Stop developer burnout

gm legends, happy Wednesday.

On-Call Health is for actually seeing how cooked your on-call rotation is, Oz lets you run serious agent workloads in the cloud instead of melting your laptop, and Willow for Developers lets you talk through code and prompts out loud instead of hammering away at the keyboard.

On-Call Health is a free, open source tool that shows how heavy your on-call load actually is. It pulls data from Rootly, PagerDuty, GitHub, Linear and Jira, mixes it with quick self check-ins, and tracks everything against personal and team baselines before people hit a wall.

🔥 Our Take: Teams ship dashboards for everything except how cooked their incident responders are. This finally turns all the “yeah, this rotation feels rough” vibes into hard numbers you can’t shrug off in retro. My guess is a lot of graphs will just confirm what the most tired person on the team has been saying for a while.

Ilai from Pretty Prompt started a thread about how the tough bit is not shipping, it is getting people to actually care. He talks about the shift he has seen recently where users are posting about the product on their own, explaining how they use it, and basically doing the kind of word of mouth you cannot fake.

His main point is that you do not get that with clever marketing flows. You get it by doing unscalable things, talking to customers a lot, and making them feel looked after. Then he flips it back to the community and asks how other builders turned regular users into people who recommend their product without being asked.

Oz is Warp’s cloud orchestration layer for agents. It lets you spin up hundreds of coding agents in isolated cloud environments, index as many GitHub repos as you need, and keep long running jobs going from the desktop app, web, CLI, or even your phone. You can wire it into your own systems over CLI or API to build things like bug triage, incident response, or scheduled report bots on top.

🔥 Our Take: Local agents are fun until your laptop sounds like a jet and you have no idea what any of them are doing. This is the grown up version of that experiment, with real sandboxes, logs, and enough compute that you can treat agents as background infrastructure instead of a toy you babysit in one terminal tab.

Willow for Developers brings Willow’s voice dictation into AI IDEs like Cursor, Windsurf and more. It is tuned for code work: file auto-tagging based on your open workspace, variable name recognition, and technical term handling so you can talk through changes and context instead of typing long prompts by hand.

🔥 Our Take: Vibe coding sounds fun until you are writing mini essays to explain a flow to the IDE. Talking through what you want in one go is much closer to how you actually think about a change. If the accuracy is as good as they claim, this quietly turns “prompting” into just speaking your plan out loud.