Elicit is an AI research assistant that uses language models to help researchers figure out what’s true and make better decisions, starting with common research tasks like literature review.
I'll join the chorus here and say that Turbo Pascal was my stepping stone from beginner tinkering (BASIC on Apple II and DOS) to professional development with C. The dev environment was the perfect mix of "serious and you can make real stuff including compiled executables for distribution" with approachability for a relative beginner.
Our team's experience has been the exact opposite: Catalyst let us create a beautiful, fully-native Mac app[1] that shares the bulk of its code with its iPad counterpart. A few bumps here and there but overall a good experience.
We designed the architecture to allow full Signal-style e2ee, but given that there are only two engineers on this we had to cut scope on that for the first release. Now we'll be listening to what our customers ask for in terms of where to prioritize in the roadmap. We have quite a few medical professionals and attorneys using Muse in their professional lives, so we've already heard from some of them on this topic.
Thanks for the review. Agree with most of your points here, briefly:
iPhone app is just capture right now, intended for "I just had this thought that I need to muse on later." We don't market it much because it's so limited. Future version will build on the sync infrastructure to offer better capture and ability to look up stuff on your boards.
Board rename on Mac is a quick fix we squeezed in before launch. Will make that nicer soon.
OneNote is pretty great actually, ahead of its time. It does still have the "big list of documents" element and in general has the Microsoft level of design polish, which is to say, not much. So certainly our target market with Muse is folks that prefer something better-crafted and from a small/indie team. But that only emphasizes the importance of those fit-and-finish bits you mentioned. Now that our sync engine is finally out in the world we can spend time iterating on that.
Yes, we do need some way to link/embed/transclude documents between applications. OLE was one failed attempt, OpenDoc is another (in)famous example. This is something we're researching at Ink & Switch, but will be hard to innovate here without OS-level support.
Anyhow thanks for taking the time to kick the tires. I will happily take "Muse is awesome but could be more awesome" as validation of the work we've done so far, and motivation to keep improving.
> Anyhow thanks for taking the time to kick the tires. I will happily take "Muse is awesome but could be more awesome" as validation of the work we've done so far, and motivation to keep improving.
Thanks, that’s the spirit it was intended in. I’m looking forward to it.
Also, while I didn’t mind too much on iPad, I really need Dark Mode support on macOS - I have Dark Reader extension installed in every browser, dark mode in every text editor, and now basically every app I use has dark mode on macOS (except VPN apps, sigh…) but to get the most benefit here I have to expand this white Muse window to two thirds or more of my widescreen monitor and… it’s blindingly bright ;-)
Also on the (very long) list of things we want to build. It's more challenging design- and technology-wise than you might imagine, due to the canvas model and relatively arbitrary content folks can bring in. (Actually realizing this will be a good topic to discuss in a future podcast episode.)
Much appreciated. Sharing the journey is a big part of what has made this project fulfilling.
As another commenter pointing out, the Mac app alone is somewhat limited without the iPad as a kind of ink/annotation accessory--but I'm curious to see if people get use out of it. There are plenty of desktop-focused spatial canvas apps (Scapple and Miro come to mind) so clearly there's potential for utility here.
We researched the name heavily as part of registering our trademark. There are hundreds of hardware/software/internet products and services that use the word "muse" in their name somewhere--and that's true for almost any other single english word or even combination of words.
The practical reality of naming things online in this age is that you try to avoid obvious collisions e.g. don't name your photo editing app Photochop, don't name your computer company Newton's Apple. But there's not a lot of risk of customer confusion between different domains like music sequencer vs thinking workspace.
Simply using MUSE by itself with no other words exposes you to trademark litigation, even from FOSS software. It’s the same business category, according to the USPTO, and the existing project simply named “muse” has an automatic first use to the claim. Merely showing prior use by th undermines your trademark application. I dealt with precisely this issue nearly 20 years ago.
Your assurances are not a legal justification and read like “it’s ok” when it’s really not.
I'm a big believer in tools that make us feel good about our work. Pretty handwriting as an output is in that category.
You'd be surprised how much engineering goes into point simplification, bezier smoothing, etc. One of our team members, Adam Wulf, has worked extensively in this space and even open-sourced a lot of the results: https://adamwulf.me/open-source/
...but point remains that we can make the inking experience better in Muse. It's on our list to work on.
I consider Muse’s ink engine pretty good, although I’ve never found anything that comes close to GoodNotes. So that would be the one to emulate – including nice features like their “pause and hold for straight line”, which I really miss in Muse
I’ve felt the same way with GP; IMO the big issue is that the thing that I’ve written or drawn changes during writing. This removes the pen feeling of the Apple Pencil and feels more like a graphic tool? So like drawing circles feels fine, but writing text and seeing that the text feels wobbly while writing gives me the perception that the pencil is not meant to write text in Muse.
Changing while you write sounds surprising. All ink engines have some amount of retroactive smoothing going on, but typically (and this is the same for Muse) it should only be the last 50ms or so of data, and often that part is still hidden by your hand or stylus tip.
Text formatting (possibly Markdown) for sure. Code highlighting is something I would personally like but depends on how many software engineers decide to become members. Latex is probably out for its complexity, but in the long run some kind of plugins system might make it possible.
Elicit is an AI research assistant that uses language models to help researchers figure out what’s true and make better decisions, starting with common research tasks like literature review.
Front-end engineer: https://elicit.com/careers?ashby_jid=b5e218b8-8730-4254-b026...
Senior software engineer: https://elicit.com/careers?ashby_jid=aa99e2e9-5b15-4cd3-ac9d...