Tiago has made significant contributions, helping people manage their notes, but his recent AI overview was too shallow for me.
Tiago Forte
As someone interested in knowledge, understanding, and learning science, Tiago has been a high profile figure in my internet environment for years.
He created the course, Building a second brain, which I have taken. And has shared various frameworks and terms for concepts related to managing notes.
PARA which stands for Projects, Areas, Resources, and Archives, for organizing information.
CODE which stands for Capture, Organize, Distil, and Express, a framework to consume on the internet.
Concepts like progressive summarization to help distil ideas from sources.
The work he has done, has always been done with attention to detail. A focus on the nuance.
Recently, Tiago’s time has, correctly in my view, shifted towards his family. Spending less time in his notes and work.
In a video, just released as of writing this, he mentions using Notion AI (artificial intelligence) as a solution to his time constraint.
Putting aside the potential conflict of interest with the sponsorship of the video, and his personal and business investment with Notion, AI as the solution seems a bit odd.
Titling the video “Using Notion AI to Save 12+ Hours of Work”, is understandable. After all, titles are designed to be intriguing.
But pairing that with the early comment using Notion to “distil thoughts and insights“, makes me curious.
How does AI save time in the distillation of ideas, while maintaining understanding?
Idea refinement
During the video, Tiago uses the example of a book he took notes on, “$100 million offers“.
He listened to the book rather than reading it, which caught my attention.
In the past, highlighting multiple times was part of the process he calls, progressive summarization.
A past blog post of his covers 4 levels. You don’t need to follow the same steps or go to the same depths.
But the idea is to progressively go through notes and ideas, to refine and distil your thinking about the source.
Instead the notes are taken on his phone using the Notion app.
The questions that come to mind are things like:
How does he decide what to write down?
Does he do a second or third pass over those notes?
What does he write down, quotes or summary thoughts?
Although the video is about AI summary, the quality of the notes being taken would surely impact the summary he gets.
Moving forwards in the video, Tiago brings up something we probably are all too familiar with.
This is where AI comes in to save the day. AI to be used for refinement.
A reminder, refinement is where you “need to add structure, you need to change how they are presented“.
And now, so Tiago claims, with AI, we can distil and/or refine, all of those notes to bring back the context with the push of a button.
What AI can do
What AI has done, and is doing to all areas of the tech industry is incredible. It is truly a revolutionary tool.
But can it do what Tiago is claiming?
The first example goes through the summary option in Notion. A brief 3-4 sentence summary of a 3000+ word note, which is created in seconds.
You would expect all of the main points from the notes to be included.
You would probably like to see some overarching themes or concepts.
And you would likely want to have the most pertinent information shown that is relevant to you.
On the surface, that is what AI seems to do.
However, this summary would be different from a quick internet search from an article or video, it is from his notes after all.
At least that is what I would hope, happens.
Otherwise he could just do an internet search.
The second example searches for action items.
As an Obsidian user, this seems easy as that can be done through search, queries, or plugins.
AI is doing a search which is nothing new.
This is where Notion and Obsidian differ with their feature set, so AI isn’t saving any time here for those that use the tools that already do the job.
Obsidian is one of many.
The original claim of AI saving 12+ hours is still resting in the summary prompt.
Language translation is the third example used.
Moving past the sample size of users that will use this, and that AI doesn’t speak every language, with potential limitations of training data in the future as well.
I am not sure AI is saving time, but doing a similar job to a translation tool.
More pertinent to this conversation, translation in my mind isn’t a form of distillation.
You could argue it is a form of refinement, as the notes are being restructured, but it is for the benefit of others.
That is where the line between personal refinement, and product creation starts to blur.
This is where I struggle to grasp Tiago’s argument for AI.
Saving time to me, means getting a similar result quicker.
But is the AI result quicker while maintaining the quality. Or is it doing something else?
AI compression
What Notion AI is doing, is compressing the notes. Converging the notes together to generate a summary.
From my experience, and by looking at other examples, AI summaries tend to do just that, summarize.
In Taigo’s words, “If we compress a note too much … we make a summary that is too brief, we lose the context and it loses all meaning.”
So either Tiago is shifting his view, and that a brief summary is more beneficial than he originally wrote about.
He still holds his original view but, AI compression is ok for some reason.
Or, like many of us, he is excited by the possibilities and still navigating their use cases.
Which is my personal irritation with the video.
It was too shallow.
Shallow explanation
The compression of AI writing in the summary is not as much of an issue when you use a prompt. A custom prompt.
Or multiple prompts.
Something Notion AI does afford for.
But Tiago doesn’t discuss that.
However, if you use prompts, it takes time to make them, or takes time to dig deeper with them.
So AI may be reallocating time to another task.
Then, there is no way to be sure the AI has picked up all of the things you may want, unless you check it.
So is Tiago suggesting to blindly trust AI. I don’t think so.
But that is how I interpreted his use of the tool with the note summary.
Which could mean added time to confirm the quality of the notes.
This is where I think a more nuanced explanation, one I am more accustomed to, would help.
AI is a powerful tool, but I think the time save in one area, is moved elsewhere.
And like with most products, convenience comes with a cost.
In this case, quality and depth.
