>>107719697
linux > macos > windows, but framework has good linux support. it basically comes down to what your preference is, and the big memory chunk is nice if you want to play with local LLMs

How small are we talking about? I would say its mac or itx build catered to your needs.

>>107719697
Depends what you need said performance for...

>>107719697
For that price if you don’t want or care for windows, the Mac is the only choice.
Before this month I would have called you a retard for buying the framework, but considering the price of individual components, it’s probably not a horrible deal these days, still have to deal with soldered on bullshit on a PC, but whatever at this point.

>>107718275
>Why do people cling to that and the reddit man so much?
Dumb people prefer repetition to new things.

>>107714079
Are we pretending like a smartphone is full of hardware bloat taking up massive space? The thickness of a smartphone is 100% dependent on the battery. How are we getting "super-thin" exactly?

>>107714091
FPBP

>>107715009
>tfw I asked anon to draw the Gentoo-tan picture in the bottom left
Damn, time flies...

>>107717976
It’s literally just Nano Banana but made by Elon Musk’s company

Any good LED drivers?
Got some shitty one off amazon and when I turn it on I can hear loud buzzing in my headphones connected to my laptop
Something with a transformer prefably that it won't put noise in the entire electrical grid

>>107716775
Honestly neat, all this lora stuff looks interesting but I don't have the time or willingness to do an autistic deep dive into the DIY of the hobby. A prebuilt phone-like thingy that just works looks cool. Though the ultra low end MCU doesn't inspire too much confidence for having a lot of functionality (what's the camera even for, isn't lora super low bitrate?)

The other main issue is having people to talk to

>>107720557
Yeah, I know there's really shitty ones and there's good ones. And the good ones aren't even that much more expensive, you pay a bit more of course but it's still like $30 for one or some shit so IMO very worth it
However it has been ages since I've looked at this so I don't remember exactly what the different types are called or which ones are good. But if you do some research you should be able to get an idea relatively quickly.

>>107715958
late to the ai game here... how can i use ai in my tech job to do less work/make my life easier?

>>107720668
What's your job?
If it's programming, download either cursor or claude code. Get a subscription or ideally get your work to pay for a subscription.
If it's something else then you'll have to figure it out I dunno man. AI can do a lot nowadays, but everything needs supervision. Imagine having access to a monkey that has read every book every written and is well trained to behave intelligently, but still sometimes thinks smearing poop over things is a good solution because it's a monkey. Now imagine hiring that monkey for $20/month. What kind of low-level tasks can you give it where if it does them correctly, it makes your life easier, and if it gets a nonsensical monkey idea and you have to throw it away and gently go "no monkey, good attempt but that's not it" and probably do it yourself, it's not a big burden? The key is to minimize verification effort.

I use Windows because I have nothing to hide.
Oh, you want me to post all my private info for you? Ok, provide me with a quality service like Microsoft has for the past ~3 decades and I'll give you all my info. Can't do it? AHAHAHAHA didn't think so.
Your privacy has been compromised the moment you connected to the internet btw, a non absolutist stance to this is just pure vanity.

>>107720495
>I'm cucked by Ranjeed Shekelsoft and if you provide me with gibs i'll be cucked by you too.

>>107709926
I have this laptop. It's OK. CPU performance is enough for web browsing, but not a whole lot more. GPU is basically just good for hardware video decode. When I first got it, Linux support was questionable. It took a fair amount of work to get the touchscreen working, and sound didn't work. But these days everything just works at least with newer versions of Ubuntu. Only other weird thing is that the mouse buttons are mapped to the numpad, so you will need to remap them. Battery life seems pretty good.

I had the bearings go out in the fan after a couple months. So I just unplugged the fan and added some more thermal pads to utilize the metal case as heatsink and the thing stays perfectly cool (and a lot quieter too). The biggest problem with this device and probably all mini laptops is that the screen is just too small to easily use most applications with reasonable scaling. I don't need a laptop all that often, so I can live with it. But don't expect to do extended work on the laptop's built in screen. And there is no multifunction USB C port, so don't expect to be able to use a docking station either.

File: images.jpg (8 KB, 331x152)

8 KB JPG

>>107712111
I miss the 90s/early 00s retro-futuristic trends in Anime. Combo of then contemporary computer forms with big words like Internet, AI, holograms etc was kind of charming. Sadly it pretty much died a natural death.

Corrector Yui, Rockman.exe, first Digimon etc were pretty much build around these aesthetics, but it was kind of everywhere during that era.

I've been looking for this but I haven't found it:
>x86 tablet (no keyboard) for full compatibility with older software
>8 to 9'' screen, or considerably small
>Powerful enough for anything that isn't games
>A digitizer (pen support)
I think such a small size is retarded, but the person looking for it seems obsessed with the idea of having a device like this because 10'' is already too much to carry or something, I don't know. A tablet is specially unusable with Windows at that small screen size, I'm guessing that's one of the reasons pen support is requested.
I found some devices like this but they were old and definitely not powerful. I found that a very powerful small device is the Lenovo Legion Go if you don't use the gamepads, but it doesn't have a digitizer. I found somewhat decent tablets with a digitizer (I guess the fucking Surface ones) but they're all too "big". You see where I'm going with this?

