Show HN: Lathe – Use LLMs to learn a new domain, not skip past it • Hacker News Board

Re: Show HN: Lathe – Use LLMs to learn a new domain, not skip past it

#91

Post by dchuk » Tue, Jun 09, 2026, 8:07 PM UTC

edot wrote: post #23

dchuk wrote: post #12

I’ve been using this general pattern - a custom cli app for deterministic tasks, skills for the agent harness, run the skills in the agent and it produces artifacts for you by using the cli and its own agentic reasoning - a lot lately for work. Things like “give me an executive brief of the activity in these teams backlogs over the last month” and in 5-10 minutes I have a few page doc I can read that is cited with th…

Can you give some examples of the deterministic tasks? So in your example, was the deterministic task “fetch this team’s backlog”? And then the LLM parts are “process each backlog” and “combine a summary”?

Generally it’s things like that yes, but also stuff like preparing the pulled content to be summarized and then taking those summary batches and preparing for synthesis in a templated format

Re: Show HN: Lathe – Use LLMs to learn a new domain, not skip past it

#92

Post by danielmai » Thu, Jun 11, 2026, 2:05 AM UTC

[flagged]

Re: Show HN: Lathe – Use LLMs to learn a new domain, not skip past it

#93

Post by robinvdvleuten » Sun, Jun 14, 2026, 1:49 PM UTC

[flagged]

Re: Show HN: Lathe – Use LLMs to learn a new domain, not skip past it

#94

Post by williamsss » Mon, Jun 15, 2026, 9:43 PM UTC

This is similar to how I’ve transitioned to learning new programming languages. I used to build stuff to learn. However, with AI assisted coding I don’t have those organic moments I’m stuck and need to search online for the answer.

My learning process now uses Claude Code to serve as a tutor that creates bite sized lessons writing parts I want to re-review, found confusing, what lessons were reviewed and what is next into a long lived CLUADE.md file.