Enshittification aside, any new technologies you find yourself relying on/using regularly?
This can be anywhere from hardware or software/apps.
I recently started up a CalDav/CardDav service (radicale, think like your own private Google Calendar and Google Tasks that can also be synced on multiple devices) on a VPS. One step closer to degoogling myself.
For something that’s sure to be enshitified, I use Perplexity regularly, especially since Paypal gave me a free year of its Pro plan. I’m finding it considerably more effective than traditional web search when I’m looking for something specific, though to be clear, I’m looking for an existing web page rather than the output of the LLM. It’s also pretty good at providing the exact command line incantation for some one-off task and producing short code samples for some API I’m probably never going to use again. Sometimes I pay Anthropic for the latter.
Some other stuff:
- Cleverkeys, an open source Android keyboard with open source swipe typing (no Google library dependency).
- Rio terminal - GPU accelerated, written in Rust.
- Lemmy - you may have heard of it.
Not actually new, but more people should know:
- KDE Connect - notification sync, shared clipboard, remote control, etc… between phones and PCs. Supports Android, iOS, Linux, Windows, Mac, and more.
- Syncthing - sync the contents of a directory between multiple devices. Syncthing-fork to do it on Android.
Cleverkeys is such a good find from you. Thank you so much for it 😃
I’ve found KDE connect to work better than the native Windows Phone Connect when I was using windows. I switched to Linux Mint a few weeks ago since I primarily value stability.
That cleverkeys keyboard is… suspicious? Only 2 stars? 0 issues. Never heard of it. Seems like such a keyboard would be much more known, no? Do your due diligence before you install it, it is a keyboard, with potential to cause some serious damage.
It’s fair to be concerned. I have known the author online for several years, and he has reputation to lose if he publishes malware. Of course, you have no reason to trust me either.
It’s a recent fork and rebrand of Unexpected Keyboard.
I’m going to be that guy; I actually use LLMs to get more done as an IT technician. It’s a starting point, not a cure-all. And being honest/direct in one’s prompts is a huge part of the output, too.
But nobody wants to talk about tokenized data and how word groupings reduce false positives. It’s only going to replace window lickers.
Window lickers? Never heard that term.
I do all the time as well. Has changed my life in a very useful way. But as you said, knowing what it’s place is, how to use it, and what it’s limitations are (as well as my own) are key. I have solved many many problems I’ve been working on for years on in the digital world.
I also sympathize with the AI hate, and really struggle with the energy usage as well as the bubble. It has power and capability, but not what the “public” think it does.
I just deal with the online hate as it’s not shit people says to my face, and it’s driven of ignorance like much is these days.
And as you said there are development in the pipe which will further change our lives. Knowing how it works and why, as in using the critical thinking in synthesis with an LLM and what comes next is going to be valuable.
The power usage isn’t even that much on an individual basis once it’s trained, it’s that they have to build these massive data centers to train and serve millions of users.
I’m not sure it’s much worse than if nvidia had millions of customers using their game streaming service running on 4080s or 4090s for hours on end vs less than an hour of AI compute a day.
It’d be better if we could all just run these things locally and distribute the power and cooling needs, but the hardware is still to expensive.
You have apple with their shared GPU memory starting to give people enough graphics memory to load larger more useful models for inference, in a few more generations with better memory bandwidth and improvements, the need for these data centers for consumer inference can hopefully go down. These are low power as well.
They don’t use CUDA though so aren’t great at training, inference only.
I got a smart watch a while back. Had a lot of fun setting up workout routines, tracking my run, measuring my activity and sleep, putting in reminders for water and movement, setting alarms, journaling my period, syncing music to headphones, connecting calls from my phone etc.
Then I got bored with running and stopped logging in online to transfer my data, got annoyed with all the reminders that always popped up at the wrong time, turned off bluetooth and wifi within a week because it drained both phone and watch on battery, so no longer connected to headphones or phone, and of course I quickly forgot all about tracking periods etc because I had to turn on wifi and sync to do so, and figured out the sleep tracking was bogus.
