๐จ SCOOP: Claude Sonnet 5 is releasing later today:
- Knowledge cutoff January 2026
- Launching at $2/$10 per Mtok promotional rate*
- 1M context variant available
*Some changes have been made behind the scenes to prepare for a promo price offering, preliminary values
for folks who feel their Claude Code got nerfed, here's what I've set in my ~/.claude/settings.json to make my CC's behavior more stable. snippet and explanation in thread below -
@karmaniverous I also want to mention that openai api is very slow during America daytime, if you are running on apple m chip it could be easier to use a local model, embeddings are not a rocket science anymore
Iโve been working on ai memory before and I like how spacebot does it, I literally stole it:
- 8 types of memory is recorded, each fact is weighted by importance
- there is a recurring job that decays memory. If you never talk about again, piece of memory fades (gets multiplied by 0.95)
- their golden idea is โthe cortex bulletinโ, a piece of prompt that is composed from the most relevant and unfazed pieces of memory
1) Counter Strike (okay, ready to move out!)
2) Silicon Valley (you just brought piss to a shit fight! ๐คฃ)
3) Arnold Schwarzenegger (I'll be back)
4) Mr. Meeseeks from Rick and Morty
Link: https://t.co/NsbGzSpxZM
The biggest upgrade for my vibe coding workflow is OpenPeon.
I shows HUGE notifications once the agent is done or needs an input and plays warcraft 3 orc peon sounds like "ready to work" or "work-work".
My favorite sound packs so far: ๐งต
Today I stumbled upon solo term by @aarondfrancis and this is a total boost to my workflow:
Claude code now controls all the services that i need to develop the app that I'm working on. (I have web, server + lang graph)
๐งต
Now, paired with Google Chrome extension, my Claude can:
- interact with browser
- check back-end logs
- make a fix
- restart the servers
- try again, until it fixed the issue
All by itself!
This is truly a new level of productivity.