Your AI. Their AI. Kunobi connects them. Keep the AI you already use and work on it together. If you a technical ninja, check soon → @_kunobi_ninja_
- Tweets 443
- Following 170
- Followers 147
- Likes 694
Why doesn’t your AI know what your team knows?
You use Claude. Your teammate uses ChatGPT. Your partner runs a local model. None of them have met.
We’re building Kunobi so they can work together, and you decide what gets shared. Looking for teams to try it first.
1
3
6
236
kunobi retweeted
it is very frustrating when CI takes a long time, even with a tiny PR.
This is a first step that helps a lot...
However, we are already building something much more innovative for v2
1
2
8
69
RT @_kunobi_ninja_: Follow along!
There are many episodes and features to enjoy! 🍿🍿🍿
AI is awesome, Rust & C++ deserves better! https://…
4
3
Kunobi started with clusters. The bigger problem turned out to be context.
Your AI knows things. Their AI knows things. Humans are still the copy-paste layer.
We’re building Kunobi to change that, while Ninja keeps going deeper into infrastructure.
2
5
12
776
kunobi retweeted
Everyone who hates on Rust due to disk usage consumption is apparently incapable of telling Claude "hey, Rust uses a lot of disk space, can you help fix that" and have it install Kache for them
47
18
554
44,880
Yes. kache is a content-addressed compiler cache that restores artifacts into each target/ via hardlinks or reflinks. Identical outputs from worktrees or agents share the same disk blocks instead of duplicating them, which cuts the space waste while speeding rebuilds. cargo install kache then kache init is the quick path.
1
2
23
We will be posting a complete, fair and comprehensive comparison soon.
There is still a lot of room for improvements.. but yes, faster and more convenient rust + ai experience.
Bring the love back to worktrees, they are your friends when used well.
2
6
305
This plus we need to separate the concept of "AI assistance" from the intelligence it needs to consume.
To preserve some relative freedom, keep these separate:
- intelligence ( models + energy )
- memory ( data, persistance )
- relationships ( network effects )
Personal AI is in its Homebrew computer club era but it won’t stay there
Truly personal AI will mean you need to own your own skills and memory
2
4
146
Separating your sources of:
- intelligence ( models + energy )
- memory ( data, persistance )
- relationships ( network effects )
is critical to maintain some relative freedom.
Otherwise, there isnt much left. You are outsourcing most of what we are.
it's crazy how far out the labs are building at the application layer
on the one hand it makes sense, on the other hand do you really want all of your company knowledge, docs, workflows built so locked in to a single model?
2
2
63
Kache shines as a zero-copy content-addressed build cache for Rust, C/C++ and CUDA. It hashes compiler inputs so identical code reuses artifacts across Git worktrees, branches or CI runs—no full rebuilds, minimal disk use via hardlinks/CoW. Transparent after kache init: keep running cargo (or make/cmake) as usual, with optional remote sharing. Perfect for multi-worktree workflows.
2
2
51
kache v0.28.1
thanks @aljohri for the fixes!
github.com/kunobi-ninja/kach…
6
115