I still use it all the time, to check the time without pulling out my phone and risk getting distracted. It’s great to have this technology that shows the time right there on your wrist!
Kinda annoying to have to charge it weekly but well worth it for keeping the phone tucked away safely. I even finally started automatically checking my wrist instead of reaching for the phone when needing the time. Pretty neat!
Regular wristwatches: are we a joke to you?
Which watch do you have, if you don’t mind me asking? I’m actually on the verge of getting a new one since my old one broke months ago.
For how I use it, I recommend just getting a classic analog watch… I had a pretty cool one as a child with Bugs Bunny as a background. It was neat.
I like checking my pulse for fun and use the alarm for reminders during the day on my smart watch, but that’s all the extra features I still use out of all the extras. It’s a Garmin.
If you have a high refresh rate display, Lossless Scaling is life changing. It’s a $5 app on Steam and it can take any application and introduce frame generation. It requires a decent gfx card, but it doesn’t need to be a recent one. It works on Linux now too, or at least on the Steam Deck.
The use case for me is primarily for older games that are locked to 30 or 60 fps, including PAL region games locked to 50 that normally have microstutter. Games from the PS1 era at 120 fps with geometry correction never looked or played better. But even some newer games are fps locked and can benefit. It also makes the app full screen eliminating my need for Borderless Fullscreen too.
I’ve been enjoying the payment app and personal information verification app in the country I moved to quite a bit. The only previous exposure I had to a centralized app-based payment and/or verification system was China’s WeChat, so I had quite a bit of negative stereotypes with them… but I’m liking these a lot
For additional information: the two are separate apps. Both are private companies (I think?) heavily regulated by and strongly promoted by the EU I believe, and the latter is the de-factor verification system that is used for governmental stuff as well
It’s Me?
Yes. I’m impressed at how well it works
What I love even more though is that you are not forced to use it. I just got a 15€ card reader for my PC and can use that instead of the app as well.
Yt-dlp to mirror YouTube channels to my Jellyfin instance.
Does that still work? Last I tried it, YT broke it.
I update it once a month with the inbuild -U switch
Well, fuck. Just at this exact second you’ve taught me that I’ve been doing that the hard way for ages, by actually going to the project’s github page.
Anyway, another shout out for yt-dlp regardless. I get a giggle every time I see one of those sporadic news articles involving the music recording industry still whinging about piracy. Er, the record labels themselves pathologically post every single track ever recorded to Youtube to rake in that ad revenue, and it’s all free for the taking. If you decide you’d like to be proud owner of any of them forever you can just hit it with the ol’
yt-dlp -x.I am continually amazed at the number of non-Youtube sources that yt-dlp Just Works with as well. It seems any video content posted online that you’d like to gaff can be handily vacuumed up with it, regardless of the site operator’s desperate attempts to prevent you from doing so.
How did I not know about the -x param? I’ve been doing --list-formats and then choosing the high quality audio only. This will save me a step. Thank you.
Same way I didn’t know about the -U update parameter, I’ll bet you. You ask yt-dlp to list its flags and arguments and it spits out a listing into your console that’s about nineteen miles long because apparently it can do anything.
The only command line tool I use regularly that’s comparably capable and even more byzantine is ImageMagick.
3d printing. Custom parts for the house. Replacement plastic bits that broke or got lost. Custom toys for gifts.
I live in a pretty busy car dependent city, and I usually put my daily commute into Google Maps just to get advance warning about traffic jams or wrecks.
Same, my commute is about a half hour, with 2 main routes I could take that are about the same time. Just knowing if one of those is screwed up can be the difference between an okay drive in and hell.
linux
“new”
“unix”
“TempleOS”
“paper”
“tablet”
Telegram, the ‘Saved’ channel is very useful, good for quick note taking and sending files across devices. I consider its contents public and don’t upload anything important tho.
I would say youtube autocaptions have gotten pretty good. I often download and read the transcript instead of watching the annoying video.
