[1] https://www.henrikkarlsson.xyz/p/good-ideas
At least read Gatto’s work in education because he knows what he’s talking about as an actual teacher who has put more students through schooling than anyone writing these articles. His work contradicts some of this because he studied very non-exceptional childhoods of exceptional people. He has a better answer to how to think for oneself as well.
To actually get to the bottom of things: I think most normal folks are concerned more about getting by and making decent money in “the age of AI” than they are about being brilliant whizkid prodigies coming up with original ideas. A lot of those end up being poor anyway. But the desire to live a quality life is a more universal thing. No amount of “mind training” will help here. Just steer clear of paths that AI can dominate (they’re expanding), and failing that, use it to your advantage as best as you can.
Right now working with one’s hands seems to be in vogue because it’s one of those things that people are unaware of is actually dominated by robotics in the industrial/manufacturing sectors, so the ignorance there can probably get people through some hard times. But eventually even that will be shown for what it is and we’ll have to find better ways to spend our time.
I don't know how many times I've read a variation on this. It took me a very long time, but now I pretty much made my peace with that: I use Emacs (for certain things), I use VS Code (with Emacs bindings), I use Apple Notes.. I don't find that it's possible or reasonable anymore, the desire to be "pure" and use only ONE tool to rule them all. The same for messaging apps, chatbots, etc.. I now embrace extreme diversity.
The trick is to stick to as few packages and as little configuration as possible. And when opting to install a package, sticking to something popular and well maintained. This leads to a small and robust setup with little churn. Most built-in packages work out of the box. Most defaults make a lot of sense. Emacs is really tidy these days compared to where it was one decade ago. Package management has been key facilitating this.
Personally, I use major packages like AUCTeX, Org, Magit, or gptel with little to no customization and I avoid installing lesser known packages that build on top of them as I have found this to be a major source of fragility. You can get a lot of functionality from a boring 50 LOC .emacs/init.el that consists of a few straightforward use-package directives.
On the flip side, my experience with Emacs has been quite different. You don't need a ton of plugins to get the most out of it; I've been using the same configuration of under 200 lines for the past six years without encountering any breaking changes. I rely on Magit, Org-mode, Org-roam, and Org-agenda every single day.
That said, using Emacs does require some commitment to reading the documentation. While I agree that it has some outdated defaults, you only need to make those adjustments once.
I think this SO question demonstrates this well. The question is how to select a window quickly. You can install umpteen different packages and have several black-box soltions, if you want. But it really can be as simple as this if it fits your need:
``` ;; Select the 3rd window in the `window-list' (select-window (nth 2 (window-list))) ```
Emacs almost always allows you to find a solution in-between.
The more deeply you think, you train your brain harder, but also improve the utility of the AI systems themselves because you can prompt better.
I doubt it. Obsidian is not open source, and the core is maintained by a small group of people, rather than a community. What happens when the company dies?
That said, I am willing to have more faith in Obsidian, than many other things since they are not [VC funded](https://stephango.com/vcware)
Also stuff like Bases[0] might be the thing that entrenches Obsidian even further as an IDE for knowledge work (more or less).
If not, someone might make an api-compatible oss clone, because lots of the value is in the myriad of plugins.
Obsidian's ace however is it's great wyiwyg text editor if you ask me, enabling friction-free writing.
I would prefer to buy from those commercial players that have a clause in their license saying upon sunsetting the commercial offering or closure of the company the source code becomes open source. In the absence of such a clause, I prefer open source solutions.
[RMS was right saying "Free as in 'freedom' is not about payment." There can be paid-for open source software, and there can be free-of-charge commercial software, but the freedom to edit and recompile is the most important aspect of "being free".]
But yeah, your vision is the next best thing I like to day dream about sadly.
I had the same workflow with Markdown files if you prefer the closed source Markdown editors. At the time, I was using VIM for editing and viewing the Markdown files.
There is a lot to be said for the value of simplicity, if one of the goals is portability.
> The app will eventually become obsolete. It’s the plain text files I create that are designed to last.
The context switch between digital and analog is compelling. There is something satisfying about throwing a piece of physical paper away after pushing your commits.
I tend to hoard information, so having a medium that is highly constrained keeps me honest with what I hold onto over time. Not being able to do full text search over my notes means I prefer to keep no more than ~one legal pad active at any given time.
I also tend to get distracted with shiny technology tools. I can take my notes anywhere. I don't need an internet connection. I can fold the piece of paper and store it in my wallet.
I think I got over that sometime while Clinton was in office ))
The goal to index your obsidian and also persist anything from llm chats, and browsing history.
To your own sqlite database so you can still own your data across providers.
