A programmer can trace any line of code to its author, date, and historical state. Lawyers reading federal statutes lack this convenience, especially as legal research increasingly relies on language models unable to track legislative lineage. The US Code, outside titles enacted as positive law, is only prima facie evidence, with the Statutes at Large governing discrepancies.
The Office of the Law Revision Counsel publishes release points, each current through a named public law, with amendment notes recording which public law wrote a sentence. Effective dates often reside in separate notes, with default enactment on the day of passage. For example, the small business debtor definition in 11 USC § 1182(1) was created in 2019, modified by the CARES Act in 2020, extended to 2022, lapsed, restored retroactively, and adjusted to $3,424,000 by April 2025.
A study from the Technical University of Munich found language models fail when applying superseded rules or ignoring temporal validity. Retrieving statutory text improved accuracy from 24% to 87%. However, generating old text from amending laws produced 93–99% similarity but only 50% exact matches for single amendments, with legally significant errors.
To address this, a system was built applying public laws in date order to older text, comparing results to newer release points. Of 365 changes across nine titles, 361 matched exactly. Each change includes the law and effective date: 261 by default, 72 by effective-date note, 29 by amendment date.
One depends on a pending agency determination. Users can now select a date to view the text in force, with changes marked and colored by the Act that wrote them.
Source: Hacker News · Summarized by HeadlinesBriefing