Post by @brewsterkahle@mastodon.archive.org

@brewsterkahle@mastodon.archive.org

2026-04-29 14:50 UTC

Claude Code is good at doing research! In this case helping find the disposition of 126,000 digitized US Supreme Court dockets (cert denied or full opinion), and then reporting why in a archive.org review (using like a wikipedia discussion). This took real hand-holding and QA to be sure, but it is super helpful. Looked at court listener, the supreme court site (and the old one via the wayback machine!), the Caselaw Access Project at harvard. So good. https://archive.org/details/us-supreme-court

Replies (5)

@subterfugue@sfba.social 2026-04-29 14:59

@brewsterkahle@mastodon.archive.org imagine a tool actually designed for such work rather than that crap anthropic produces

Open ##2652600
@sb@metroholografix.ca 2026-04-29 15:11

@brewsterkahle@mastodon.archive.org Its really disheartening to hear that you are using those tools, built on the theft of human culture - enriching those who are most able to and desire to cause global death and destruction. It's pretty fucked up, actually, given your mission statement.

Open ##2652601
@morgan@leds.social 2026-04-29 16:04

@brewsterkahle@mastodon.archive.org also good at assisting with war crimes! https://futurism.com/artificial-intelligence/pentagon-ai-claude-bombing-elementary-school

Open ##2652610
@loadhigh@bitbang.social 2026-04-29 19:54

@brewsterkahle@mastodon.archive.org Nothing LLMs produce can be trusted. The companies behind them want to monopolize and control access to knowledge. I honestly find it odd that you would consider using their tools to produce content for the Archive, which is supposed to be a trustworthy repository of *human* culture.

Open ##2652611
@hackerfactor@noc.social 2026-04-29 20:06

@brewsterkahle@mastodon.archive.org As long as a human vets the results, I agree with you: AI doesn't replace the human, but sure can speed up information tasks.

Open ##2652612