@kunchenguid@Ozolmada Isn't this true for innovators too? Historically, the bottleneck was the time and capital needed to build and find distribution. AI collapses both.
@plainionist@unclebobmartin I wanted to make it easier and faster to run multiple sessions on the same project. Theoretically it should be faster, but I don't see it in my day-to-day work. I will probably benchmark it vs. the original approach.
@plainionist@unclebobmartin thanks, I tried to recreate swarm in claude code only, and modified the approach a bit: each feature session gets its own feature branch in its own worktree, and merges back through a rebase-and-regate step. I am wondering now if I did not make it too complicated.
@unclebobmartin@plainionist If you run multiple swarms on the same project, do you run worktree for each pack plus indyvidual sub tworktrees for each agent type?
@unclebobmartin@plainionist Out of interest do you use different class od models(ie. Sonnet as coder, opus as architect etc?). I tried some different set ups to reduce the token usage, but I don't have enaugh experience to really asses impact on code quality, of each change.
@MateuszChrobok O czym ma być to opowiadanie? Daj mi kilka słów — bohater, klimat, gatunek (np. fantasy, sci-fi, obyczajowe, grozy) — a napiszę je od razu.
@KubaHajkus@radekspace Grupa mojego syna w przedszkolu też oglądała, dzięki za świetną relacje. Książka go na razie jeszcze nie kręci ale już czeka na półce :).
Jeśli nitka wydała Ci się interesująca, standardowo prośba o retwitt.
To nie kosztuje tyle, co MSR, a potrafi ponieść nitkę na Marsa i z powrotem 😉
18/18
@elonmusk You really compare ruthless invasion with things like Bucha massacre happening (https://t.co/kP2Ay4kkYR) with illegal immigration issue? You are a brilliant man, but this is really living in your own bubble.