GitHub had its Satellite event this week and my feeds have been a single colour ever since. Codespaces, the in-browser dev environment, got the loudest cheer, and an official gh command-line tool turned up alongside it. By midday every third post was a screenshot of someone spinning up a full environment in a browser tab with the caption "this changes everything".
It might. I genuinely do not know yet, and that is the honest position to hold on day one.
I want to be clear that I am not being sour about it. A dev environment you can summon without forty minutes of toolchain yak-shaving is a real and good idea, and I have lost enough afternoons to "works on my machine" to appreciate why people are excited. The official CLI is the bit I am quietly happiest about, because the unofficial gap there has been irritating for years and I have three half-finished shell aliases that exist purely to paper over it.
But there is a reflex I have learned to resist, and a release like this is exactly the trigger for it. The reflex is to adopt immediately, to rip up my working setup the same afternoon because a demo looked smooth. I have done that. It rarely ends with me being more productive. It ends with me three layers deep into someone's preview feature at 11pm, discovering the edges the keynote did not mention.
The thing about a launch-day demo is that it is the happy path, polished to a shine, run by people who built the thing. Your path will not be the happy path. Your path involves a private registry behind a proxy, a repo with a build step nobody fully understands, and a network policy that exists for reasons lost to history. None of that is in the demo, and all of it is in your Tuesday.
So my plan is boring. I will play with Codespaces on a throwaway repo, because curiosity is healthy and you cannot judge a tool you have not touched. I will install gh properly, because a stable CLI is low-risk and immediately useful. And I will leave my actual working environment exactly as it is until the new thing has earned its place by solving a problem I actually have, rather than one a launch told me I should care about.
Maybe in six months Codespaces will be how I work and this post will read as needlessly cautious. That is fine. Being slightly late to a genuinely good tool costs me very little. Being early to every shiny thing on launch day has cost me whole evenings, and I would like some of them back. The feeds will move on by Friday. The good ideas stick around, and they are still good when you adopt them calmly.