A giant hedge fund. #EndTheFed

@jb55 @fiatjaf Has anyone tried writing a Postgres adaptor for relays? Insane? How about a simpler CRUD layer?

Sold all my Bitcoin for copper. Not financial advice.

Sunday project: zaptask.

GitHub

zaptask/docs/PRD.md at master · shawnyeager/zaptask

Nostr-native todo manager with encrypted storage and zap bounties - shawnyeager/zaptask

Nutshell

Nashville is under a thick sheet of ice. Only two branches down in my yard, so far. And me with no ice skates.

Amazon just prompted to upsell me on insurance. For drum sticks.

gm, good people. ☕

Just set up OpenCode as a persistent background service on a home server. The aha: OpenCode's web UI is fantastic. Runs in any browser, full TUI capabilities. So I set it up as a systemd service, accessible from anywhere via Tailscale. Now I can: - Start a long-running AI task from my phone while taking a walk - Close the browser, come back later—session persists - Attach from terminal (`opencode attach`) or browser, same session - Auto-restarts on crash, survives reboots Resource cost: ~330MB RAM idle. Basically free. The unlock: fire-and-forget AI tasks from anywhere. No need to keep a terminal open, no need to babysit. Works on any always-on Linux box—laptop, home server, $5 VPS. Wrote it up as a gist if anyone wants to replicate:

Gist

Run OpenCode as a persistent systemd service with Tailscale access

Run OpenCode as a persistent systemd service with Tailscale access - opencode-systemd-tailscale.md

> $ nohup opencode web --port 4096 --hostname 0.0.0.0 Add Tailscale. Boom.

I'm bringing it back.

Podcasters who do guest interviews: 1. How much time do you spend researching before you hit record? 2. What's the most underrated tool in your workflow? Working on something and looking to chat with anyone doing 2+ episodes/month. 20 min, no pitch — just want to understand your process. DM me if you're up for it.

Live vibe coding sessions, but where you hear me hurling obscenities at Claude.

gm, good people. ☕ Have s great weekend.

Trending in my inbox: "inclement weather".

If you're replacing an entrenched product (say Airtable for Excel), features may win. If your product is new, novel, difficult to understand, *always* lead with benefits. How do you understand benefits? Talk to prospective users about their problems and their pain.