You have an amazing tagline. This is the first time I read a tagline and thought: this is exactly what I was looking for.
But the product seems much more narrow than an actual tool run the whole business in markdown. I was hoping to see Logseq on steroids, and it feels like a tool builder primarily. I love the tool building aspect, but the fundamentals of simply organizing docs (docs, presentations, assets etc, the basics of a business) are either not part of the core offering or not presented well at all.
I love the idea of building custom tools on top of MD and it's part of my wishlist, but I feel little deceived by your tagline so I wanted to share that :)
This is great feedback, thank you. I will say that IS our goal... but we only really launched last week and are still figuring out what resonates with people and what they really want! It sounds like you're saying that the organization aspects are not there, which is very helpful to know... I am not quite sure I understand if you also think the toolbuilding is lacking?
If you are open to it, I'd love the opportunity to hear more. Here or email (alex@moment.dev) or our Discord (bottom right of our website) or Twitter/X... or whatever you prefer.
I have this as a skill Claude created to run the rest. It mentions each skill in turn, see below. It’s not deterministic but it definitely runs each skill and it’s raised a bunch of issues, which I then selectively deal with. Where I can, once an issue is identified, I make deterministic tests.
Text includes:
Invoke each review/audit skill in sequence. Each skill runs its own comprehensive checks and returns findings. Capture the findings from each and incorporate them into the final report.
IMPORTANT: Invoke each skill using the Skill tool. Each skill is independently runnable and will produce its own detailed output. Summarize findings per skill into the unified report format.
Set it up and never managed to have it work. Only thing it did was renaming my sessions on my main cc instance. Mobile did nothing, not even an error message.
This is broadly how I worked when I was still using chat instead of cli agents for LLM support. The downside, I feel, is that unless this is a codebase / language / architecture I do not know, it feels faster to just code by hand with the AI as a reviewer rather than a writer.
Would that book be useful as a reference to introduce data journalism students to AI? I'm less interested in the basics of using the API or claude code etc than best practices for workflows dealing with unstructured data, entity extraction, automated pipelines (with evals)? Although I do have some decent workflows around this I'd be interested in reading from someone who lives and breathes this kind of work. Pure data analysis to me is also something where I haven't found a good bridge between the current "generate a python script for me that I'll double check" paradigm and the spreadsheet centric world of most data journalists.
The book is likely a good fit to this type of work. The chapter on structured outputs shows how to extract out data from text, walking through prompt engineering and k-shot examples to generate json, to pydantic, then batch processing with the different providers.
It also shows how to set up evals in different parts of the book. (Depending on what you want to do, the structured outputs has evals show comparing models/prompt changes to ground truth, and the agent chapter has evals LLM as a judge.)
What's the hook for switching out of plan? I'd like to be launch a planning skill whenever claude writes a plan but it never picks up the skill, and I haven't found a hook that can force it to.
Would you be able to share more? I lead a tiny non-profit org doing data literacy mentoring and I've been meaning to move more of our process docs to Logseq. Although I probably don't need a tool of the level of sophistication of usm.tools, I could take inspiration from your core ideas for our homegrown system.
To understand the approach, you need to first understand the method it is based on.
I have written a simple introduction about it that you can download for free from simpleusm.com, no sign-up required.
Simple homegrown system for processes is not that difficult to do. You basically model the USM process model, templates as instances which you then copy as a basis for editing and make a UI around the editing.
You could even just use JSON files and git, but while the data model is not complex, it is still not simple enough for editing by hand in an editor.
Then the question is what is the benefit. I would say that just using USM to define your services is helpful.
By this approach you can build various stakeholders views to your services that are always up to date and do not require manual labor.
> one is celebrated art, the other is virtually unknown and at best 'one among many baroque northern-talian painters'.
Who claims that 'baroque northern italian painters' are not artists? If anything, an unknown painter is much closer to art with capital A than Banksy, in the traditional hierarchy. So this is a weird framing.