>>107711020
happy new year tohru maid. unfortunately, 15 inch with mechanical keyboard is better, bigger is better. I can only get satisfied with 15 inch.

>>107717209
I had a gpd win 1 and a politely loved that little guy. He died a few years later and I didn't think to check out the steam deck which came out a few years after. I always listed after handhelds but ignored it until I started traveling for work and then I decided I was tired of raw dogging international flights and bought an Asus ally x.
Im now never going back to gaming rigs and laptop. Handhelds all the way. My new laptop is a 14inch elite book whose sole purpose is just general computing. No gaming. That's what the ally is for. My next handheld is going to be a legion go 2.

>>107719986
why do people use this shit? What is it really, i use runit and it works fine and it is fast, you can write a simple script that will be your whole init system. Systemd is over-complicated and i have no desire to learn something like that. I only use simple, well written software.

Recently built a small webserver in raspbian. When setting up the hostname, I find out shit'emd fingerprints my hardware and shares the fingerprint with any user. Wtf.

>>107720229
what if a new package needs to start, will your package manager just happens it?
what if it strictly needs graphics or network on start? what if it depends on a service?
what if it needs a specific disk to exist?
can it parallelize?
can it create a report of detailed timings of every component?
can your well written software handle all this, automatically, without explicit intervention by a simple invocation of pacman -S or dnf install, maybe only with a restart in case it needs early boot conditions? or does it need constant babysitting and manual adapting for thousands of different combinations of packages

>>107719356
Did not meant to quote

>>107716966
Sounds great. I'm definitely interested.

(defun eshell/args (&rest args)
  "Experiment with ARGS parsing."
  (interactive)
  (eshell-eval-using-options
   "args" args
   '((?a "all"    nil show-all "show all things")
     (?h "help"   nil nil      "show help message")  ; nil nil lets me get usage on -h
     (?n "number" t   number   "pick a number")
     :show-usage
     :usage "[OPTION]... [ARG]..."
     )
   (unless args
     (setq args (list ".")))            ; How doess eshell/ls work with no args?
   (eshell-with-buffered-print
     (eshell-buffered-print

Comment too long. Click here to view the full text.

>>107719361
>TLDR: recursive check for .gitmodule file over each submodule in repo, the operation is too expensive for emacs due to having to wait for a response from remote to asses failure, the constant failures and retries kill performance on tramp.

You'll need to disable submodule merging for all repos read over tramp by setting project-vc-merge-submodules to nil
> project-vc-merge-submodules: Non-nil to consider submodules part of the parent project.

Something like this maybe? Not completely sure haven't used it myself.

;; Disable submodule merging for all remote connections
(connection-local-set-profile-variables
 'remote-project-profile
 '((project-vc-merge-submodules . nil)))

(connection-local-set-profiles
 '(:application tramp)
 'remote-project-profile)

Comment too long. Click here to view the full text.

>>107717855
you misunderstood, it's no different than installing nix on guix distro manually from source
there's no integration between the two package managers, they behave independently

File: sanakan-meth (emethmeth)-(...).jpg (135 KB, 1036x1666)

135 KB JPG

>>107720311
>we are talking about corporate headquarters that can take entire buildings
yea and kings used to have palaces. there's a lot of flunkies, what can i say.
>then you are not using them enough to notice the difference.
o am well aware ai never sleeps. is there a video where an llm completes some sort of task? little crappy demo it stole off github isn't a use case. i try to get chatgpt to write the features i actually need, and it never fucking works. at best it can just suggest a framework, if the feature already was solved.

>>107719987
nigga, try the new AI tool by google
The bot don't read the whole text, instead it do text search by making scripts. It can also do simple interactions with browser. With moore law o algo it won't be long before the bot could use the whole computer or phone

>>107718779
Why the FUCK do you want everyone to use AI to wipe their fucking ass?

I can't find a fucking use for AI, other than local model girlfriend. It's fucking useless, at best it's as good as googling shit five years ago, at worst it's like talking to a retard. It can't reason, it can't think, it can't imagine, and as years go on it keeps getting gutted, censored and enshittified. You need 5 prompts to google shit through chatgpt now. I can't believe the entire world economy depends on that shit fucking company.

File: sanakan-nihei tsutomu-saf(...).jpg (267 KB, 1450x2048)

267 KB JPG

>>107720437
>before the bot could use the whole computer or phone
...to do what? i've most recently tried to get chatgpt and claude implement or at least describe iterator logic for a very specific vector-like container i have. they just can't. they can tell me about existing things which are outwardly similar to what i want. they can't actually come up with an implementation, even if i give them lots of hints. even when i straight up described what to do they implemented it with errors. i am almost 100% certain this is an architectural issue in ml, or animal brains would've worked differently.