Which tool(s) do you use?
To get the transcript? Either copy/paste from the web page, or download with yt-dlp and clean up the.vtt with a python script.
“new”?
The fediverse is really nice. Some cool advances in programming, but nothing major.
Obsidian is really nice, with plugins can do virtually everything you’d want from a note taking or even writing app.
localsend is a syncing/sending solution via wifi for smartphones or computers in the same network.
If you like obsidian you might check out the FOSS project Silverbullet. I have almost no experience with obsidian, but I love Silverbullet.
Love Obsidian. I’ve been using it to study for IT certs and organize my worldbuilding ideas. Been looking for a FOSS alternative that won’t enshittify. Hasn’t happened with Obsidian yet but they do paywall useful features like accessing an existing vault on mobile from a file share.
Logseq maybe?
No it’s scope is more limited vs Obsidian. Logseq is primarily an outliner whereas Obsidian is anything you can express in Markdown
It’s wild following the Obsidian subreddit. People are turning it into a whole ass OS or into a desktop environment at the very least.
It’s like the new emacs hah
Oh yeah. My favorite (and only) plugin so far is the https://github.com/twibiral/obsidian-execute-code
Let me explain: Obsidian is basically a very fancy wrapper around a folder with markdown files in it. (which makes it git compatible, which is one of the upsides). In Markdown, you can define codeblocks, with syntax highlighting, because of course you can, programmers will improve their own tools first. Now, there are two cases when you would do this:
- you want to execute the code because it’s actually driving something. Like some kind of interactive, “this is the manual, but also, you can just do it right away by executing this code” and then they give you the code.
- you’re actually building it as a document, and you want something in your document that is actually the output of some program that’s producing some output. Like… analyzing numbers and creating a graph. You can now just put the code in the document, hit “execute” and you get your output in the document right then and there. And that concept isn’t new, it’s what “jupyter” also does, but jupyter uses a weird bytecode, xml zip format or something, in obisidian, because of the markdown base, it stays just code. (which again, makes it git compatible where jupyter isn’t) AND you can do it not just with python but with…
- JavaScript
- TypeScript
- Python
- R
- C++
- C
- Java
- SQL
- LaTeX
- CSharp
- Dart
- Lua
- Lean
- Shell
- Powershell
- Batch
- Prolog
- Groovy
- Golang
- Rust
- Kotlin
- Wolfram Mathematica
- Haskell
- Scala
- Racket
- Ruby
- PHP
- Octave
- Maxima
- OCaml
- Swift
AirTags. I have 7. One for my keys, a Find My compatible third party card shaped one for my MagSafe wallet (also third party), one for my second wallet, and the rest for my bags and suitcases. They have saved me so many times.
As a blind person with ADHD they’re invaluable. I’ll forget where I put something mere minutes after setting it down and can’t see where it is.
I’m using owncloud for that but (don’t hate me) I’m having to use chatgpt a LOT for the initial configuration. There’s a lot of stuff I’m having to learn to get back into it and it’s really good at searching docs by fuzzy match, pointing out what part of a function block isn’t working, and drafting individual function blocks to experiment with (I’ll comment out the original, paste in the chatgpt one, test it, etc). It’s a much better experience than a decade ago and I haven’t gotten emotionally abused by a more experienced person on stack overflow or reddit once yet.
I’ve got owncloud running as a lan only service so it just auto syncs my calendar, contacts, and a few folders of documents / pictures I like to keep with me but only whenever I’m home. Right now I’m just also configuring the same old computer to do some basic media / retro gaming on the TV and the hard part right now is configuring magic mirror as it’s idle display to keep hubs and I straight on our calendars and tasks (and show some family photos). Nothing major just as I get back onto Linux after ten years it’s nice to not get called lazy and stupid while I’m learning.
Usenet for obtaining movies and TV shows, it’s so much better than torrenting. I only pay for one streaming service these days (and only because they’re fairly small and I really like their content).
That’s 50 year old technology! But still better than the stuff being foisted on us now.
