My girlfriend is in another data center

Also in v2.7.0: comprehensive Oban Pro support. Rewrote oban-pro-basics.md from 80 to 358 lines: - Pro.Worker APIs, args_schema, process/1 vs perform/1 - Workflows, Batches, Chunks, Relay - Smart Engine config (global/local/rate limits) New Iron Law: "NEVER USE attempt TO LIMIT SNOOZES" From a real production incident — snooze rolls back the attempt counter, creating an infinite loop. Caused 72k orphaned jobs before someone caught it. Thanks to @random for flagging the issue.

I love LogSnag but my fave part of Elixir is not needing third parties or external systems - PubSub instead of Redis etc So I built FYI: product event tracking Slack/Telegram pings for Phoenix. No external services. mix fyi.install github.com/chrisgreg/fyi #MyElixirStatus

claude-opus-4-5-20251120 This endpoint was just added to claude's webapp. One of the most insane weeks we've ever had in AI.

Jason Fried, on the benefits of a light grip. He uses a drum analogy to say: "When I first started, I gripped the sticks too hard, and when you grip the sticks really hard, you don't play as well for lots of reasons. You can't get the same tone and you don't have as much movement. There's no ghost note. Lots of things become harder. Your hand cramps up, so you can't play as long. But by relaxing your grip, you get better." @jasonfried

I wish to see more novel beautiful apps for a billion people created — aren't those the most exciting ones to build? Think back when Twitter, Gmail, YouTube, and Spotify got started (2004–2006). Or WhatsApp, Instagram, Pinterest in the mobile era. All generational companies that changed the way we live. And even the ones that didn't become huge like Vine, Flipboard, or Path were hugely creative and exceptionally well-crafted, laying the groundwork for how we compute today. AI puts everything back on the table. Yet, I see few founders take a swing at it. It does require more craft; you do need to build a more compelling experience from the get-go. But it's thrilling to start the next household name among consumers, and the opportunity is massive.

All the great breakthroughs in science are, at their core, compression. They take a complex mess of observations and say, "it's all just this simple rule". Symbolic compression, specifically. Because the rule is always symbolic -- usually expressed as mathematical equations. If it isn't symbolic, you haven't really explained the thing. You can observe it but you can't understand it.