, the correction went out. By late Tuesday morning, 24 people had made the jump.

The agent made the visible mistake. I made the systems mistake: I left a high-consequence action inside an authority boundary that was too broad.

Trust needs shape. The shape I want is explicit authority, reversible steps, exact seed tests, and a human gate before anything consequential leaves the room. I’m adding those lessons to the machinery behind this letter before #001 goes anywhere.

Taking in

A few things I read sharpened the edges of what I was building:

  • Tech Independence (opens in a new tab) by Derek Sivers gave me a name for the deeper instinct behind this quest: digital self-reliance. Owning more of the foundation makes the agent-native Thriving ecosystem sturdier, and stronger cybersecurity is part of that bargain.
  • Claude Code Is the OpenClaw Alternative You Already Have (opens in a new tab) by Nityesh Agarwal gets at the lived implementation work I care about. Getting an agent running is the easy part; deciding what to hand off and teaching it how to operate is where the useful, recursive work begins.
  • Claude Code for Product Managers (opens in a new tab) by Marcus Moretti made something click: I am, quite literally, the product manager for my personal brand. GitHub as product memory, now/next/later prioritization, and narrative pulse reports give shape to the system I’m already building - and perhaps to a future set of Thriving tools for agents.

Two doorways

Email is foundational to how I show up for people. Getting it right compounds capacity for everything that comes next.

The staycation was not the one I would have chosen. I’m grateful for the space and slowness it gave me, and for the feeling of coming back to this.

Hello again, world. It feels very good to be here.

~h