MUSE FEATURE ALERT: y'all have been asking for this one. your @muse is getting its own email address. add it to a thread or forward it stuff and it’ll handle it for you.
muse mail!
Update: it takes like 3x longer to reason and latency is like 5s/request but we get 97% cost savings with the same evals. Nice work @Zai_org
https://t.co/kfx9bmHzIW
Amazing research done recently amongst all the benchmark gloating. I will be testing out GLM 5.3 Flash on some projects I have using Sonnet 5
https://t.co/4MRpcMnl5k
I used to be all for keeping docs present tensed, but I’m getting a huge benefit from agents leaving traces of trials and things it was doing no matter how verbose
I've been loving how Codex implements subagents, backtracking, and forking (especially through /btw) to manage context. I'm fascinated by how tools manage to cater to their power users without alienating their casual users and I think this is a great example
A friend and I have been building a project together for 104 days. Some stats:
- 376 merged PRs
- 58 deploys
- 200,679 LOC churn
- 36% of the codebase are tests
We live in such cool times