announcements

2026-01-27T14:42:24.287119Z

I just released the first version of https://clojars.org/org.tcrawley/datastar-pedestal-adapter, that only uses Pedestal's SSE abstractions, and not the SSE abstractions from the underlying webserver. The motivation was wanting to use Datastar in an existing Pedestal/Jetty service.

🚀 1
🤩 1
🎉 16
Alex Miller (Clojure team) 2026-01-27T15:29:40.298329Z

Clojure CLI 1.12.4.1602 is now available • TDEPS-275 Use sha256sum in linux-installer instead of shasum

👍 1
👍🏼 1
🎉 1
7
borkdude 2026-01-27T15:31:45.982619Z

I went through this once with bb.

# macOS only have shasum available by default
# Some Linux distros (RHEL-like) only have sha256sum available by default (others have both)
if has sha256sum; then
    sha256sum_cmd="sha256sum"
elif has shasum; then
    sha256sum_cmd="shasum -a 256"
else
    >&2 echo "Either 'sha256sum' or 'shasum' needs to be on PATH for '--checksum' flag!"
    >&2 echo "Exiting..."
    exit 1
fi

😵‍💫 1
fogus (Clojure Team) 2026-01-27T15:49:01.467479Z

surely one could vibe-code a Bash-native impl. 😉

Alex Miller (Clojure team) 2026-01-27T15:52:36.233419Z

yeah, I'm going to wait for feedback and see if I need to sniff it in the posix-installer

borkdude 2026-01-27T15:52:59.666819Z

I'm sure Rich would approve... https://chatgpt.com/share/6978df4e-0628-8012-9607-d50195928b98

🤓 1
2
adi 2026-01-27T17:27:05.659749Z

oh i wonder where it copied the program from... surely at least https://gist.github.com/Qix-/affef08b50686e54e1f2ca18f97a6ff7

adi 2026-01-27T17:33:04.532249Z

(but still... it's wild that "just some matmul" can map squishy english strings to anything comprehensible at all, let alone a plausibly-working program... blows my mind every single time... even though I don't use LLMs)

nathanmarz 2026-01-27T22:13:06.137899Z

Released Agent-o-rama 0.8.0. This release enables structured and unstructured https://github.com/redplanetlabs/agent-o-rama/wiki/Human-feedback on agent runs, with time-series charts automatically created for all human feedback. Runs can be manually or automatically added to "human feedback queues" to organize review across your team. This addition combined with the existing features for dataset management, experiments, and automatic evaluation provides everything needed to systematically track agent performance and improve them over time. Discuss in #rama. https://github.com/redplanetlabs/agent-o-rama/releases/tag/release%2F0.8.0

🎉 7