The Painful Foundations for Tools of Thought
10 min read

The Painful Foundations for Tools of Thought

Notes and takes on note-taking apps and why some foundations take precedence over features

The Notion of a Second Brain

One of my oldest dream projects had been to collect all my ideas and thoughts and create a Wikipedia of sorts. Later, I found the right term for it is personal knowledge management or as Tiago Forte calls it Building A Second Brain.

Over the years, I'd done this in physical journals (back in the day) and then moved on to Evernote and OneNote. It wasn't perfect, but it got the job done.

Then I came across Notion. Everyone seemed to be talking about it and it got praised unanimously. I was smitten by its simplicity and ease of use. It could do a little bit of everything and soon, it became the central tool for my personal Wikipedia. It seemed so much better than Evernote, which at the time had stagnated while Notion kept introducing features and offered a lot more. It seemed to be exactly what I needed. Well, almost.

Notion's slow motion

I was somewhat of a feature nut and like the vast majority of people, I always wanted more. That was until I read Brilliant Hardware in the Valley of the Software Slump where Craig Mod talks about the significance of speed:

Speed and reliability are often intuited hand-in-hand. Speed can be a good proxy for general engineering quality. If an application slows down on simple tasks, then it can mean the engineers aren’t obsessive detail sticklers. Not always, but it can mean disastrous other issues lurk...

Notion wasn't really known for its speed. The spinning wheel was always there to greet you whenever you opened a page. They also chose to go with an Electron web wrapper instead of a native app, since this made for easier cross-platform development. As expected, the experience was worse on mobile.

The friction introduced would have been enough for many to quit the app and leave a barrage of 1 stars. However, like Comrade Dyatlov from Chernobyl, most of us didn't make much of it and thought it was not great but it wasn't terrible either. And besides, nothing else came close in terms of features and ease of use. Our collective meltdown was inevitable.

A refreshing take on notes

While waiting for a Notion database to load, I did what anyone would - browsed Twitter for the latest memes on Notion's speed. Quite a few responses mentioned how they'd switched to this new kid on the block called Craft, and they weren't looking back. Intrigued, I went about exploring Craft.

A few months ago, Tiago outlined what he wanted in the ideal note-taking app. This, I'm sure, is the ideal version for most people:

And here we have Craft's marketing copy:

I was impressed. I tried it on my Mac and iPad and it was beautiful. Being native-first, the iPad experience was particularly great and seamless. It also utilised the capabilities of the Apple Pencil, while Notion seemed to ignore its existence.

The Craft team had an interesting response to all the talk about Notion:

While we have seen a lot of Notion users switching to Craft, we're not trying to be an alternative to Notion.

As a matter of fact, they tried to distinguish themselves on launch by stating they are as comparable to Notion as WhatsApp is to Slack (personally, I feel Telegram and Slack is a better analogy). This is true since Notion markets itself as a tool for work as opposed to Craft as a tool for thought.

But I wasn’t a huge fan of the wall styled homescreen and it lacked databases - a key feature for any Notion user. Even though the experience was great, I dismissed Craft as it was limited to the Apple ecosystem - I’m a green bubble Google Pixel user (not for long). And back to Notion I went.

Crafting foundations over features

A few weeks later, I came across their post on data ownership and accessibility, and I was fascinated. This excerpt hit home on the state of Notion:

Working on Data Ownership and Accessibility is what we call foundational work - meaning it’s a lot of work in exchange for small visibility - and users often don’t notice this until the point they (desperately) need it - which might be months/years after their initial engagement. It rarely drives direct signups or revenue. At the same time, the core principles impact every bit of product development - meaning it’s extremely hard (often impossible) to backward implement it. Startups will typically ignore this question, as their focus is on shipping features - and if the product becomes a success, there is little to no incentive to change the underlying principles and take on new constraints of this magnitude.

The agonising slowness of Notion was overlooked because of the features it provided. As more people jumped on the bandwagon because of the features, it kicked off a vicious cycle where Notion was incentivised to push more features while the foundations took a back seat.

Craft touts its speed wherever possible in its marketing copy:

users often don’t notice this until the point they (desperately) need it

Notion had a worldwide outage lasting a few hours, which is a relatively long time, considering that Notion marketed itself as the knowledge repository for teams. Twitter was ablaze with calls to introduce offline support to lessen the impact of future outages. There wasn't much progress to this end and the outages continued. The desperation then lead to disbelief as to why it was taking so long to implement something as simple as offline support.

it’s extremely hard (often impossible) to backward implement it

An important point is how initial choices impact future development. In Excel Never Dies, Packy McCormick writes how Lotus was the industry leader in spreadsheets by a huge margin. Microsoft's spreadsheet didn't take off, not until they rebuilt from scratch with a fundamental change - instead of recalculating every cell each time a cell changed, it recalculated only the affected cells. That gave it a huge speed and performance advantage.

Things like making stuff faster or accessible offline look simple at face value, but incorporating those features might involve rewriting the whole codebase. Uber tried to rebuild its app from scratch and it was almost catastrophic for the unicorn. From the Uber Engineering blog:

In early 2017, Uber made the decision to rewrite our driver app. This is the sort of decision that Joel Spolsky, the CEO of StackOverflow, once called “the single worst strategic mistake that any software company can make.” Rewrites are incredibly risky, resource-intensive, and take a long time to deliver a tangible benefit for users. For this particular rewrite, hundreds of engineers contributed in some capacity, not to mention designers, product managers, data scientists, operations, legal, and marketing. In practice, our rewrite took a year and a half to implement and roll out globally.

