Preparing for the AI revolution: How obsidian could turn into one of the most valuable source of info for the personal assistants of the future

The AI revolution is here. With tools such as ChatGPT, we get a glimpse of what the future may hold and start to better understand what large language models will eventually be able to do for us.

One key aspect of large language models is the data they are trained on. This effectively is what defines what a model will be able to know, and consequently, do. It as such becomes critical to feed the network with the correct information, as this alone can unlock features and capabilities impossible otherwise. Until now, all large language models available to the public have been trained on mostly non-personal data, such as subsets of the internet. This however will not cut it if we strive for powerful and tailored personal assistants.

Tailoring a large language model to an individual is not a mystery, all that is really needed is access to information relevant to the individual in question. Some of this data is (theoretically) easy to retrieve if present. Examples would include:

  • Emails
  • Online calendar/schedule
  • Browsing history
  • Programing time (a number of solutions already exist which can be integrated into most IDEs)
  • Contacts/social network (friends + activity on social media)
  • List of todos
  • Finances (sensitive, but a number of large banks already provide means of allowing third parties to access such information)
  • Health data (smart watches, connected scales, sleep trackers)
  • And much more (smart homes, phone GPS, etc…)

Those alone already allow for getting a solid grasp of an individual and should be enough to achieve a high level of customisation. It does however exclude a significant aspect of an individual; their “knowledge”.

Why is Obsidian so special?

While it is possible to partially derive that from one’s browsing activity, all of the above sources do not allow for properly grasping the extent of one’s knowledge, which I believe to be fundamental to properly understanding and assisting someone. Allowing your personal assistant to better grasp your knowledge base could enable the following:

  • Understand which areas you are knowledgeable in, and which areas represent weaker points: This could in turn be used to determine the level of details and complexity that should be considered in responses to prompts. A good teacher isn’t one that just knows everything, but also understands what you don’t know.
  • Detects areas of misinformation: Being able to access a knowledge repository enables for detecting errors/misunderstandings, which could in turn be reported and notified to the individual
  • etc…

These personal data repositories we have all been creating could be significantly enhanced further through the inclusion of effective time tracking. Being able to properly log the amount of time spent on various notes would provide a tremendous amount of insight into the interest, capabilities, or workflow of a person. This user activity could allow for a new form of metaphorical linking too, which I would refer to as “activity links” or “desire links”. Akin to “desire paths” (see the figure below), those links would be created by the natural evolution of the user in his pkm environment.

To clarify, I am not talking about necessarily creating actual links, but rather capturing this natural relationship between notes, to be eventually leveraged by large-language models and users (a dashboard of one’s activity in obsidian would provide valuable insight into how time is split and allocated to various aspects of work/life)

A plea for help!
I have previously made a post on this topic on this forum, describing how logging the opening, putting in focus, and closing of notes could be coupled with the note’s path and tags to effectively track this data with no extra effort. The true value of such data only increases over time, and every hour spent without it is fundamental information missed-out on. If anyone with the right skillset could take on this challenge, this plugin could allow for starting to harvest this invaluable information, while remaining in the obsidian ethos of a local-first, and privacy-centric approach. Even better, building this directly in obsidian to ensure a reliable and accurate log could prove even better.

I am excited to be on for the ride with you all and look forward to hearing your thoughts on the topic. I would like to then leave you all with a question to further brainstorm on this topic: as we stand at the edge of a new revolution, what else could be done with obsidian to further anticipate and prepare ourselves to truly unlock AI-assisted knowledge tools?

To conclude, as we enter this new era, obsidian happens to find itself at an exceptional crossroad, and stands with the potential to radically change the capabilities of the upcoming tools of the AI revolution. Obsidian marketed itself as a second brain, and might end up living up to it more than anyone ever thought possible!

6 Likes

I think we will need to pay careful attention for what we will use AI-like tools.
According to me they need to “remain just tools”. Not substitutes of our higher order mental processing activities.

I think using automation for searching, linking, classification, summarization, rewriting, content creation and so on, the most valuable parts of our knowledge workflows, is dangerous for our brains.
It can be ok for lower-level tasks, instead.

It’s like making training if we pretend to practice a sport. Training is hard and tiresome, but it is the training that develops our skill. Choosing what to read, reading, thinking how and what to link, how we summarize, the our act of writing and so on.
If we demand our thinking processes to an AI, we lost the ability to effective apply them with the time.

I will not use this tools. For me the friction caused by doing my knowledge activity manually is the thing that make me think, learn and create.

2 Likes

While I absolutely see the merit in this application of training an AI using the concepts I have placed into my PKM space, I don’t really like the idea of it when applied to myself. Maybe when I’m long gone from this world, someone might be able to feed in all my journal entries, my PKM writing and my photo journals into something that could effectively capture my essence. But I think, as @andy76 wrote, there’s a real merit to developing knowledge on your own, and grappling with the information before adding it to a knowledge database.

I think having an AI as a means to suggest places where gaps may exist could be beneficial, but I do not want an AI telling guiding me how to think for myself, if that makes any sense. In a way, a PKM is how I choose to display my learning autonomy and what I consider valuable, interesting, and useful information. To then designate that synthesis process to an AI would be to throw away the reason why I maintain my PKM systems in the first place. If an AI goes and links things in my stead, then am I really making the connection, or am I falling into a weird variation on the Collector’s Fallacy?

