adriano_f

This is a great point. I highly recommend "Liminal Thinking", which explains how what he calls "Battles for the Obvious" like this get started.

What I've been doing when I want to avoid this "unexpected leading", is to tell the LLM to "Ask me 3 rounds of 5 clarifying questions each, first.". The first round usually exposes the main assumptions it's making, and from there we narrow down and clarify things.

I've read you comment about all the things you tried, and it seems you have much broader experience with LLMs than I do. But I didn't see this technique mentioned, so leaving this here in case it helps someone else :).

I also tried Zed on Linux a few months back, and had GPU/driver issues, so it was either slow or didn't run. Tried it just now and it worked right out of the box, and it's incredibly fast.

I will keep playing around with it to see if it's worth switching (from JetBrains WebStorm).

I don't promote often enough, but my writing platform is basically built for working this way:

https://gingkowriter.com

(came up with the concept when struggling to write my PhD thesis)

Hope it helps!

Thanks for taking the time to add these details. Will add to the arsenal of tools. Best of luck!

My son (11yo) has had this as well for the last 2.5 years. Seeing an Audiologist for treatment.

Can you elaborate a little on "listening to violent noises" approach? When do you do this, for how long? Is it graduated in intensity, like exposure therapy?

Thanks!

Hilarious!

(For comparison, here's GPT-4 getting it on first try: https://chat.openai.com/share/9e17ed25-d9ea-4e72-a9d8-a139ca... )

I've started teaching my (homeschooled, 10yo) son Python, and he's been briefly exposed to some 3D software (including Blender).

I'm curious: how do you imagine this working? Can you name a few examples of the learning possibilities that you have in mind?

Ok, glad it's mentioned somewhere.

Couldn't find any mention first time I checked it out.

I'm the creator of the app that inspired this one, http://gingkoapp.com

They say imitation is the best form of flattery, but I'm surprised Ray that you didn't credit Gingko at all. It's practically identical.

Thoughts?

---

Edit: My bad... it's not on the website, but I see Gingko mentioned and linked in the Readme. Thanks Ray.

Actually, it does Markdown, which includes images, and anything you can embed as HTML.

Also, supports LaTeX.

My wife and I, and our 3 yr old son, got rid of most of our possessions, stored the rest, and went of for a 6 month stint of the digital nomad life.

We both have location independent businesses, so we could have continued indefinitely. But nearing the 3 month mark (which was the longest we'd gone in the past), it started to wear on us.

Constant novelty can be just as destructive as constant monotony. You need a balance, and only you can find what that is for you and yours.

We don't regret the trip at all, though. It made us realize what we really missed (family, green spaces, people who loved our son, and wouldn't mind watching him ;). As a result, we ended up moving back to a suburb closer to family, instead of the decidedly cooler city, 2 hours away, we used to live in.

Quite a pendulum swing, from digital nomad to suburb life, but really, as long as we have each other, and love our work, location doesn't matter that much. Which means, in the end, that being free to move around indefinitely also means you are free to stay as long as you like in any one place.

Interesting approach.

I agree with the premise (code is structured, and shouldn't be handled as flat files). I've played with the idea of using my own card-tree editor as a LISP editor before: http://blog.gingkoapp.com/features/gingko-as-a-lisp-editor

It's an avenue I definitely want to explore.

Lots of interesting points, but I disagree with the logic that "we are not designed to handle more than one layer at a time" necessarily implies that a tree-editor is impossible/undesirable.

Yes, we are only built to think of an idea/code/story/essay at one "level" at a time. You are either thinking about the overall structure of the code, or a particular line. A particular scene, or the overall plot.

But that's precisely why a tree editor, one that separates the levels for you, is so useful. You can always "zoom in" to the details by going right, and "zoom out" to the big picture by going left.

Have a look at Gingko, for our approach to solving this problem: http://gingkoapp.com

This is pretty much exactly what we're trying to achieve with Gingko.

It's a full tree editor, that is being used for everything from PhD theses, to television screenplays.

http://gingkoapp.com https://www.youtube.com/watch?v=egCKZHsICm8

Thanks for the feedback!

Yes, more than three layers deep is our highest priority, along with offline mode.

And you are right, I meant "depth-first"; makes more sense.

