@varunkumar@psankar can't you have a standard file something like this
# Worktree Standard
All git worktrees must be created under:
~/worktree/<project>/<branch>
Examples:
~/worktree/core/feature-agent-memory
~/worktree/app/feature-voice-routing
and route the agent to it.
@psankar@varunkumar Quick tip if anyone does it. Iโve modified my whole machine into agent friendly files. Like a dev folder which is common across all the machine and commit that folder into git.
A dev got so frustrated watching his AI agent write 500 lines for a 5-line problem that he built a fix.
He called it Ponytail. Named after the guy every team has - long ponytail, oval glasses, been there longer than the version control. You show him fifty lines; he looks at them, says nothing, and replaces them with one.
Now your agent does the same. Before writing anything, it looks for a reason not to.
80-94% less code. 47-77% cheaper. 3-6x faster.
The best code is the code you never wrote.
GitHub Repo: https://t.co/WnFp9YNY53
Ricky Ponting getting a 135km/h bouncer to the face and immediately telling Javagal Srinath to โget back and bowlโ is the elite mentality we miss! ๐ฅ No physios, no drama, just pure Aussie grit๐คฏ
Is modern cricket getting too soft, or was this era just dangerously reckless?
Every engineering graduate should be taught to read research papers early in their career. Sadly Iโve encountered many kids who have no clue about it. #AIEngineering
Frontier labs and their outcome defines the tech industry at the moment. China has an edge over all other competitors combined interns of outcomes. The race is on. #AIEngineering
@psankar This has become a nuisance. Iโm appalled by the fact that the so called academics canโt understand child psychology. Recollecting our school days and we probably were last of the generations who had a stress free school days.
HF storage buckets are so underrated and makes life so much simpler if you're doing anything with data at scale.
Before this, it was such a pain when switching between compute providers like Azure, Modal, or AWS.
Azure storage is expensive, and the egress costs are insane. I once spent thousands of dollars just to migrate TBs of data off Azure.
Now I can simply mount a storage bucket on a Modal compute instance or an Azure/AWS VM and not worry to much
If we ever figure out how to load ONLY the active params of an MoE into the GPU instead of the full weights, it's game over.
Data centers would see a 100x efficiency boost.
And we could literally run 1T models like Kimi locally on just 32GB VRAM.
Yeah I know it's basically impossible right now, but who knows what the future holds. Let me dream.
Rejoice, for finally I need not carry my laptop at all!
Kittylitter supports:
- Codex
- Claude code
- Opencode
- Pi
And soon Droid.
I can finally easily use my local models as my agents from a slick app! No more GPT app. No more discord.
I canโt believe itโs free and OSS