the brief

A quiet day still pushed on practical edges: a paper on distilling capabilities from closed LLM APIs; a minimalist Bash wrapper to talk to models without Python or Node; and a useful correction on POSIX shell assumptions. Simon Willison amplifies a healthier 'agent-in-the-loop' framing for reviewable automation.

the poursit · sip · 4 items

pulse

(01)
  • hn/frontpage· feedJun 28, 07:43 PM

    Bash4LLM+ simplifies terminal LLM calls

    Single-file Bash wrapper using curl and jq adds prompts, chats, streaming, and JSON session logs—handy for lightweight CLIs and server boxes without Python or Node.

    Show HN: Bash4LLM+ – A lightweight, dependency-free Bash wrapper for LLM APIs — Bash4LLM is a single-file Bash wrapper for interacting with LLMs from the terminal. I created it because I wanted something simple that worked without installing Python, Node, or any other runtime. It uses only Bash, curl, and jq. You can send prompts, start a small chat, process files line by line, stream output, and save session metadata in JSON format. I tried to make it safe and predictable: no use of the syst...

findings

(02)

voices

(01)
  • simonw/blog· feedJun 28, 09:57 PM

    Shift to agent-in-the-loop

    Simon Willison highlights Jon Udell’s framing: keep developers in control, require reviewable diffs and tests, and make agents collaborators—not unchecked PR machines.

    Quoting Jon Udell — <blockquote cite="https://blog.jonudell.net/2026/06/28/doctor-it-hurts-when-agents-create-unreviewable-prs-dont-do-that/"><p><strong><del>Human</del> Agent in the loop</strong></p> <p>I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the same way we always have, now we recruit agents to join the team. An agent-assisted process need not be a black box that takes in prompts and emits features...

    signal 6hype 1agentshuman_in_loopdev_processsource ↗