daniel_grady

Enthusiastic second. I’ve been daily driving Marimo with Claude for several months, works great, also with custom access points.

Congratulations to the Zed team! What a great project.

The newer layout that came along with the parallel agents feature is very nice; even without using parallel agents regularly, this is a breath of fresh air.

Wow I was excited to see the TextMate icon in the screenshot at the end. Good memories.

Oh my gosh. Everyone can always do better, but also, has no one else here had to use Microsoft Teams recently?

Although it's not part of the standard library, toolz is wonderful for rounding out these modules.

https://toolz.readthedocs.io/en/latest/

I made a serious effort to integrate Muse into my work starting around late 2021 or early 2022, and in fact bought an iPad Mini and Apple Pencil specifically to use Muse. The work that comes out of Ink & Switch is always interesting, and I was excited to try some of it out in the real world. Over a year or so, I used it to read and review PDFs (mostly journal articles for work), wrote and presented a lunch-and-learn from Muse, dropped PNG plots from Jupyter for scribbling or easy comparison; I even got one of my colleagues interested enough to use the collaboration features semi-regularly.

It hasn’t stuck though, and I’ve stopped using it; subscription will lapse later this year. I’m sad; like others here I really wanted to like this and for it to make sense to keep using.

I don’t a have clear set of reasons for why it didn’t stick. Just thinking out loud. Partially, I was fighting against my organization — my immediate team is science / Apple / Python, but the larger company is Teams / Windows / PowerPoint, and that’s always friction. Partially, it was a workflow thing — most often I wanted to review PDFs, which live in Zotero, and then it’s like, did I copy that one over yet? Where are my notes about that one? Muse’s PDF excerpting feature is really wonderful; the lack of being able to zoom a PDF, or support for table of contents, was a bummer. Large PDFs like textbooks could be problematic. Partially it was that Muse on iPad vs macOS felt like two incomplete halves — can’t type on iPad, can’t ink on macOS. Partially: things I did in Muse, felt stuck in Muse; not literally true, but copy or export out of Obsidian vs Muse feels very different. Partially: always that nagging concern from lack of E2EE sync, and after Apple launched E2EE for iCloud, Obsidian + iCloud offered the sync I wanted with a subscription I already had anyhow. (Collaboration features aren’t as good, though!)

Anyhow. Muse did so many things well and first in this space, it remains impressive. Many iPad apps (in my opinion) are incrementally different versions of Apple Notes; Muse is a standout example that supports Apple Pencil as well as Apple first-party apps but targets a very substantially different use than drawing. Although I’m setting it aside, still optimistic about what this year will bring for Muse, and wishing the best to Adam Wulf!

The classics never go out of style. Mike Bostock has a nice article about the use of Make for data workflows: https://bost.ocks.org/mike/make/.

Amazing, I can check off “learn something” early today!

> For instance, macOS is a Unix™. It passes the tests, and Apple pays for the certification. But it hides most of the real Unix directory tree, its /etc is relatively empty, it doesn't have an X server – it's an optional extra. And most of all, it's not case sensitive.

What does it mean for the operating system to be “case sensitive”? Certainly APFS is case sensitive, so this must refer to something else?

That’s a great point; I didn’t know about that!

Those are interesting pointers; appreciate it! My own experience over the past three years has been similar. I tried using Pipenv, and then Poetry, for internal projects at my company; in both cases the tool seemed overly complicated for the problem, slow, and I had a hard time getting co-workers on board. About a year and a half ago, I saw [Boring Python: dependency management](https://www.b-list.org/weblog/2022/may/13/boring-python-depe...), which recommends using the third-party `pip-tools` library alongside the standard library’s `pip` and `venv`, and switched to that for the next project. It’s been working great. The project has involved a small team of scientists (four or five, depending) who use a mix of macOS and Windows. We do analysis and development locally and write production-facing algorithms in Python packages tracked in our repository, and publish releases to Gitlab’s PyPI. For our team, the “get up and running” instructions are “clone, create a venv, and pip install -r requirements.txt” and for the software team that manages the production systems, deploying an update just means pip installing a new version of the package. Every team’s got different constraints, of course, but this has been working very smoothly for us for over a year now, and it’s been easy, no pushback, with everyone understanding what’s going on. Really impressed with the progress of the core Python packaging infrastructure over the past several years.

> pip's dependency management doesn't seem to (for Pytorch, specifically)

That’s interesting — I’ve also had difficulties with PyTorch and dependency resolution, but only on the most recent versions of Python, for some period of time after they’re released. Picking Python 3.9 as a baseline for a project, for example, has been very reliable for PyTorch and all the related tooling.

What are some of the reasons that teams use conda (and related tools) today? As a machine learning scientist, I used conda exclusively in the mid-2010s because it was the only framework that could reliably manage Python libraries like NumPy, PyTorch, and so on, that have complex binary dependencies. Today, though, pip install works fine for those packages. What am I missing?

Congratulations on this release! Your writing at bost.ocks.org, D3, and Observable have been big sources of inspiration over the years, and it’s always exciting to see new ideas from this team.

Absolutely. There’s no secret here.

It’s exciting. Between this and lambdas, Java will soon be as good as Python was ten years ago.

> Now `charAt()` allocates memory and has variable latency.

It’s Java, right? Everything is variable latency.

Strongly agree.