Artificial Intelligence comes to NotePlan

The developer of the NotePlan app continues his push for world domination :slightly_smiling_face: with his latest release by adding two ways to use Artificial Intelligence in NotePlan as well as an Obsidian migration tool billed as “An easy way to move from Obsidian to NotePlan.”

2 Likes

That’s an impressive update.

This is an intriguing use for me, because it’s applying LLM technology against one’s own content. I have no interest in LLMs for general knowledge writing, search, etc. but this could be useful.

I received the email announcing this update and promptly downloaded NP (again). I think this software is pretty great, but the subscription is pricey and I already own tools that replicate much of the functionality, albeit not in an integrated manner. However, I always loved that NP keeps notes and tasks in the same document and Eduard’s consistent development and improvement of NP are to be applauded (and supported).

I came to the forum to start a conversation about this that I thought you’d want to comment on @karlnyhus given your own use of NotePlan, and I see you beat me to it :joy:

I’m not keen on NotePlan seeming to be transitioning more to PKM than task manager - it’s value to me is that it’s one of very few markdown/plain text notes-based task managers, and I’m not keen on the app losing sight of that (I.e. it’s not Obsidian).

But, this opens up interesting uses for its ongoing role (for me) as a scratchpad alongside task manager. I leave NotePlan open to the side of my screen much of the day (mostly just because I have the real estate to do so, I have an ultra wide monitor with a portion of desktop that I rarely use), and I regularly jot down fragments of thoughts so that I don’t lose track of stuff while I’m doing something else. This new Open AI integration could speed up the processing of some of these fragments (e.g. if I need to make a proper note about something, or write a new set of tasks around an issue I’ve spotted). I am also superficially impressed by the simple markdown table creation that was demonstrated!

However, I don’t know if I’ll actually use the AI, and I don’t agree with some of the use cases demonstrated in the video by NotePlan here. We’ve discussed the ethical issues around using AI for content production and we’ve discussed disclaimers for written material. People expect that if they’ve received a personal message from someone, it was written by the individual, not AI. I think this is a slippery slope that will likely devalue written communication :mountain_snow: (And given that for most office workers written communication is the main way we talk to others, this seems not a great path to go down.)

I’m still thinking through my thoughts around AI, but I’m starting to feel there is a very clear boundary forming between “AI for my own use” versus “AI that will venture beyond my personal domain”. So in the context of the NotePlan demonstration, AI to format a markdown table or tidy up some notes for myself is fine, but AI to write an email or an article still has ethical issues.

[Given the topic, I feel I should add a disclaimer that these are my own rambling writings and weren’t written by AI, which would probably have been more succinct :grimacing:]

2 Likes

I thought it would be fun to run my reply through NotePlan’s OpenAI integration just to see what it did. I instructed it to “shorten this, keep it relaxed”. Here it is:

I’m not overly thrilled with NotePlan moving more towards PKM rather than task management, I value its task management feature. I also find its new AI integration could be useful for processing note fragments faster. Still, I’m unsure whether I’ll use this feature, as I have ethical reservations around AI creating content, like personal messages that one might assume are written by an individual. I feel a boundary should exist between AI used personally versus AI venturing beyond my personal sphere. For instance, AI for formatting markdown tables is fine, but writing an email, not so much.

I have my own doubts about NotePlan’s apparent efforts to justify its hefty price tag. (Along with other previous subscribers, I’m grandfathered in at half the new price.)

I use NotePlan for its fluid and capable Markdown editor, its beautiful text display, and a flexible user interface that really works for me.

It is important to me that my notes are stored on my Mac and sync’d to my iPhone via Cloudkit.

And I appreciate the many improvements that continue to be made … as long as I can turn off or ignore the ones that aren’t important to me, like AI. :slightly_smiling_face:

1 Like

This is very much why I use it as well, along with impeccable sync and absolute adherence to including everything in plain text files.

I have no interest in using the “AI” features to write notes, and much of what I do, I’m not going to send off to someone else’s server (part of the reason I generally avoid web services).

But I do find the LLM integration intriguing as a way to turn quick notes into a list of tasks, and similar transformative effects. I don’t know if it’ll find a place in my workflow, but it’s one of the few LLM approaches that even has a chance of doing so.

1 Like

Since I don’t consider AI essential, I’m not sure I’d be willing to go over £6/month for NotePlan. I have enjoyed most the improvements done since I started my subscription, but I consider them “quality of life” improvements I think; they weren’t essential to my use of the app, but now that I have them I appreciate them (and I’m happy for now to keep paying for that).

Actually, we’re touching on an interesting topic here, about continued improvements of apps and whether apps are good at it. NotePlan and DevonThink for me are two apps where big updates generally bring something useful that I didn’t know I needed until it arrived. To a degree I look forward to their updates. Some other apps, I look at the changelog and just think “well, I won’t be using any of that”. I think there’s a temptation sometimes to push through big updates even when really all the developer needs to do is to keep on top of bugs. Some apps are fine as they are and don’t need a continual cycle of “improvement”. (And whilst I might be happy to pay for bug squashing, it can be tiresome to be paying a subscription for an app to get updates that you don’t feel it needs). I guess it’s a magic combination of the app team having a proper understanding of what their customers want (as opposed to just thinking they know what people want!) and my desires being in sync with the majority of other users.

With this update I do wonder what percentage of users want AI in NotePlan, and for those that do, what do they actually want it to do? (Do they want AI that can write an email, or do they want AI that arranges the tasks based on personal criteria, for example?)

3 Likes

“Software as a Service” is best when it is literally service as a service … e.g., there are costs to providing whatever the service may be.

One of the sins of subscription is what you’ve described: charging for future development when the product was already good enough. In these cases the Agenda model really is the best. It’s the model with the most respect for users/customers, and hence it gives the dev team a lot of integrity.

But it clearly does not generate the most growth. Alas.

(None of the above is a direct critique of Eduard, NotePlan’s developer, but the industry in general.)

1 Like

Has anyone tried this migration feature? I asked on Reddit what it does but didn’t get an answer, so far…

This is becoming more and more important as the software industry has almost abandoned “one-off” purchases in favour of subscriptions, for all the reasons that everyone knows.

There was a phase in the industry when a developer could be a “one-hit wonder”: make a fortune by selling millions of copies of something, never support it well or make new versions and quietly disappear when competition appeared. This was not always good for the user, but kept the wider world of software creative and dynamic.

We now have a much more mature market, dominated by a few developers and much harder to break into. That brings a requirement for steady revenue, which builds an expectation for more features, so things get bloated and go wrong (side-eye at Adobe)

I’d love to hear more about refinement. You don’t always need new features, but even very mature software can be refined so it is utterly robust and a joy to use.

Your best bet is to try the NotePlan Discord. (I am not migrating from Obsidian, and I lobbied successfully for a control to hide AI integration controls, to avoid accidentally sending data to another server.)

1 Like

I did a test and was underwhelmed (see my report here: Reddit - Dive into anything

They reached out to me and said they wanted to release an update.

I particularly hope they not only remove the duplication of headings but change how the core of noteplan determines the name of a note from “what the first line says” do what the first level 1 heading says, regardless of where it is.

There seem to be three approaches to note titles: filename, first line, first tip-level heading (presumably with some kind of fallback if there is no top-level heading)

Sadly, these aren’t very compatible. In an ideal world, plaintext and markdown apps that take portability seriously will provide a user preference of some sort.

Not sure I’m following. Where is “there”? In the world of note-taking apps?

1 Like