Thanks for the feedback, and I'm glad you like it.

My cofounder is working on full offline support right now, as it's been a much requested feature.

"Hypertext" perspective shift... I like that. What's your project called? Can I see it?

PS: My ancestry goes: Italian > Argentinian > born in Kuwait > living in Canada. "Ferrari" means iron-worker in Italian, so it's not a rare surname in Italy.

I love overthinking things!

It's true that if every card had a title, then import and export while maintaining the tree structure would be easy.

But as it stands, if you have the following cards:

    A  A.1
       A.2
    B

this approach wouldn't be able to maintain the structure if the cards didn't have titles (it would just list all these as A, A.1, A.2, B).

It's for this reason alone that we've considered making titles required for at least the first card in a group.

Or am I overthinking again... ?

> that sometimes, the tree is more like a graph

I agree, but I think it's better to start with a tree, and then add graph-like features, than to start with a graph and try to make a tree out of it. Mostly because I believe hierarchy is a more natural structure for our brains.

For reading non-fiction, yes the parallax can be distracting. But I think A) the interface can be improved. B) For non-fiction, it's outweighed by the benefit of having context & details available.

We do have plans for a distraction free mode, similar to "Zen mode" on Github. But offline & deeper trees are a priority for now.

I have studied Scrivener, but didn't know about Ulysses. Thanks for pointing it out.

Most of the site is much more toned down, but I thought I'd share what I believe about Gingko's potential.

I respect Nelson, but his intensity is off-putting... don't want to make the same mistake.

I see. Again, thanks for the insider info.

I didn't consider that presenting too much concept art & story details might detract from the experience of letting the concept blossom in the studio exec's own mind.

I'll have to sit down with some screenwriters again, and see if Gingko is something that they'd be able to use or not.

I agree, the UI needs improvement. No one feels the gaps more than I do. But they say, "launch before you're ready!" ;)

For one thing, scrolling on its own (with mouse scroll), doesn't automatically highlight whichever card is now centered. So it breaks the flow. Clicking, or scrolling with keyboard are the only way right now to keep the context clear.

> the hyperbole is a bit of a turnoff for me.

I know, it's a big claim, and I don't expect anyone else to believe it. I thought about not including it, but it's honestly what I believe. Delusional founder? Check.

> This is really a different formatter for the same structure of text we've already been using

That's true, more or less. Table of contents, chapter & subsection breaks, etc, are all attempts to make the structure clearer.

However, once we go beyond 3 levels deep, this has significant advantages.

One can get an overview of a text by reading column 1, but can drill deeper and deeper into specific sections, and even into other trees that are embedded (or "transcluded") into this one. We could have all of physics, or comp-sci, in one tree, and the reader will always be able to drill in for more, or zoom out for overview.

> Why do we need something new? What's being added? What couldn't I do before that's now possible?

There are great benefits for writing this way, for one. Rearranging entire sections or subsections is just a drag and drop. Working collaboratively is also easier, because each person can focus on their particular strengths

> Hopefully this criticism is helpful.

It is, thank you. All criticism forces us to reevaluate our approach, again and again.

We explored having different color scheme to denote the node distance between two cards.

Unfortunately, it got too messy, and neither of us are designers, so we had to shelve it for now.

As for scrolling, and want the mouse scrolling to work just like keyboard (selects the card at center, reorders other columns appropriately). Haven't gotten around to it yet.

Thanks for your feedback!

That app still exists, and it's being worked on by my friend Ivan Savov, one of the original founders.

You can check it out at http://papers.mcgillweb.com

Yes we could definitely improve the way you interact with the tree.

For one thing, the looping back when holding the arrow key is annoying to me too, and a simple fix we will get to.

As for the branch jumping, we tried other approaches, but haven't found one that works well in all cases.

The core of Gingko, navigating tree structures, is something we are constantly trying to improve.

Thanks for your feedback!

For three levels deep, no we don't need a new standard. Larger trees, linked together will probably need something new.

Thanks. You can contact me through main menu in app (gear at top right).

UPDATE: My cofounder has just found the bug and deployed a fix. Refresh and it shouldn't happen again.

We attach files using dropbox links. It's a workaround, but it's been useful so far.

Offline & import/export are features in the pipeline.