I'm pleased to announce hledger 2.0 preview release 3!
Highlights:
- conventional file layout for rules/data/prices
- get command
- commodity aliases
- print aligns by decimal mark
- more intuitive precision behaviour
- more robust lot detection
- average cost basis methods
- customisable report titles
- prices --summary report
- a more robust roi report
- and many other improvements and fixes.
https://hledger.org/relnotes.html#2026-06-24-hledger-1993
https://hledger.org/install
@simonmic The output of a genAI model is subject to the copyright restrictions as a derived work of any and all of the training data.
Yes, models can't contribute the creativity to establish new / additional protections, and prompts can't contribute the creative effort necessary to establish new / additional protections, but that doesn't free the output from all copyright.
Instead the output treated like a work created though a non-creative process, and is still restricted by the input(s) to that process.
At least, that is my current reading of the documents prepared by the US copyright office.
The lack of new copyright to model or prompter is covered in the second document. The output as a derivative of training material is in the third document (published preview).
I hear you. As a practical matter I think how copyright law will be applied is very much still TBD. If you haven't already seen, these might be of interest:
- https://hledger.org/AI.html#generating-code-with-ai-violates-copyright-of-the-training-sources-
- https://hledger.org/AI.html#law
Simon Michael
Boyd Stephen Smith Jr.