I always wondered if Justin Kan’s Atrium closed door prematurely by just 2-3 years. It would have been cool to see a “technology” driven law firm and how it would have adjusted to LLMs.
Self-hostable legal AI as open source is a useful direction in principle. Hard to tell how mature the actual implementation is though, the repo is pretty fresh and the marketing site is doing a lot of heavy lifting compared to what's in the code right now. Will be more interesting to revisit in a few weeks.
Interested to try it out!
Some feedback on the homepage there's nothing above the fold, or directly below that says its a Legal AI platform. I would like a legal AI tool, but I'm not familiar with the space don't know what Harvey or Legora are. It was only the hackernews title "Mike: open-source legal AI" that gave the context.
Presumably this is an issue for the commercial competitors too, but in light of the recent court ruling in United States v. Heppner that AI chatbots can break attorney-client privilege and/or work product doctrine, what kinds of things can this be safely used for? (I would assume you want to avoid sending anything with client-confidential information in it to a service provider like OpenAI or Anthropic.)
Potentially if used with a local LLM and not a service provider, this might protect attorney-client privilege?
For a moment I thought it was some open-source LLM trained on legal. It's not, it's a web app wrapping major LLM providers and streamlining legal workflows, uploading documents, and having the LLM providers interact with them.
Why don´t you put a direct link that redirect users to some proprietary AI providers instead of making it look fancy. (If I ask whatever AI model will produce same outputs/forms, structured as you wish, and even locally).
To qualify as some wrapper you need to add a layer of creativity by you on top of the existing ones.
I believe this is the direction enterprise software is generally going. An open-source base with a very permissive license that then each company can adapt (with claude, codex, etc.) for it's own needs. It's either running it on it's own infrastructure or in hosted environment by the author. I've built a similarly extensible codebase for an ERP: https://github.com/lambdadevelopment/lambda-erp
Thanks. The demo itself is prerecorded - so no LLM calls. It's just replaying from the sql db. In an actual request it isn't much slower though in finding what tools to use, etc. You can test it btw. I've allocated an hourly budget - anything that you query after the demo goes to an actual LLM.
I'm a little puzzled by what this actually is supposed to be. The marketing material on this website suggests that it's meant to be used with a firm's Gemini or Claude API keys. ("A chat interface that reads your documents, cites verbatim, runs multi-step workflows, and drafts and edits contracts end-to-end. Plug in your own Claude or Gemini keys, and keep full control of the models you use.").
If that's true, how does it actually achieve anything with respect to client confidentiality or anything else? (For example, there's the claim "the assistant keeps full context across every conversation and every document." --- but isn't that a function of the model one uses, which is on Anthropic or Google? Ditto the claim "Documents never leave your perimeter. Compliance, residency, and privilege stay under your control." But this is only true if you're not piping them to Anthropic or Google...) Is this just a user interface?
It would be nice if these product webpages included an easy way to find documentation so that one could figure out what the product actually does. I can't find any obvious way to discern if it can be easily used with a local model running via ollama or something, for e.g.
One question I have about legal AI startups/products, is how do they maintain or improve upon billing practices of law firms?
Having worked with a bunch of lawyers, I know that I'm often paying $500/hr to that firm. That work is actually done by a paralegal who is being paid $40/hr, and then I'm being billed through the partner for an extra $460/hr. This is a gross oversimplification, but you get the point.
If the partner needs to bring in $5M a year, how does any addition of tech solve that?
If I'm the customer of the law firm, I would love to have a more cost efficient way to get legal advice. But, I don't understand how those incentives are matched by the partner? I don't really think they want a more efficient result for their customers, they want a better way to get more billable hours. Adding "tech efficiency solutions" does not solve that issue at all.
Inevitably, customers will use LLMs on their own, and as people have noted, lose attorney client privilege (and often get hallucinated bad advice). There will probably be some very comical court room dramas when people try to represent themselves with an LLM on their shoulder.
Am I misunderstanding something fundamental about the legal world that will make a major law firm adopt this tech? I feel like there are some strong reasons they will universally avoid moving in this direction. Long term it will win and there will be blood on the floor, but why would any large firm adopt this stuff right now?
There is a real need in the space and a real opportunity for a solution like this but this is a complete nothing burger of what exists in the underlying code.
The requirements for this kind of product are extensive and complex. The shape of the data layer is complex and nuanced. Absolutely none of this is considered or implemented in the project but it sure is blowing up.
Maybe it's just me, but seems strange to not include what countries legal system this is for, and having it prominently front and center?
Since this is HN, I guess it's fair to assume it's for the US, but since English is used in more countries than the US, wouldn't it be a good idea to say outright what countries legal systems this actually understands and supports? Or is it maybe meant to be country-agnostic somehow? If so, that isn't very clear either.
40 comments
[ 4.8 ms ] story [ 55.1 ms ] threadThat may be confusing on the naming.
That, plus an Anthropic-like logo.
Potentially if used with a local LLM and not a service provider, this might protect attorney-client privilege?
Cool project regardless!
If that's true, how does it actually achieve anything with respect to client confidentiality or anything else? (For example, there's the claim "the assistant keeps full context across every conversation and every document." --- but isn't that a function of the model one uses, which is on Anthropic or Google? Ditto the claim "Documents never leave your perimeter. Compliance, residency, and privilege stay under your control." But this is only true if you're not piping them to Anthropic or Google...) Is this just a user interface?
It would be nice if these product webpages included an easy way to find documentation so that one could figure out what the product actually does. I can't find any obvious way to discern if it can be easily used with a local model running via ollama or something, for e.g.
One question I have about legal AI startups/products, is how do they maintain or improve upon billing practices of law firms?
Having worked with a bunch of lawyers, I know that I'm often paying $500/hr to that firm. That work is actually done by a paralegal who is being paid $40/hr, and then I'm being billed through the partner for an extra $460/hr. This is a gross oversimplification, but you get the point.
If the partner needs to bring in $5M a year, how does any addition of tech solve that?
If I'm the customer of the law firm, I would love to have a more cost efficient way to get legal advice. But, I don't understand how those incentives are matched by the partner? I don't really think they want a more efficient result for their customers, they want a better way to get more billable hours. Adding "tech efficiency solutions" does not solve that issue at all.
Inevitably, customers will use LLMs on their own, and as people have noted, lose attorney client privilege (and often get hallucinated bad advice). There will probably be some very comical court room dramas when people try to represent themselves with an LLM on their shoulder.
Am I misunderstanding something fundamental about the legal world that will make a major law firm adopt this tech? I feel like there are some strong reasons they will universally avoid moving in this direction. Long term it will win and there will be blood on the floor, but why would any large firm adopt this stuff right now?
go look at the auth - it's a call to supabase.
go look at the migrations - it's like 5 tables.
There is a real need in the space and a real opportunity for a solution like this but this is a complete nothing burger of what exists in the underlying code.
The requirements for this kind of product are extensive and complex. The shape of the data layer is complex and nuanced. Absolutely none of this is considered or implemented in the project but it sure is blowing up.
Since this is HN, I guess it's fair to assume it's for the US, but since English is used in more countries than the US, wouldn't it be a good idea to say outright what countries legal systems this actually understands and supports? Or is it maybe meant to be country-agnostic somehow? If so, that isn't very clear either.