Comment by brundolf
2 days ago
Feedback: the UI feels like a direct reference to VSCode, which is familiar to software developers, but not to lawyers. If you're hoping this will be adopted by lawyers, I would focus on making the UX familiar to them. Look at software that they already use, and mimic those idioms insomuch as it makes sense to do so. I would also have the base web domain link to a normal home/info page, not to the demo directly. And maybe prefill the demo with some actual content (documents/etc) so people can really see what it does and how
Good luck!
Great feedback; and I do agree. The HN link goes to the app itself because we're impatient, but there is an actual landing page most visitors hit.
I've gone back and forth on the UX idea, and while I do agree, it's important that Tritium selects for users that are going to be able to quickly adopt the newer concepts. Just simply presenting a "better Word" isn't really going to move the needle. It's really a shift in expectations. That said, I have recently backed off defaulting to dark mode to make it feel slightly more familiar.
I think software people tend to underestimate the value of superficial familiarity. By all means, adhere to your new concepts and mental model. But even things like coloring, placement of the menu bar, the icons that you use, the organization of the UI, etc can go a really long way
Think about programming languages- ones that introduce radical new concepts may still employ familiar syntax/naming to smooth the transition for newcomers. Rust mimicked C++, TypeScript extended JS, etc. These languages were made to introduce powerful new ways of thinking about code, but by appearing as similar as possible to what devs already knew, they freed up more brain cycles for people trying to adopt them. They didn't muddy their concept-space for the sake of familiarity, but they didn't introduce any more unfamiliarity than they actually needed to for the benefits they wanted to give
No attorney who is flummoxed by this UX is going to touch an AI product in any meaningful way. Making legal tools for lawyers who would otherwise be using cuneiform tablets or the dictation pool is a waste of conversation. Looking similar to the tools a seventy five year old lawyer uses is like making an F1 car that would look familiar to Jackie Stewart: yeah, it’d probably help him adapt, but not enough to be competitive with an actual car.
Dig the idea of this product, will give it a whirl tonight.
Source: attorney, former dev
2 replies →
> Rust mimicked C++
if anything, it didn't mimick /enough/
I really like the idea. I could see some of my academic collaborators use something like this because it has features typically only supported when working with plaintext. A lot of academics do not love working with LaTeX.
But I would push back a bit against the UX and it being a "better Word". It is not immediately clear from looking at the website whether you support tracking changes. If you support editing Word documents, why aren't basic editing features, like font selection, size and weight, exposed in the UI. (I am viewing it on mobile Chrome and I might have missed it because your page doesn't support pinch to Zoom.)
You don't have to make it look like Word but it must be designed to facilitate common interaction patterns needed for working with Word documents.
(If you are building it on top of VSCode you could use its multiplayer features, which could be a good sell. )
Figure out how to make it uncompromisingly productive for power users and then dumb it down, not the other way around way around
There's a couple ways to skin it. In fields where people are happy with products, i think familiarity is good. In fields where people hate the products, you get to go in tabla rasa. In this case they take advantage of the form of the interface through multisearch et al. Instead of resembling legal software, I wonder if they should resemble a court or a briefcase?