new feature in opencode v2: low verbosity mode to hide all the crazy stuff that goes on
before: <thousands of shell output, etc>
now:
+ 1 command, 1 edit, 2 thoughts, 3 reads
kit finally finished the blog post
he covers how OpenCode 2.0 is able to modify everything about itself in real time even during running sessions
btw post whatever easter eggs you're able to find
key thing is to not go crazy with it
just have it available, with access to email, calendar, etc over your primary chat interface (like imsg)
when you have a need, try to use it, if it can't do it tell an agent to build out whatever is missing so it can do it
you end up solving some of your main pain points only unique to you. for example I have a million tennis texts spread across threads and finally have a way for it to give me a summary on schedules and add to calendar
I don't build anything else that I don't immediately need
to do this you should build from scratch too btw, don't use something else. only works for tech people, but this is why the problem is so hard to build a product
I setup a “personal assistant” to help summarize all the things I needed to do the day.
Day 1 was amazing, at least I thought
From day 2 to day 10 ignored the whole thing lol
Turns out I just don’t want to make a bunch of decisions, and an AI agent doesn’t change that. 👇
my wife hates our bot
but I think this is going to win her over: we casually talk about the meal plan for the week over text and it automatically figures out ingredients and makes a kroger order and it shows up at our door
I tested Jev on the most annoying problem in personal finance tools: getting a good payee description from raw bank data
It probably 95% of the way there on the first try, and there's lots of different ways I can improve how to ask it
working on different verbosity modes in opencode
always pays off to build little tools to iterate faster! I've probably adjust 100 little things in here because I can quickly see it with real sessions
we're making worktrees better in opencode
plugins can now register different strategies for managing them, so you can use tools like rift, lane, etc
we'll have a tool too so agents are good at managing worktrees how you want. this used lane!
opencode v2 comes with a terminal builtin. let me show you how to use it!
it's super fast thanks to libghostty (thanks @mitchellh and team)
they are scoped to your session, survive across restarts, and provide a quick way to run anything
I love our little simulation tool. you can ask it to record a video and it can even resize. all from a slack bot
here's an improved prompt input in opencode at small sizes
man I love telling ai to take a large piece of work and break it up into small stacked PRs that you can land individually
it lets you build something big quick but then you can do a final pass to make it production ready by focusing on individual pieces in order
breaking it up was so much work before
I finally get the sandboxes use case. has nothing to do with security
it’s unbounded resources (only limited by $)
Now that I know how to use agents I’m constrained by a single device (3 typechecks at once bring it down)
This clicked when I was wondering why our slack bot felt so frictionless and we started doing so much work there
It’s because you don’t have to care at all where it’s working… so now I want that locally