I have adhd myself and I am designing assistive technology for myself.

My software is basically "self spyware", a (mouse/key/browser history/etc) logger for linux. In a way like an open source (though not yet) microsoft recall but even more extensive and with better search and no AI. The idea is to:

1. Be able to put in small notes easily and quickly without having to think of name, tags, etc.

2. Be able to recall what I did, when I did it and what else I did around that time, with _time_ being the main link. So you don't have to semantically link things together like in e.g. obsidian. You can, but using locality of reference should already provide good results.

I wrote some code already for the keylogging parts. Where I got stuck is when thinking about/designing the parts that require cryptography. When I have more free time I definitely want to get back to it.

What I found I need most due to my adhd:

1. A way to do anything "in the moment". Let's say I'm about to procrastinate on my phone but know I have to write an email. Then I must have the ability to do that very quickly right from my phone. If mentally the task of "turning on my laptop" and doing it there seems like too much, I won't do it. So the "preparation friction" to do any task must be as small as possible. This is somewhat a difficult problem but technologically it requires everything to be cross platform and easily accessible.

2. A way to very quickly confirm what I already did before or confirm that I already did something. I have to constantly confirm it because I can't remember if I did. This part is what my software tries to address.

As ADHD dev that thought about something similar (but never worked on it) I'd love to see the code. Please make a post when you intend to release it as opensource, it seems super interesting

This is unsurprising. They do this a lot with especially short usernames. I recently saw an interview with a German youtuber (@dima) where they talked about how youtube gave them this channel. In that case the channel still contained private videos and information of the previous owner, which I find somewhat shocking.

Easy fix: use youtube without a channel, e.g. using newpipe or yt-dlp.

I used it just with web components and pure html/js. Honestly don't even have a need for any framework with it, it's a great ssg like that already.

Interesting. Do you just build web components instead of React/etc, and add the client:load and Astro gives you the static (fast) initial page load, and lazy loading of web components?

> "illegal"

If legality is where one draws the line, faith in the united states would be considered long dead.

I prefer kotlins trailing lambdas, which do mostly the same. But the way they are written allows code to almost look like JSON/some declarative DSL.

But use is nicer to avoid callback hell and all the indentations/scoping.

That's one thing. Another would be introducing homomorphic encryption in order for companies and people using their models to stay compliant and private. I can't believe it's such an under-researched area in AI.

That is the only implementation I can think of that might make me trust a third party with confidential data. Of course these massive transformers are already insanely computer intensive and adding FHE would make that orders of magnitude worse.

I agree but that's why it needs to be researched more. Some changes in architecture may be able to address some performance problems. It could lead to algorithmic optimizations or even specialized hardware for this.

100% on board. This would be a paradigm shift for cloud services.

(And take away a, for many significant, source of income - data they can sell, train on, etc - So I’m afraid the incentive to research an implement it will be lacking)

I agree. And you probably neither need Omarchy, CachyOS, etc. Choose boring and get peace of mind.

CachyOS is basically arch on easy mode. I used to spend countless hours tinkering with arch but then I got older and don’t have much time. Plus there are helpful meta packages for gaming that work great out of the box, which for a gamer like me sans Windows is pretty awesome.

I've tinkered at most an hour on my arch install and it's just been running smoothly. The installer these days is easy to go through as well. It's the same for bash, very little customization, then it runs smoothly for years.

I'm not saying CachyOS is bad, just that it is in my opinion another layer of complexity that may change/deprecate/etc.

That’s fair, the more steps you take away from the source distribution, the more variances become a potential for trouble. Also it’s been over a decade since I used arch, it’s probably a whole lot better now.

>I used to spend countless hours tinkering with arch but then I got older and don’t have much time.

Have your lost your old configs that you worked hard on? That's a shame if so. I love moving my configs I've worked hard on to new machines and instantly getting up and running in a now-familiar environment. It saves so much time and effort.

> Have your lost your old configs that you worked hard on?

I mean… it was so long ago. Actually it’s more than 15 years ago, on used hardware no less. Needless to say, I don’t even have remotely the same hardware anymore - I’m on AM5 / DDR5 right now.

Only tangentially related but maybe someone here can help me.

I have a server which has many peripherals and multiple GPUs. Now, I can use vfio and vfio-pcio to memory map and access their registers in user space. My question is, how could I start with kernel driver development? And I specifically mean the dev setup.

Would it be a good idea to use vfio with or without a vm to write and test drivers? How to best debug, reload and test changing some code of an existing driver?

Sine you are mounting and not syncing the files, what happens when you edit a file offline? And what if on another offline device the file is also edited?

Fair question. Conflicts happen, which I'm fine with.

Realistically speaking, most files I have in my cloud are read-only. The most common file that I read-write on multiple devices is my keepass file, which supports conflict resolution (by merging changes) in clients.

Also used to happen when I tried editing some markdown notes using obsidian on PC, and then using text editor (or maybe obsidian again?) on android, but I eventually sort of gave up on that use-case. Editing my notes from my phone is sort of inconvenient anyway, so I mostly just create new short notes that I can later edit into some larger note, but honestly can't remember the last time this happened.

But yes, if not careful, you could run into your laptop overwriting the file when it comes online. In my case, it doesn't really happen, and when it does, Nextcloud will have the "overwritten version" saved, so I can always check what was overwritten and manually merge the changes.

P.S. If anyone wants to set this up, here's my nixos config for the service, feel free to comment on it:

  # don't forget to run `rclone config` beforehand
  # to create the "nextcloud:" remote
  # some day I may do this declaratively, but not today
  systemd.services.rclone-nextcloud-mount = {
    # Ensure the service starts after the network is up
    wantedBy = [ "multi-user.target" ];
    after = [ "network-online.target" ];
    requires = [ "network-online.target" ];

    # Service configuration
    serviceConfig = let
      ncDir = "/home/username/nextcloud";
      mountOptions = "--vfs-cache-mode full --dir-cache-time 1w --vfs-cache-max-age 1w";
    in {
      Type = "simple";
      ExecStartPre = "/run/current-system/sw/bin/mkdir -p ${ncDir}"; # Creates folder if didn't exist
      ExecStart = "${pkgs.rclone}/bin/rclone mount ${mountOptions} nextcloud: ${ncDir}"; # Mounts
      ExecStop = "/run/current-system/sw/bin/fusermount -u ${ncDir}"; # Dismounts
      Restart = "on-failure";
      RestartSec = "10s";
      User = "username";
      Group = "users";
      Environment = [ "PATH=/run/wrappers/bin/:$PATH" ];
    };
  };

However, why is that even surprising? Tailwind is essentially a frontend css stylesheet. What business could there possibly be around that?

I understand, they have UI kits, books, etc. but just fundamentally, it was never going to be easy to monetize around that long term, with or without AI.

Tailwind also has a compiler of sorts (so you only include in the bundle the exact styles you need) and a bunch of tooling built around it. In an alternate universe it could have been a fully paid enterprise tool, but then it might not have caught on.

The comment you are responding to said their revenue is down 80%. So they did monetize training and services, and I don't see how that would have been a problem long term if AI didn't come along and make all of that unnecessary.

Yes. The point I was trying to make was that after the initial hype disappears, sales in those categories would probably taper off regardless. But it is purely my opinion.