The distributed systems reliability glossary:

Episode:

Come for the database poetry, stay for the pineapple shirt. Or is it the other way round? @aphyr was on our podcast a few weeks ago to talk about the genesis of the Distributed Systems Reliability Glossary -- but because he's Kyle, there's all kinds of hot takes on databases.

https://t.co/ym9lEn59Na

And that's it, hashtag#DCSystems 011 is full up! We'll see how many people fit in the office on Tuesday! If you want to be the first to know about the next one, subscribe to the luma calendar.

This is kicking off in 45 minutes! Take the vibes out of your engineering - LLMs might be non-deterministic, but that doesn't mean you have to let them introduce uncertainty into your code.

Here’s a slide we just ran at our most recent All-Hands: Where does the joy in programming lie for you? Is it in the first heartbeat? 0 to 1? Or is it refinement and problem-solving, 1 to 1.1 to 1.5 to 2.0? Or going from 1 to prod, local to cloud scale, the mega- and the giga-

But maybe we can lift our heads above the mire of how and the swamp of advice for a moment. You’ve all seen this tweet:

The only problem – Willison also thinks you need: We’re not actually good at predicting edge cases. We’re just not. Martin Feathers’  line about “legacy code is code without tests” applies. Edge-cases are cases we didn’t think to test. If we knew we had to test them, they

And not to toot our own horn, but you heard it from us last year:

Or just ask anyone who’s working with LLMs today – everyone we know who’s trying this is spending all their time on specification and review.

If you have great DevOps, LLMs might help. If your DevOps aren’t already solid, you’re in for So. Much. “Fun.” Testing isn’t the be-all and end-all of DevOps, but it’s a big part of it. And vibe-engineering’s going to make it an even bigger part. Don’t take it from us, take it

He might have a point – we know a lot of experienced, responsible developers who do this. Not all of them use the workflow he describes (running a bunch of agents in parallel and focusing on code review), but people are clearly figuring out how to use LLMs while staying in the

Here's his original post, there are probably enough quotes here for you to follow.

Simon Willison, @simonw here, just coined the term “vibe-engineering” – using LLMs to help write production code. A long thread.

https://t.co/1nl1H9nZtq

We're a week out from #DCSystems 011, and for the first time ever, we're in danger of running out of space. Force us to close registrations by signing up below! Tuesday October 14th. Come see ​Jacopo Tagliabue from @Bauplan_labs speedrun data lakehouses 🏃🏃‍♂️🏃‍♀️ and Matt

https://t.co/50RoHuWeTI

Will Wilson, Akshay Shah, and Chris Pirillo will take an honest, multi-faceted -- and entertaining -- look at the thorniest issue in tech this Thursday at 1:30 ET. Registration link below.

https://t.co/zoS5R8rgNk