As for time, this is both time taken to create and time spent practicing to reach a certain level of artistry. A speed painter is still an artist, and they reached their speed not by using an AI shortcut but by spending long hours practicing.
The underlying question is how do we tie art and legitimacy: society has always tied both, which is why we have institutions tasked with assigning legitimacy (museums), a hierarchy of art forms where the longest lived are seen as superior (painting over photography), and artists gain prestige not from a single art piece, but from a consistent production of works that are tied together by a shared identity.
On the other hand, a lot of the "pro" AI art discourse I've seen often boiled down to attempts to disconnect art from legitimacy. That's a tough hill to climb.
Pretty strong statement, an as such needs a non-tautological proof. Rich people buying rare things as what society should consider art may not exactly fit that bill.
The thing is what is considered art at any particular time is very nebulous and quite often tied to what the rulers of a country would allow. Trying to say that modern institutions get to decide what art is and isn't is also going to cause definition problems. Does folk art not recognized by museums count at art. The said people who like it would say it does.
Does a person who spends a small amount of time creating something that others consider art, even though that's not what they do, nor will they do it again, have they actually made a piece of art?
Simply put trying to put these rules on the ethereal concept of art quickly devolves into pedantry that makes actual enemies in fields were factions say their ideas are the only true art, and other factions that attempt to destroy the concept altogether.
I was pointing out that time and skill are not universal markers of 'worthy' art. The fact that a random graffiti guy is celebrated, a "big A" artist is unknown is a direct indication that time and skill needed are of little concern in the long run.
> a hierarchy of art forms where the longest lived are seen as superior (painting over photography),
I think this is a bold take - comparing an art form that has been around in a meaningful way for 2000 years to one that has been around for 100 years. Also, if that was true, and not just survivorship bias, shouldn't we consider sculptures and cave wall paintings superior to oil paintings?
Photography, by the way, was considered 'unworthy' by 'real artists' for decades because 'there is no art involved in pointing a box at a tree and pressing a button'. That sounds awfully like the AI debate of today, doesn't it?
> artists gain prestige not from a single art piece, but from a consistent production of works that are tied together by a shared identity.
Or is it that an artist does produce more than one piece of art over their lifespan, so they can, in fact, survive, and, once they become popular with one painting, their other stuff is retroactively elevated?
Art one-hit wonders (or low-hit wonders) do exist. Van Gogh is known for the night sky and the sunflowers, virtually nothing else (unless you are an afficiado). Da Vinci is known for the Mona Lisa - if you are an enthusiast, you might know the Salvator Mundi - the Vitruvian Man I consider less art and more technical drawing. Dürer is known for the hands, the rabbit, and a self-portrait. Shepard Fairey is only known for the "Hope / Yes, we can"-poster.
> On the other hand, a lot of the "pro" AI art discourse I've seen often boiled down to attempts to disconnect art from legitimacy.
That may be related to the circumstance that most of the Anti-AI backlash comes from mediocre artists who do mostly derivative works. If your portfolio consists of furry porn and a broody Heath-Ledger-Joker sitting down, with 'HA HA HA' scribbled over it, sorry, that makes you a 'media / creative', but not an artist with a bit A ... You are essentially doing what the AI is doing: Take an idea, rehash it, minimally, then put it down on paper/your Wacom tablet. If all you bring to the table is 'I suffered for this', your market just shrank to people who enjoy your suffering.
Art is not artistry, art is the idea. As such I find the concept of a 'Street-art Darth Vader' covered in colorful tags that's basically an AI image directly or post-processed more interesting than the 'real artist with colored pencils' Darth Vader in a classical pose that the artist got from a superman comic book cover.
But the product seems much more narrow than an actual tool run the whole business in markdown. I was hoping to see Logseq on steroids, and it feels like a tool builder primarily. I love the tool building aspect, but the fundamentals of simply organizing docs (docs, presentations, assets etc, the basics of a business) are either not part of the core offering or not presented well at all.
I love the idea of building custom tools on top of MD and it's part of my wishlist, but I feel little deceived by your tagline so I wanted to share that :)
reply