It's about how filesystems as they are (and have been for decades) are proving to be powerful tools for LLMs/agents.
Production grade systems might be written by agents running on filesystem skills, but the production systems themselves will run on consistent and scalable data structures.
Meanwhile the UI of AI agents will almost certainly evolve away from desktop computers and toward audio/visual interfaces. An agent might get more context from a zoom call with you, once tone and body language can be used to increase the bandwidth between you.
IMO it's insulting to the audience, it says your time and attention is not worthy of the author's own time and attention spent putting their own thoughts in their own words.
If you're going to do that at least mention it's LLM output or just give me your outline prompts. I don't care what your LLM has to say, I'm capable of prompting your outline in my own model myself if I feel like it.
It is easy to spot the compacted token distribution unique to each model, but search engines still seem to promote nonsense content. =3
"Bad Bot Problem - Computerphile"
https://www.youtube.com/watch?v=AjQNDCYL5Rg
"A Day in the Life of an Ensh*ttificator "
The problem I think with AI generated posts is that you feel like you can't trust the content once it's AI. It could be partly hallucinated, or misrepresented.
> That's not a technical argument. It's a values argument. And it's one that the filesystem, for all its age and simplicity, is uniquely positioned to serve. Not because it's the best technology. But because it's the one technology that already belongs to you.
That's a bit vague. Was the article was written without the aid of LLMs? Yes or no.
There are a lot of unique aspects of the writing in this post that LLMs don't typically generate on their own.
And there's not a "delve" or "tapestry" or even a bullet point to be found.
Also, accusations and complaints like this are off-topic and uninteresting.
We should be talking about filesystems here, not your gut instinct AI detector that has a sky-high false-positive rate.
I swear there needs to be some convention around throwing wild accusations at people you don't know based exclusively on vibes and with zero actual evidence.
It's not a website you go to — it's a little spirit that lives on your machine.
Not a chatbot. A tool that reads and writes files on your filesystem.
That's not a technical argument. It's a values argument."
Are you saying this post is a few edits away from becoming a New York Times bestseller?
But you're right, it did hit the front page, and that says more about my sensibilities not lining up with whoever is voting the article up.
The problem today is that we build specific, short-lived apps that lock data into formats only they can read. If you don't use universal formats, your system is fragile. We can still open JPEGs from 1995 because the files don't depend on the software used to make them. Using obscure or proprietary formats is just technical debt that will eventually kill your project. File or forget.
It’s proven several times over that it’s the correct approach. Abstractions (formerly Google photos, currently Immich) should just be built on top - but these proprietary databases are only for convenience.
For work, I’m having the same experience as the author and everything is just markdown and csv files for Claude Code (for research and document writing).
Ostensibly, things like MacOS Spotlight can bring real utility and value to the file system, and extended attributes through the sidecar indexing, etc. But Spotlight is infamous for its unreliability.
The other issue with file systems is simply that the user (potentially) has "direct access" to them, in that they can readily move files in and up and around whimsically. The "structure" is laid bare for them to potentially interfere with, or, such as the case with the extended attributes, drag a file to a USB fob, and then copy it back -- inadvertently removing those attributes.
And thats how we end up with everything being stuffed into a SQLite DB.
What are your thoughts on XMP sidecar files? I'm torn right now between digital negative + external metadata versus all-in-one image with mutable properties. Portability vs. Durability etc.
Relational is better. Hell, and kind of unique identifier would be nice. So many better ways to organize data stores.
All files are represented in a table with rows and columns. "Directories" simply have a special "directory = true" attribute in a row (simplified).
The hierarchy is for you, the human.
Like many file systems, NTFS also contains a log for recoverability/rollback purposes.
It's not quite relational but it doesn't make sense to be relational. Why would you need more than one 'table' to contain everything you need to know about a file? Microsoft experimented with WinFS, which wasn't a traditional file system (it was an MSSQL database with BLOB storage which sat ontop of a regular NTFS volume). Performance was bad and Skydrive replaced the need for it (in the view of MSFT).
The best compromise is what modern OSs have: a tree-like structure to store files but a database index on top for queries.
One thing directories solve: they're great grouping mechanisms. "All the Q3 stuff lives in this directory"
I bet we move towards a world where files are just UUIDs, then directory structures get created on demand, like tags.