I’m sure someone with much more time than me will win with a better version of it
I started my personal knowledge base in the days of Mac OS 8 and still use it. Classic Mac OS with it's spacial finder and 1-to-1 file and window mapping works so well for my brain I still use it today for things I want to save and refer to.
I use & subscribe to AIs, but my personal knowledge is kept in the system I can use like the back of my hand, which happened to be the Macintosh Finder for me.
But that's not why I mentioned ekg, the reason is that it does embedding out of the box, here's a quote from the repo on GitHub:
"There is support for attaching Large Language Model (LLM) “embeddings” to notes, for use in search and similarity search, via the llm package. This allows you to search based on semantics, as opposed to text matching. You can also use LLM chat in your notes, getting an LLM to respond to your notes based on a default prompt, or new prompts that you add."
These days I feel like you have lots of great options for note-taking in Emacs and you're not forced to use the org format unless you want to.
ekg repo: https://github.com/ahyatt/ekg
I add plaintext keywords to the notes I write myself, but sometimes the note app is truly the problem when search implementation/tokenization doesn't make sense.
With LLM-based AI, should one also store individual chats in personal knowledge system? Yeah, I believe that some my chats are quite full of relevant info, that can be used in the future.
Also what is the right general approach here - should I ask the same question several times (every time I need information) or should I just look up previous answer in my history? To be fair I dont store google results, I just search it again, but with chat the path to right answer is often more complex than spitting few words in google search input box.
"I went through a phase that I imagine many software developers go through at least once in their careers—a period of intense fascination with Emacs..."
What is this widespread overt fascination with the tools one use? I've always been aware of a tradeoff between why one uses a tool and any tendency to fascinate on the tool or process of using a tool. The use, the purpose, the goal outside and originally to want any tool has always been my preference. To see the forest when among trees, to see a purpose and path through. Not to navel gaze at the existence of tools in preference to using them beyond pontificating their existence. This then extends to shaking my head at the swelled population of developer influencers that over fascinate on tools, appearing to do that in preference of anything else.
This confuses a tool with knowledge itself. smh.
[1] https://publiccomment.blog/p/you-ll-never-think-alone-170518...
Maybe I'm missing the author's point, as it's early here, but I don't see how your own thoughts can possibly lack value because of AI. LLMs can only summarize the documents it was trained on, so it has no way to tell you what you're thinking (like why something is wrong). The value of AI is using RAG or semantic search to make your notes more useful to you. What the author's suggesting is outside the capabilities of current LLMs. By design, AI can only be used as an assistant.
I experimented with feeding my notes into an LLM model for RAG and was underwhelmed. The resulting output was repetitive, stilted, dry, and uninspiring. I wanted it to see if it find relationships between my ideas that I had not found on my own, but was disappointed . It did not provide me any new insights into my thinking. The style of what it did write was so foreign to my own style I found myself needing to read and re-interpret what it wrote back into my own ways of thinking that it was more busywork than help.
I'm all about automated solutions for things — but I find that my desires are typically for unrewarding physical activities. I still don't have a virtually costless robot butler, driver, farmer, chef, and maid to anticipate my needs around the house and home office; to transport my family and me around town safely and on time; to grow microplastic-free food hydroponically; to do my meal planning, shopping, cooking, cleaning, tidying, and laundry with effectively no intervention and supervision on my part. Why not?
Now that would really 10x my productivity.
Take travel for example. I have travel resources (packing lists, loyalty reward numbers, etc). I would presumably have a travel area, as it's something I'd like to be a semi-major part of my life long-term. But I'd also think whatever trips I'm currently planning are "projects". Does this mean 3 travel folders, in 3 different parents, so I have to go 3 places for one topic?
It all seemed so confusing.
GTD from the days of Palm PDAs still works fantastically if you just use it.
The key is just to choose a system that seems like it will work for you based on the published "way to use" it ... and then make it your own and stop looking.
You basically need a monogamous relationship with blinders on so you can actually become a deep expert in your own productivity tool and be able to rely on it and refer to it without thinking.
why are emacs users like this
> Call me Ishmael. Some years ago—never mind how long precisely—having little or no money in my purse, and nothing particular to interest me on shore, I thought I would sail about a little and see the watery part of the world
I believe the primary issue with nearly all note-taking tools is the lack of genuine encryption. Many claim to use end-to-end encryption, but I find this misleading. End-to-end encryption secures communication, not the data itself.
I wouldn't recommend a tool to anyone that doesn't encrypt the data itself to maintain private notes. Its like storing your passwords in plain text.
[1] https://forum.obsidian.md/t/local-file-encryption-of-obsidia...
They have a dedicated page that compares with Obsidian here: