You presented the concern from my adjacent comment perfectly (“LLM performance: common law vs civil law” essentially). So is AI possibly just growing the “Reverence for Professional Experience” factor that plays such a big role for legal compensation here in the US?
Genuinely wondering (aka not snarky): Has anyone found frontier models to provide useful research in the context of European civil law systems?
Your comment made me wonder if there are any halfway-acceptable model benchmarks for law tasks? Specifically I’d love to know how the frontier models’ abilities compare between common law vs. civil law systems. My guess would be that an AI in a common law context should have a clearer idea of how a specific case is interpreted/accepted by (common law) practitioners, whereas trying to rely on AI in a civil law context, like Germany, can be daunting. In a few Germany-specific recent examples, the models feel like they present only (maybe too stubbornly?) the “civil law”-based laws. All while negating much of AI’s research benefits because civil statutes are portrayed as being absolutely accurate, binding, and their enforcement (and thereby the legal reality) being uniformly applied. Am I making this interpretation up? If so, how can I prove myself wrong?
I don't know of anything that is working yet, but I know some people working on it (intentionally vague).
Roman law, on which the legal systems of Germany, most of Europe, Turkey etc. are based relies more heavily on statutes than Anglo-American case law, but cases do play a role there, too. That's why a practically useful system also needs to have access to court decisions, commentary etc. - and while the statues are in the public domain, a lot of the other knowledge is owned by specialist publishers (such as: C.H. Beck, Nomos, Mohr Siebeck, De Gruyter, Otto Schmidt, C.F. Müller, Manz, Verlag Österreich, Schulthess, Stämpfli, Dalloz, LexisNexis, Lefebvre Dalloz, Giuffrè Francis Lefebvre, CEDAM, Giappichelli, Il Mulino, Aranzadi, Tirant lo Blanch, Marcial Pons, Dykinson, Tecnos, Lefebvre, Almedina, Gestlegal, AAFDL, Kluwer, Boom juridisch, Wolters Kluwer, Larcier-Intersentia, Anthemis, Sakkoulas, Nomiki Bibliothiki, Universul Juridic, Hamangiu).
Internationally, legal systems appear to undergo a sort of convergence, which means statutes increasingly matter in the U.S. and cases increasingly matter in European law (according to a law partner friend).
Hi,
As for benchmarks for law tasks: I also have a concern regarding how the more popular benchmarks (like the tasks in Harvey Legal Agent Benchmark) can accurately reflect actual work done by lawyers. Not only in civil law systems, but also in non-BigLaw type of work areas.
Most of the law tasks should inherently reflect the legal system they are based on. It is an oversimplification that lawyers in one country tend to do the same legal work as lawyers in another, that's a form of dishonest framing from LLM providers working in this area (it is that important for them to be able to scale and disregard traditional fragmentation of the legal services).
The legal systems of common law countries are quite similar in both how they do legal training and in many areas that are important for business. Like in contract law or tort. That means not only similar concepts and foundations (things first taught to law students), but also a similar approach to what is seen as a legal problem, and what kind of answers people expect from a lawyer. Also, what they understand under "legal research" - what do you actually research and what is expected from a qualified lawyer to know by heart, how are you expected to find the legal default in a detailed question, how certain that default is, what is up to professional judgment.
But even the differences within common law systems is huge in other areas that go beyond these basics. Like those affected by statutes, codes (like CFR or NYCRR) or even local regulations etc. The popular legal benchmarks do not dare going into these more detailed legal work directions. And these laws change frequently, and they are very different from one jurisdiction to another, even within the US, let alone in other countries building on English legal traditions.
(Nobody I know of is building a benchmark that affects these areas of law.)
And the diversity is even bigger within legal systems lumped together as "civil law systems". Even those based on Roman law, or on German jurisprudence in the 19th century... There is not much common in how French and German students study law, even if both are members of the EU and many areas of law are harmonised or subject to the same legal acts of the EU and the same EU Court of Justice.
If your LLM has access to up to date legal databases of those countries, commentaries etc., the answers are still surprisingly useful in research. That said, no professional should rely on those LLMs that merely use public legal databases, scraping latest laws and cases via self-operated MCP servers.
I think the main difference is how lawyers from a civil law system use the LLMs compared to someone from a common law system - and that is based on the different training they receive as law students, how they see "law" etc. Maybe there is no such a thing as absolute accuracy in the civil statutes either, but at least civil law lawyers have a stronger notion of "law is knowable". There is a higher likelihood of having one "correct" answer, even when talking about more lifelike, detailed questions.
(I'm qualified in both a civil and a common law system but not specifically in German law.)
