“In other cases/situations, I’ve tried just giving a model access to all the docs and saying “write a brief arguing X,” but it’s always terrible at this.”
in “Astra for Law” · 11 voices
MixedAccuracyUnverified voice
10 more voices in this thread
- Hacker News · Sep 18, 2026ClaudeValuesRecommends cross-checking legal documents across multiple models (Claude review, GPT drafting, Gemini, Copilot) and still having a real lawyer verify. (summary, not a quote)open
- Hacker News · Sep 18, 2026
ChatGPTCriticizes“Since this is a thread about using ChatGPT for law, I should note that this is actually one of the most important reasons you'd rather hire a lawyer than ChatGPT to represent you.”
open - Hacker News · Sep 18, 2026
LegoraCriticizes“We also tried Legora's tabular review, there were also numerous halucinated provisions in our due diligence exercise.”
open - Hacker News · Sep 18, 2026ValuesHaving several AI models critique each other's work has raised output from a high-school intern to a second-year lawyer in a year, though it still does not match experts, and AI will increase litigation volume rather than replace lawyers. (summary, not a quote)open
- Hacker News · Sep 18, 2026ChatGPTCriticizes
“I find that when I feed contracts to ChatGPT (which I do all the time) it will try to find 50 things wrong with it.”
open - Hacker News · Sep 18, 2026LegoraValues
“Switching from ChatGPT Enterprise to Legora at my firm was a godsend.”
open - Hacker News · Sep 17, 2026ClaudeValues
“The work still must be reviewed by an attorney - Claude is nowhere near making the kinds of judgments a lawyer would make about this content - but it has increased their throughput from 2-3 documents an hour to 8-10 documents an hour by killing the busy work.”
open - Hacker News · Sep 17, 2026ChatGPTValues
“If both parties use ChatGPT to find the relevant laws or read contracts, they could come to an agreement without expensive legal fees.”
open - Hacker News · Sep 17, 2026ClaudeMixed
“It's good for research and data organization, but terrible for drafting.”
open - Hacker News · Sep 17, 2026NeutralNotes that Astra for Law scores slightly below Claude and Muse on a legal research benchmark, without praising or criticizing it. (summary, not a quote)open