Now a startup called Judicata is looking to change that, by developing a search platform that makes sense of legal precedent. Not exactly a prediction engine — since each legal case is unique — but certainly one that can help attorneys better play the odds.

Judicata yesterday announced $5.8 million in VC funding led by Khosla Ventures and its newest partner, Keith Rabois (who previously seeded the company while serving as COO of Square).

So I spent a few minutes discussing the deal with Rabois, and have posted an edited transcript below:

FORTUNE — You were an attorney before transitioning into the tech world. Had you been keeping track of innovation in the legal industry, or lack thereof?

RABOIS: I had to revisit it, and was surprised to learn how similar things still were. It’s been 13 or 14 years since I last practiced law, but it seems that people doing legal research today are doing it pretty much like I was.

So what is Judicata doing that traditional players like Lexis-Nexis or Westlaw are not?

One thing they do very well is reduce, by an order of magnitude, the number of hours and amount of money lawyers need to spend on research. Right now so much of it is still done through keyword searches that only sometimes produces relevant results. You can throw human editors at the process, but that’s obviously very expensive and painful. The key is extracting meaning from thousands of cases at scale, which is what Judicata is doing.

For example, say you’re looking for differences between similar cases in which the plaintiff was a male or the plaintiff was a female. It sounds simple, but right now the best way to do it is to go to old books and spend countless hours taking notes and then comparing those notes. Judicata can help organize and make sense of all that data.

The reality is that what they’re doing is very challenging from a technology perspective. It’s not surprising that the background of the team is from places like Google (GOOG).

Where are they in the process?

When they were launching it was mostly based on a theoretical approach, but they now have a working prototype based on having gone back over a substantial body of law. They will release a beta in September that will work on certain types of law and for California law.

Why begin with limited subsets and California? Does the system work better for certain types of law, or is it just a scaling issue?

I think it’s scale and time. The logic behind the technology should work in all areas of law.

I’ve seen Judicata described as the Palantir of law. Fair?

I personally wouldn’t adopt that label. I don’t think great companies make that sort of comparison, because they each have their own culture, own market and value proposition.

There have been fewer VC-backed startups aimed at innovating in the legal market than in almost any other major U.S. industry. LegalZoom, I guess, is the notable exception to that. Are entrepreneurs just not interested, or are VCs ignoring them?

I’ve only been investing professionally for 10 weeks, so I don’t want to come off like an expert on VC history. But, from my perspective, it’s probably a bit of each.

To be successful in this domain you need some appreciation for the nuances of the profession, and it’s rare to find someone who has that and advanced computer science ability. What’s striking about Judicata is that a majority of the team has both a JD and CS degree. Sometimes I’ve gone by the office at lunch and seen them have the entire company take bar exams and debate the answers. It really is an unusual mix that would be very difficult to replicate.

Sign up for Dan’s daily email newsletter on deals and deal-makers: GetTermSheet.com