It pulled back Plan 9, and I was shocked: this is exactly what we need today, as I'm convinced we need to think about minimizing agent permissions the exact same way companies do. Plan 9 was just too early.
The article gets some fundamentals completely wrong though: file systems are full graphs, not strict trees and are definitely not acyclic
The challenge is how to structure messy data as a filesystem the agent can use. That is a lot harder than querying a vector db for a semantic query.
The code bases we’ve been using agents in had been pruned and maintained over years, we’ve got principles like DRY that pushed us to put the answer in one place… implicitly building and maintaining that graph with all the actors in the system invested in maintaining this. This is not the case for messy data, so while I see the authors point and agree that a filesystem is a better structure for context over time, we haven’t supplanted search yet for non-code data.
I've been researching and building with a different paradigm, an inversion of the tool calling concept that creates contextual agents of limited scope, but pipelines of them, with the user in triplicate control of agent as author, operator of an application with a clear goal, and conversationally cooperating on a task with one or more agents.
I create agents that are inside open source software, making that application "intelligent", and the user has control to make the agent an expert in the type of work that human uses that software. Imagine a word processor that when used by a documentation author has multiple documentation agents that co-work with the author. While that same word processor when used by a, for example, romance novelist has similar agents but experts in a different literary / document goal. Then do this with spreadsheets, and project management software, and you get an intelligent office suite with amazing levels of user assistance.
In this structure, context/task specific knowledge is placed inside other software, providing complex processes to the user they can conversationally request and compose on the fly, use and save as a new agent for repeated use, or discard as something built for the moment. The agents are inside other software, with full knowledge of that application in addition to task knowledge related to why the user is using that software. It's a unified agent creation and use and chain-of-thought live editing environment, in context with what one is doing in other software.
I wrap the entire structure into a permission hierarchy that mirrors departments, projects, and project staff, creating an application suite structure more secure than this Filesystems approach, with substantially more user controls that do not expose the potential for malicious application. The agents are each for a specific purpose, which limits their reach and potential for damage. Being purpose built, the users (who are task focused, not developers) easily edit and enhance the agents they use because that is the job/career they already know and continue to do, just with agent help.
It's people realizing that you can use a coding agent for any cognitive work, and it's better since you own the file system while easily swapping the model or harness. I personally use a graph like format but organized like a simple text file, each node prefixed with [id] and inline referencing other nodes by [id], this works well with replace, diff, git and is navigable at larger scales without reading everything.
Instead of reading multi-meg data into memory to determine what to do, I used the file system and the program would store data related to the key in sub directories instead. The older people saw what I did and thought that was interesting. With development time factored in, doing it this way ended up being much faster and avoided memory issues that would have occurred.
So with AI, back to the old ways I guess :)
My life's data, including all the official stuff (bank statements, notary acts, statements made to the police [witness, etc.], insurance, property titels), all my coding projects, all the family pictures (not just the ones I took) and all the stuff I forgot, is in files, not in a dedicated DB. But these files are a definitely a database.
And because I don't want to deal with data corruption and even less want to deal with synching now corrupted data, many of my files contains, in their filename, a partial cryptographic checksum. E.g. "dsc239879879.jpg" becomes "dsc239789879-b3-6f338201b7.jpg" (meaning the Blake3 hash of that file has to begin with 6f338201b7 or the file is corrupted).
At any time, if I want to, I can import these in "real" dedicated DBs. For example I can pass my pictures as a read-only to "I'm Mich" (immich) and then query my pictures: "Find me all the pictures of Eliza" or "Find me all the pictures taken in 2016 on the french riviera".
But the real database of my all my life is and shall always be files on a filesystem.
With a "real" database, a backup can be as simple as a dump. With files backuping involve... Making sure you keep a proper version of all your files.
I'd say files are even more important than the filesystem: a backup on a BluRay disc or on an ext4-formatted SSD or on an exfat formatted SSD or on a tape... Doesn't matter: the files are the data.
A filesystem is the first "database" with these data: a crude one, with only simple queries. But a filesystem is definitely a database.
The main advantage of this very simple database is that as long as the data are accessible, you know your data is safe and can always use them to populate more advanced databases if needed.
Were Haiku mor mature/stable would have been a nice fit for the OS for the LLM/Ai personal use cases.
[1] https://arstechnica.com/information-technology/2018/07/the-b...