My book would explain to lawyers how to properly typeset different types of legal documents, including court filings, contracts, wills, letters of engagement, and invoices. It'd also teach lawyers how to use the freely-available tools to collaborate with colleagues and clients in the preparation of documents.

Does this sound interesting to anyone?

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451045)

Even if the typesetting has no legal effect on the document (besides compliance with the jurisdiction's formatting rules), clients love to see something that looks nice. For example, clients love receiving elaborate-looking stock certificates for their newly-formed corporations.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451056)

lawyers hate dealing with TOCs and all that stuff, and the rendering on TeX is simply beautiful

i have only ever met two lawyers IRL who were aware of TeX, and they both commented immediately during my interviews on it (my resume was rendered in LaTeX)

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451068)

in addition to looking cool, it immediately conveys that the resume was written in LaTeX

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451075)

lol @ the lies + fraud that is forming a bidness

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451082)

http://www.hotdocs.com

http://www.contractual.ly

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451061)

contractual.ly is interesting but it looks like, after launching as a document generation startup aimed at lawyers, they are now focused on contracts for procurement depts.

i've been posting for a couple years now that there should be semantic markup for contracts and a contract-specific text editor that isn't Word.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451085)

I'd personally store every standard contract clause in a document store (like mongodb or something) and then have a script generate the LaTeX based on the information in the database. It's also a good approach to cataloging issues and holdings to facilitate writing memos, briefs, and motions, but that's getting more to firm management software (which needs to be improved).

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451107)

the OP's proposal overlaps a bit with this but appears to be targeted toward typesetting which is more of a formatting issue

but because TeX is so customizable (and resembles computer code, iirc) it would probably be easy to grab the data and have google pull all ur published case cites etc. or to otherwise parse it and do useful things with it

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451078)

i have no idea if this is actually useful from a corporate lawyer perspective but it would be kinda cool to have libraries of indemnity clauses or wutever and just insert them

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451095)

you could do something similar with discovery responses: feed a template into your customized OCR software, then feed it the incoming discovery request and use something like a neural net to generate a rough draft of the responses (which are always canned and essentially the same for every client/partner pairing)

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451100)

FWs are not

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451101)

welp there goes the biglaw business model

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451104)

Automating responses to discovery requests is even harder, because there's no machine learning approach that's good enough to approximate the human mind in the manner you suggest. Because there aren't enough examples of request-response pairings, you'd need an algorithm that understands the language, rather than finding patterns in the requests that yield particular responses. But stuff like document review can be automated using semi-supervised techniques that require an expert to tag some documents and then evaluate the algorithm's performance as it iteratively adapts.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451135)

probably more like tens of thousand if not more

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451170)

christ, there is so much that could be done if biglaw hired software engineers. inefficiency is to their benefit though.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451192)

one time i did a doc production w/an outside vendor. i wanted the vendor to separate thousands of files depending on their filename pre-fix. the vendor finally finished after 14 hours. i later learned that the vendor had been DRAGGING AND DROPPING INDIVIDUAL FILES, one by one, for that entire time. even tho they could have just done a find ./ -name "prefix*" -exec mv (or whatever the reg expression would be).

nobody complained about the bill even tho it was ludicrous

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451223)

but, it's almost all confidential WP and the system might complain if u were to pull it all at once

perhaps u could mine the data if u were brought in as an inhouse consultant

but keep in mind that most biglaw firms don't have efficient ways to search for their prior work product, leading to crazy situations: i.e. rather than digging out a memo that already was answered, there is every incentive to get some associate to re-research the issue and rewrite it from scratch

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451206)

http://deepdive.stanford.edu/

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27454887)

the only place where efficient software is welcome is to handle non-billable stuff like firm management, and for contingency fee cases.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27451134)

You're solving a problem they don't have. This is typical when tech doods or aspies think of how to apply technology to a business because they don't understand the pain points. Ask any solo what keeps them up at night or prevents them from scaling their business. It's all about getting clients in the door to spend $500 on real estate paperwork or handle some small insurance matter. The next biggest thing blocking scale is simply how much time it takes to correspond with clients, meetings, phone calls, etc. Physically typing and formatting documents is not the problem (yet).

Automate getting clients and replace their expensive secretary with Siri and you might get them to that point.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27709992)

Automating law firms is pretty much my goal; there's no justification for most of the profession's absurdities beyond risk mitigation and making a comma someone else's nightmare.

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#27710187)

I've put considerably more effort into the creation of a massive database for all of my research (reference materials that I scan at the library, motions that colleagues have sent me, etc.). The database has become pretty massive, and I'm currently considering a dedicated text search engine for it rather than relying on postgresql's full text searching. Sphinx, for example, provides pretty awesome advanced text matching operators (phrase matching, proximity matching, strict order matching, sentence matching, paragraph matching, etc.) that will let me retrieve documents using the familiar Westlaw/Lexis boolean search operators (e.g., "prestigious w/s (law w/2 (school | center))").

(http://www.autoadmit.com/thread.php?thread_id=2825093&forum_id=2#36838791)