Another sore point for Notion is privacy. Something as private as your second brain should most certainly be yours and yours only - in terms of data ownership, encryption, accessibility at all times and non-proprietary file formats. People were spooked by the scenario that Notion employees may read your personal notes, as it almost did with Evernote.

Apart from accessibility and security, data ownership is particularly relevant to corporate customers.  Data shouldn't be held hostage by abusing lock-in advantages, an extreme case seen is what HP did to the Pentagon.

Craft tries to address privacy concerns with the ability to store on external locations instead of their servers. For data ownership, they have committed to open source converters for parsing their data format and multiple export options. Moreover, they state their drive to minimise friction in note-taking extends to reducing friction while migrating to competing products.

However, with all that talk on foundations, it was disappointing to see that end-to-end encryption and API integrations were to be implemented at a later date. But the folks over at Craft seem to be mission-driven as evidenced by their thoughtful post and I would wager it coming sooner than later. Though I'm not using it because of its unavailability on Android, I religiously track their changelogs to see if they actually follow through on their promises (spoiler alert - they mostly do).

Reflecting on Roam

At various points, I found myself wanting to tag things because they didn't just fit in a single category. Evernote thrived on tags while Notion ignores the concept entirely. While fiddling with workarounds for tags, I came across this thoughtful note by Andy Matuschak suggesting a different approach.

Let structure emerge organically. When it’s imposed from the start, you prematurely constrain what may emerge and artificially compress the nuanced relationships between ideas.

Enter Roam. The closest thing to the Wikipedia of my thoughts that I always dreamed of having.

Almost single-handedly, Roam popularised the concept of bidirectional links to such an extent that every new tool for thought now has it as a core feature. Everyone who has spent a little time with Roam has jumped onto the bandwagon of networked thought instantly.

I developed a system where I would write all my thoughts and link them in Roam. When I had arrived at something bigger and concrete, I would make the final structured version in Notion. It was a good workflow, but only in principle.

Like Notion, Roam had the same issues with load times and the mobile experience was quite bad. The Roam Cult is abound with workarounds that seemingly solve these issues, but it does not bode well for a service whose premise is to capture your thoughts. Lightbulb moments don’t always wait to accommodate load times. The team’s response wasn’t exactly confidence-inspiring:

It’s been more than a year since that tweet and there hasn’t been a considerable improvement in the speed and the mobile experience.

In the same article, Craig provides the iPhone camera as an example of good software providing a seamless experience with great hardware. In a world where an M1 laptop can render huge video files at a fraction of the time it takes for full-fledged desktop monsters, I find myself staring at Roam’s ancient astrolabe logo while it loads, trying hard to remember what I wanted to note in the first place.

The best camera is the one you have with you and more often than not, it’s your phone. Tools for thought cannot be limited to the desktop - one has to be able to note stuff down anytime and anywhere. There is a strong case for native-first apps with offline access and Craft also demonstrates wonderfully how you don’t need to limit the feature set on the phone. Mobile devices can no longer be treated as second-class citizens and the hardware is more than capable. The real work is in designing features to translate well on the smaller screen.

Roam also has the same issue with privacy as Notion; one will be uncomfortable if the privacy of one’s thoughts is based on the company’s pinky swear that they will never read them. There will always be a few bad actors and end-to-end encryption seems to be the best way to avoid such a scenario. Obsidian and some other Roam alternatives provide the same and in fact, is their USP.

A newer tool for thought called Reflect popped on my radar and it aims to have end-to-end encryption and APIs right off the bat:

It looks like a polished version of Roam and was snappy fast. And since everything is end-to-end encrypted, I could safely write my analyses on which NFT was really going to moon.

While launching what seemed like the billionth productivity app on Product Hunt the maker started off with this - there are never enough personal productivity apps. We will be seeing more apps that do one thing exceptionally well and focus entirely on it. So it only makes sense to have an API rather than try to do everything at once. This is particularly true for the more mature market of team productivity tools, which is why Notion got their API out on priority.

In Tiago’s original illustration for building a second brain, an important reason for Evernote being at the centre of it all was because it had integrations with multiple other apps. Integrations are what really hold it all together. The possibilities with integrations are endless, limited only by the creativity of the user.

The PAINful takeaway

A pattern seems to appear in what the newer “tools for thought“ have to offer that the incumbents could not, due to the difficulty in backward implementation:

  • Privacy and data ownership
  • Accessibility offline
  • Integrations and APIs
  • Native-first and fast

These PAIN foundations are a pain to consider as they provide a considerable cushion to the prevailing competition. It may seem counterintuitive, as evident from Roam's philosophy, to focus on these principles instead of building features for the initial product. But building features with these principles in mind makes for easier development in the long run. In the long term, this also introduces a stickiness factor. Few would want to switch to the shiny new thing - not because the data is held hostage, but that there is seldom a strong case to give up a well-oiled system with solid foundations and integrations.

I'm not trying to be a person that dunks on others, there's Scott Galloway for that. Notion and Roam are pioneers that inspired a new generation of knowledge tools. I certainly hope they can fix their problems soon - my second brain currently depends on it. It remains to be seen if their early choices have set them back enough for competitors to take their place.

Be that as it may, Craft and Reflect have a refreshing approach which I hope gains traction and we will likely see more of such forward-thinking foundations.


Follow the discussion on Twitter.

I write and make stuff sometimes. You can subscribe to updates by email (~once a month or fewer).


Spare a penny for my thoughts, both metaphorically (feedback) and literally (buy me a coffee).