>>107720482
>I can't find a fucking use for AI
sounds like a "you" problem

>>107706756
YES FLOOD THE MARKET WITH THESE PLEASE!

>>107706776
>two more years
This but unironically

>>107708514
Wrong on so many levels
Tic toc jew
we can smell the rot from you

>>107708514
And here ladies and gentleman we see an example of somebody who has been deeply affected by propaganda.

>>107706912
>PowerVR
that's Dreamcast not Saturn

>>107711915
Nvidia will unironically go bankrupt when the bubble pops

File: 1766870964977306.jpg (520 KB, 817x1113)

520 KB JPG

>>107711915
Nvidia exits the consumer GPU space, AMD and Intel become the new choices. Protect Airstrip 1.

>>107712067
>>107712341
>>107712870
Living proof that "anti-semitism is the socialism of fools"

>>107720568
>anti-semitism
You used a hyphen there which refers to a different thing entirely. I know that it makes sense to put a hyphen there, but you have to understand that the hyphen was removed for a reason.
It appears you were trying to make a point about antisemitism instead. Unless you really did mean the semitic people....which would be very strange given the context of the posts you quoted.
I would advise you not make the same mistake in future as adding a hyphen to antisemitism meets the IHRA definition.

>>107716603
>Christianity was magically adopted by white people

>>107717592
>He hasn't upgraded his light switches to microcontrollers.

>>107718566
How fast can you perform a normal double-click? Wouldn't be surprised if the mouse already has internal corrections for it.

>>107718939
A "normal" double-click, a typical computer user averages between 140 ms and 190 ms for the interval between clicks.
Operating systems like Windows default to a 500 ms (0.5 second) window to ensure accessibility for broader groups such as the elderly or poop-colored "persons."
Experienced users including hardened gamers often land in the 80-150 ms range. In Windows, the fastest double-click setting is usually capped at 200 ms.
When a switch fails (due to oxidation or low wetting current), it may register a "bounce" or second click in a range of about 1-10 ms. This is far away from the 69ms maximum speed that any human has ever achieved an intended double-click with one finger (40ms may be achieve with two fingers).

>>107718566
>>107718939
Just wrote a tester. Best I'm getting is 100ms usually. Occasionally below, perhaps 95ms with high intensity.

literally just buy a 10 pack of cheap silent switches from aliexpress and get a $3 soldering iron

>>107713384
A lot of sites don't accept weirdo email addresses.

File: 1_TZ2H14_VKWlkwvVHOVxP_g.jpg (72 KB, 534x640)

72 KB JPG

>>107718314
My time in this industry is short
Definitely past time to change paths

>>107713054
>Google locked me out
you get what you fucking deserve

>>107713054
It's pretty insane that websites ask for copies of your ID. I can remember a time when giving copies of your ID was always discouraged, unless your government, bank, or doctor needed something.

>>107720612
throwing money at politicians can go a long way

File: 1738911216493722.png (604 KB, 1440x1434)

604 KB PNG

>>107719550
>dude 4:3 is totally prey behavior
fuck off with your furry shit

>>107718925
We need to stop shilling outdated crapware.

>>107718925
All displays should be 4:3 just like God intended. 16:9 is unnatural and satanic.

>>107719130
i use 9:16
vertical screen for reading.

File: 25lel.gif (2.83 MB, 260x207)

2.83 MB GIF

I wish we adopted 1:1, 2:1, and 3:1 layouts as standard

>>107718846
I'm not one to criticize taking a long ass time to get something working from the ground up, let alone a project that was born out of spite because they were having trouble with this theming thing, but this is basically GNOME written in Rust with a bunch of extensions, they're not exactly creative with this shit
Maybe someday it'll get better, maybe.

>>107718596
>You are welcome to work for free for system76

>>107718596
No thanks
Learn how to write software

File: bazz.gif (209 KB, 800x854)

209 KB GIF

>no native image viewer

>>107715418
> Why did it fail?
- Rust kills productivity.
- Wayland is a shit API.
> It's been over 4 years in development...
If they would have used X11 and Tcl/Tk it would have taken 4 weeks of development.

>>107719223
tidied up a and used writev instead of fwrite, a bit better now https://files.catbox.moe/fqehlc.c
happy new year

>>107719970
Happy new year!
>tidied up a and used writev instead of fwrite, a bit better now
This one doesn't work for me, it gives EINVAL whenever a bucket has more than IOV_MAX=1024 vectors.
In the Rust code I don't make an iovec for each individual line, that was slower when I tried it. I allocate a big concatenated buffer per bucket in the parallelized section and then do a vectored write of those buffers.
The vectored write seemed to be just barely faster than one write per bucket but that's plausibly measurement noise. Being able to parallelize the buffering was a bigger deal.

>>107720378
>it gives EINVAL whenever a bucket has more than IOV_MAX=1024 vectors
huh
I'll take a look later

File: Даша изучает Rust.png (1.11 MB, 781x1080)

1.11 MB PNG

Would you marry a Rust girl?

>>107720610
>3DPD
eww