To go even more generally, I think that the practice of applying AI as a stand-in for human creativity is the big issue. I won’t go on my soap-box about why I dislike the practice of AI art and music, but I will say that the value in PKM systems is the practice itself, and not the result. PKM systems inspire creativity and imagination such that we as people can be more well-rounded, well-minded individuals. To add a secondary party, even one that has learned from one’s own writing and thoughts, steals the creative practice of generating anything, whether it be knowledge, art, voice performances, etc.

I hope I haven’t misunderstood the point of this thread. It’s absolutely a thread that encourages healthy debate about how much each of us cares to let AI do things for us. For me though, I don’t care to have AI to a majority of the heavy lifting, as I’ll never be stronger if I don’t lift the weights.

TL;DR it might work for you, but I don’t think it’ll work for me and how I use my PKM systems. And maybe my stance will change, who knows.

2 Likes

@andy76 and @photvedt cheers for the interesting replies! It is interesting to hear a different perspective on this topic. I have to say that I would mostly agree with both of your statements. I am afraid I might have slightly been misunderstood though as to what I envision the role of these future tools: I do not believe that Ai would replace higher-order thinking or reasoning.

AI is “simply” a very powerful search engine, capable of not only retrieving data but also synthesising vast amounts of it (something we as humans are simply not quite capable of doing) to extract meaning. This however is not new at all. Data mining is the direct result and demonstration of that; when the sheer amount of data gets too big, too complex, and too varied, letting computers take care of the heavy lifting becomes necessary. This also allows humans to be more effective at higher-order thinking. By being provided with statistics derived using a computer, we are provided with information that would not be available to us otherwise.

And this is the key aspect I was trying to emphasise. Those new upcoming tools I believe will enable us to gain unprecedented insights into areas which were impossible until then. A note repository for example, could be analysed for word frequency, connect count, tag frequency etc… All of those while insightful, only paint a very poor image of the content of the dataset. AI changes all of that by enabling us on gaining insight into the actual concepts, and ideas contained in the notes. By being able to better understand the extent of knowledge, the level of understanding, and the patterns present, we are not working less, or putting less effort, but being given the opportunity to account better for our cognitive context in our thought process.

To go back to the training analogy, it would be like comparing a “regular” person going to the gym or for a run to a professional athlete, undergoing the same training, but with a personal coach monitoring progress, and providing feedback and insight to make the most of the training. Though again the analogy here partially fails to highlight my thoughts, as a coach actively guides training, while the AI in our case is only here to provide insight (akin to a very fancy smartwatch for your knowledge instead of your health). It still needs to be queried for information. By getting access to this novel information, you wouldn’t be lifting less weight, but being given the opportunity to lift larger and more varied ones.

I hope this clarifies my perspective :slight_smile:

1 Like

I share your vision, and I think it is the future.

I’ve been actively trying to find something to get a quick POC.

I’ll keep it simple: markdown files → NLP model → Chatbot (Personal ChatGPT). Let me know if you are aware of any tool that can take markdowns and build an assistance bot that I can query and propose extra content to augment my notes. In that case, I’ll use it immediately, and I think that should be the future of obsidian.

And many will resist this move, but progress is merciless, adapt or perish. It is the way forward. And obsidian is ideally suited to implement such a vision.

You could take a look at GPT_index. It’s a library made to build an index from a large not erepository to then be fed to LLMs like GPT3. This allows you to then use this as a source of info when performing queries (and it can even cite it’s sources!). I believe someone was already working in turning this into a plugin for obsidian

Nice; if you can find any other pointers or resources, please let me know.

What do you think about this approach?

It seems to be exactly what we are after. It just needs to get well integrated at this point.

1 Like

I think this article that I read a couple days ago is very relevant to this.

It’s very exciting to pair these personalised PKM systems with Large Language Models.

Anyway very interesting thread!

3 Likes

I am going to give a bit of a contrarian counterpoint here and invite everybody to think what we stand to loose with these AI tools.

It is disingenuous to think that innovations come with only benefits, that every piece of tech is a net addition to out lives. My grandma is better at mental arithmetic than myself because she dealt with cash change her whole life. We have lost our orienting skills since the mass adoption of google maps. College students have middle-school level grammar skills due to the ubiquitous use of tools like Grammarly.
We are weaker and fatter since most of us don’t do any real physical labor and we eat mass-produced artificially enhanced food. The list is long.

Now AI, that’s scary. Do I want to outsource my intelligence?

One of my PhD advisor once told me that writing (the paper) is where the real understanding happens (and I came to agree with him). Organizing is understanding.
If writing, organizing, linking, resurfacing is done by the AI, which skills in my brain are anthropizing and do I really want that?

I am not against AI, (how could I be? it’s my job!), but I do not embrace it. We should think of the second order effects of these technologies.
I am afraid we may end up with a small fraction of the population net smarter, the people who successfully use these tools as a “cognitive lever”, and large fraction of population net dumber, people who need the AI tools as a “cognitive crutch” otherwise they would not function.

It is a know story at this point that Silicon Valley parents don’t let their children use social media. Will 5-10 years from now, read the same thing about AI assistants?

6 Likes

I get your point, but someone else will if you don’t embrace it; just look at ChatGPT growth.

And the tool proposed in this thread will not do the linking etc., but it would be a co-pilot for your knowledge. AI, thus far, can’t link concepts to the context the way we do. Still, it can accelerate the retrieval, augment what is missing, find key insights and even spot flaws and weaknesses in our knowledge map, and this is something that can’t be done on paper.

I would embrace AI fully until it can ultimately outperform humans, then I’ll just intellect for fun, just like humans ride horses today.

Never bet against progress; this argument has been made with every major human innovation.

1 Like