<oembed><type>rich</type><version>1.0</version><title>john wrote</title><author_name>john (npub1v7…utfk0)</author_name><author_url>https://yabu.me/npub1v7k63c6y2vktlqhsuupywt3yc7ykursujc34at964f9cv9s9y9csjutfk0</author_url><provider_name>njump</provider_name><provider_url>https://yabu.me</provider_url><html>Keep in mind that we still have human judges and juries.  Something I say often on this point is that some judges don&#39;t want the &#34;perfect&#34; memoranda or brief, they want one that will give them cover to do what they think is &#34;right&#34; not what the law demands.&#xA;&#xA;On the other hand, there is an opportunity.  Almost all Federal hearings and recorded and transcribed, and all the filings are public (for the most part).  I am sure the white shoe firms have already done this in a proprietary way, but I&#39;d be willing to bet if deep learning were deployed on  this enormous corpus of language, it would reveal serious problems with our legal system like if you wear a brown tie on tuesday _this_ judge is more likely to grant your motion.  We&#39;ve already seen things like this in the criminal context, e.g., sentencing is more harsh when the judges are hungry, etc.</html></oembed>