I’m sorry, but how is this output “crazy good” regarding its accuracy? For example, Apple Watch Ultra is barely recognizable as such. (And UI chrome appearance of the tool is the typical new Astra-ish look I find hard to steer it away from.)
Fascinating listener comment, from a librarian to ATP, about the special legal exceptions with which libraries were equipped to have a chance against copyright and DRM tech: https://www.youtube.com/watch?v=u11ar0649Y4&t=373s (timestamped deep-link)
Just for context: this is a 4 yo design study by friend and amazing product designer Jason Yuan, who also worked on Sprout.place and more recently some stuff at Apple (@jasonyuandesign). A lot of the comments currently seem to miss this and are focusing on immediate feasibility, “successful” design, or ability to deploy irl. That is not the point here. Things like this are both creative exposition, as well as corner stones for conversations for the rest of the community.
I get a lot of value from HN, but technocratic communities often fail to understand the value of reimagining everyday things.
They sometimes get so caught up in the technical bits and pieces they don't like about something you end up with responses like those you in this discussion, or HN's reaction to Dropbox, or Shallots Slashdot's reaction to the iPod.
Yeah, this whole thread is why Hacker News has the "orangesite" reputation it does. I wouldn't have shared anything like this on here, and that is a criticism of HN, not projects like Mercury OS.
I thought this was an interesting project. The site is well-designed. The UI is attractive. As a designer, I really enjoy reading about these things. The fact that this isn't something you'd build your personally customized Linux distro on is not a problem for me. (And no, I don't know the designer or any of his friends. I think the challenge the guy took on was interesting, and I liked reading about it.)
I like HN because of what people share here, but I've never liked the community itself, and this thread is a pretty clear example of why.
No, I said the OS didn't look like something you'd use for heavy-duty development work.
I was being a little sarcastic, but I was just amazed at all the people who spent ten seconds looking at this and said "this is useless, it clearly wouldn't be very good for all the extremely complicated powerful things I want to do. Can it handle fifteen different terminal windows? Fail!!1"
The interesting thing to me is that a compromise is totally feasible and doesn’t need its own OS to accomplish. Let me save a “workflow”.
What is a workflow? A collection of apps, all with their internal states saved, and their positions saved on the screen. Let me close my current flow and open a new flow, with as little friction as possible. I could have an “email” flow in the morning, a “repo 1” flow after that, a “repo 2” flow, (and repo 1+2 etc) a documentation/paper reading flow, and on and on. A few macros can probably accomplish this. Maybe it already exists?
On a more fundamental level the authors are totally right about the debilitating distractions of apps and their damn notifications. But you’d need OS-level control to address that.
That's all non obvious. As far as I can tell this was just as serious as any of the other, 'this is the future', post. See the recent humane keynote for example, and the vergecast's reaction to it.
That said, it's absolutely valid to question stuff like this. A lot of design student type stuff miss HUGE usability issues.
Speculative literally means conjectural, hypothetical, toying with an idea for its own sake, though. It's not a word I associate with product of any kind, more with a sketch than a prototype.
It is well known issue in products that require serious 'design' for human interactions that tech nerds are useless as target audience samples.
In gamedev there are 2 groups of people first is 'game designers' or 'creative direction' and second is - 'tech people'. Best games for wide audience happen when 1st completely ignore 2nd on what needs to be done... And 2nd completely ignore 1st on how invisible bits have to implement it. Often 1st can not iterate properly without code changes on every step and it kills best ideas. I have seen a lot of success when design iterations can be done in no code or very low code environment.
It's worth discussing implementation feasibility of product designs. While I think a modern Canon Cat would be awesome, there are reasons why such a device doesn't exist.
For example: When you make a non-dedicated device, you are outsourcing a lot of the functionality to third parties (e.g. via an app-store). This significantly affects the costs of a product launch.
The ability of deployment of an operating (observe the word!) system is paramount and it IS the point here!
Designers should lay off of approaching functional tasks through vague and coined phylosophical mission statements in pace of pragmatism, it is like sawing the coat to the pretty button with esoteric properties.
Forgive me being harsh but I am pretty fed up by the trend in our time when product design is approached like pompous famous fashion designers make weird clothes only suitable for the catwalk and celebrate each other on the otherwise useless piece of textile.
We can instead express creativity and converse ideas through actual usability, we must discuss operating(!) systems through actual usability, through something could be realized and deployed, actually being possible to use for its purpose!
I'm going to disagree with you here. Concepts have their place in the world, one cannot just have pragmatic ideas in a society, nor can you just have idealistic ideas, there must be a balance, they both feed each other.
While those weird clothes by pompous designers can only be worn on the red carpet, they offer cues for things to trickle down to the clothes you and I wear. They're big ideas, exaggerated themes, like an extract, you don't consume the extract alone, you dilute, you mix.
Car concepts are the same, you never see a car's concept version go all the way till production, they're the concept, the idea is to shoot big, dump everything, write it all out, you can edit later.
We've got enough people on this earth that a great many of them can stay up in the clouds as astronauts of ideas, intermittently communicating to us their findings from the vast universe of creativity, with no impact on society's advancement, in fact to great benefit.
Being corner stones for conversations means that people are going to talk about things that matters to them, like usability or ability to deploy. These are good things! That's what design studies are supposed to do!
I don't know why you're attempting to discourage this.
I don’t know how to feel about a promising designer only spending 2ish years with ADT and bolting. Perhaps if this work is paired with a personality of similar righteousness, that could be a team issue - otherwise to have someone attempt to do great things and see them immediately (in the relative product roadmap sense) leave, is troubling from the outside.
Was also about to reply with the Superlight recommendation. Cannot stand the lag of bluetooth mice and was working my way through various mice over the years, usually trying at least 5-10 alts per year. Have not bought another mouse since the Superlight. Other Logitech mice (gaming and otherwise) eventually failed in one way or another. Not only is the Superlight a superior experience — it performs just as well today as it did 1.5 years ago when I got it. Wish the dongle and charging port were USB-C and that the Logitech SW wasn't the hot garbage that it is (I eventually settled on using a spare machine to update its config bc I won't have that SW on my daily) but from the perspective of the actual mouse itself, it's amazing.
As a user of this app and long-time admirer of the developer behind it (co-creator of the fantastic Paper drawing app for iPad), maybe it would help the conversation if I illustrate my daily encounter with this UI.
I’ve been trying to limit the days on which I eat unhealthy foods to Fridays and Saturdays (yay, it’s Friday!). I’m good at sticking to plans like that without help from an app, but when I gave “Habits” a quick spin, I admired its mission to provide a playful, game-like alternative in the sea of habit tracking apps.
Once a day (except Fri & Sat) it sends me a notification. Tapping it launches the app straight to that “checkbox”.
While the description in the post may make it seem like this is too long of an interaction, I promise you that it really isn’t, and that the haptics and animation are indeed a satisfying experience. So satisfying in fact that I stuck with the app even though I’m certain I’d follow through on my challenge without it.
(That certainty, btw, comes from how I personally have set up habits like this for over a decade, with help from a mental trick somewhat adjacent to a mind palace/loci[1] to track and reward follow-through. I bring this up because I have often thought that it might be possible to bring the power of these rewarding mental models to a broader audience via an app. And to some extent that’s what Habit accomplishes via its unique approach. The haptics and emotional resonance of the “checkbox” are a little Pavlovian mental & sensory reward for completing one’s goal. And they lead the user to a playful/gamified representation of their progress through their habit, in the form of a small, abstract 3D scene that gradually builds as you continue sticking to your goal. If I’m in a rush, I ignore that scene. But I always enjoy tapping the notification and checking off the day’s accomplishment.)
As a designer, I could maybe quibble over details of the interaction or visual style (which is customizable btw via a season of themes — another unique component) but as a a user I’m simply excited to have this app as part of my routine, thankful that someone is daring a different approach to app design, and curious to see if there will be more apps that might introduce a wider audience to the mental hacks I’ve found useful for a long time.
If you’ve read this far, I’d encourage you to check out some of the other posts on Andy.works, incl. a great retrospective on Paper’s creation[2], as well as background on the developer’s philosophy behind the “Not Boring” suite of apps[3]. Lastly, here’s the App Store link if you want to give the interaction a try (I promise I have no affiliation): https://apps.apple.com/us/app/not-boring-habits/id1593891243
This is great, Alex! I didn't write about it in this article, but much of the app is based on leading behavioral science advice for building new habits. It's great to hear stories like yours to show that something that looks and feels great can also work really well too. — Andy
I think the press release is lacking in detail. For example, they completely fail to disclose whether recovery was attempted by rerouting power from life support to the structural integrity field.
I believe a geomagnetic storm calls for rerouting power to the subspace accelerators backing your deflector dish, and of course sufficient impulse power to maintain an appropriate heading, not rerouting to the SIF. Please get your spaceship right. It's a technical forum!
Sorry, was assuming SpaceX had quantum slipstream already. But it appears I was mistaken. In the case of a slipstream drive, integrity for entry and exit of the slipstream would be paramount. But if they’re still using warp, then your comment of course is accurate.
Impulse, by the way, is covered by the press release [1] and, I agree, would need to be a priority no matter which FTL drive is used to get the satellites to their final destination.
[1] > The Starlink team commanded the satellites into a safe-mode where they would fly edge-on (like a sheet of paper) to minimize drag—to effectively “take cover from the storm”
reply