>>43005
I always thought it was for preservation purposes. Low-sodium anything is usually more expensive than the regular one.

>>43006
>>43007
Does dry powder really need that much preservation? Especially if it's also spicy? I assume that's where most of the salt is, but I never checked if the noodles are salty on their own.

Post edited on 17th Oct 2024, 10:35am

I have been taking anti-psychotics for years now, but recently the side effects really seem go hard on me. All of my cognitive abilities seem to be worsening. Thinking generally, remembering stuff, doing things I like or just doing anything that is not being idle is very hard for me and it doesn't seem to get better. I feel brain-dead. The worst thing however is, that I have no choice, should I not want to get psychotic again. It's really a curse and I hate all of it, although taking the medication is by far the lesser evil.

>>43015
no schizo no antipsychotics but i feel brain dead anyway. at least it's not as bad as being psychotic.

>>43015
>doing things I like or just doing anything that is not being idle is very hard for me
I can at least sympathize a bit, since sometimes I feel this way (but likely not as bad as what you are experiencing). Unfortunately unlike like antidepressants, the symptoms of not being on antipsychotics seem worse than the alternative.

You could talk to your doctor and maybe ask them to try switching dosage or the drug?

>>43017
>You could talk to your doctor and maybe ask them to try switching dosage or the drug?
I have changed medication so often already that I'm glad that something finally works properly and without any major shenanigans. Unfortunately almost all anti-psychotics have, what I described, as a side effect.

I hate math. Every time math has come into my life, it's caused me suffering. There's two approaches to dealing with it. You can either dedicate your life to learning it "properly". Learning and memorizing endless theorems that took people centuries to figure out all the way down to arithmetic which just makes sense intuitively.

If you're not a turbo autist who genuinely enjoys doing that, if you just want to use math as a tool, you have to memorize all kinds of math tricks. Properties, identities, "rules" and patterns. You have to grind those into your memory and use them often enough or you'll forget. Every problem is like a secret code you have to throw your toolbox of math tricks at to decrypt. You don't understand how these tricks work though, so it might as well be magic.

If you fuck up at any point during this process, you'll get the wrong answer. You might not even know you got the wrong answer. And you wont know where you fucked up, so most of the time your only option is starting from the beginning and painstakingly retracing your steps and throwing the rest of the kitchen sink at it.

Past a certain level, computers barely help. Typing out math is an ordeal in itself because it's such a symbol clusterfuck, so you need to learn a DSL created just to depict it on a screen. You need to pay a person, or beg one to explain things, or spend hours digging through some tome looking for that one little math secret that you need for your specific problem.

Post edited on 27th Oct 2024, 6:16pm

>>43042
First approach is the way to go. You don't have to be a turbo-autist, but most people are not taught math properly and re-learning it from the ground-up starting from algebra can take you maybe a year or two. Presentation of math is also poor, especially for calculus and linear-algebra, you basically need to take things a la carte. There is also a lot of outdated baggage, like formalizing differential calculus in terms of epsilon-delta instead of hyperreals which is just a lot more intuitive.

>If you fuck up at any point during this process, you'll get the wrong answer.
What level of math are you referring to? Past college-level the math is less about computation and more about underlying structure and there isn't really an "answer" per se, just presenting lines of reasoning.

>>43043
>re-learning it from the ground-up starting from algebra can take you maybe a year or two
How would you suggest doing this?
>What level of math are you referring to?
What sparked this rant was me trying to learn about neural nets and not understanding a step in the math part of the explanation because I forgot about the chain rule, which I learned years ago and haven't used since.

Besides that, it's a generalization of my experiences throughout school. Of getting stuck on problems on a test because I either forgot one of the mythical identities of trigonometric functions I was supposed to have permanently etched into my memory 2 years ago, or failed to see the multi-step factorizing and regrouping that would get me to one of those identities.

Wikipedia articles don't help much either seeing as they're written like a grimoire.
https://en.wikipedia.org/wiki/List_of_trigonometric_identities

>>43044
> neural nets and not understanding a step in the math
Ah yeah backprop is just pretty much chain rule, and in terms of implementation (reverse-mode autodiff) it's just chain rule applied not naively to the entire expanded function (which would be symbolic differentiation) but to the dependency/compute graph. And the computational graph effectively _is_ the topology of the neural network, and since chain rule composes nicely implementation-wise you can end up working backwards from the last layer to the first: and when you're done you have the gradient of error with respect to the weights.

I absolutely agree though that neural-network stuff is sort of more abstruse than it needs to be, there is a lot of index chasing in the notation. I think for ML stuff you really need a solid understanding of linear-algebra and multivariable calculus. In particular multivariable calculus grounded in linear algebra, which is not how it's commonly taught in college (e.g. if you were never told that the derivative is a linear operator and so chain rule is just matrix-composition of linear operators, then a lot of terminology around jacobians is going to be abstruse). And actually with ML there is a sort of short-hand notation developed called matrix-calculus. The concepts of things like gradient descent are straightforward but at the end of the day there is a lot of symbol chasing and some bookkeeping stuff. But I think there's no real trick there, it's just mechanical grinding out the derivative which could even be done by a CAS. Most of the insights in machine learning come from playing with the underlying topology of the network and choosing building-blocks other than fully-connected layers (e.g. insight that convolution would be useful gave rise to CNNs, then the idea of quantization, and most recently stacking Attention blocks).

>How would you suggest doing this?
I actually don't really know, I guess it would be easiest for you if you go top-down, starting with the goal and identifying what you're missing in order to achieve it. E.g. for ML you said you forgot chain rule. That's a big warning flag to me that you missed the essential underlying insight about how chain rule is just composition of rates, it's not really something you can "forget" because the principle when understood is so simple.

Try this book from 1913 - https://calculusmadeeasy.org/ . It focuses on core insights rather than computation, and sticks with the intuitive hypperreal approach I mentioned (and doesn't get too bogged down in formalism).

>>43045
>Try this book from 1913
Thanks. I'll check it out. You said going back to algebra would be the right approach though. I've heard mixed things about textbooks like Lang's Basic Mathematics. To be honest, I hate reading textbooks and prefer explanations that are as concise as possible, which the one you suggested looks reasonably so.

>>43046
Whether you need to go back to algebra is up to you. Do you see it as just a game of symbol manipulation, or do you understand the core ideas involved? I can't really suggest a book because it's just so innate by this point that I don't even know what distinct concepts you'd need to cover.

I'm alway a bit skeptical of well-suggested books, things like Rudin (for calculus/real-analysis) and Axler (for linear algebra) turn out to be complete garbage with no real intuition. Most of my learning has been through identifying concepts that aren't clear, searching online, then finding some obscure forum post that makes the underlying intuition clear. Sadly a lot of resources have disappeared, and google is no longer useful. Stackexchange is good but you sort of have to know the terminology to search for in order to find what you want, so it's a bit of a chicken-egg issue. Maybe you could use GPT et al. to provide some intuition but I'd be a bit wary of that because unless you already have some background it's hard to differentiate a specious argument from a logically sound one and you don't want to build false intuition. (E.g. the common-but-flawed "intuition" that the chain rule holds because dy/dx = dy/du * du/dx because the "du" cancel; it's not really about "du" canceling per se [which obviously doesn't hold in the generalized multidimensional case], but the rates composing [which does hold in the multivar case as mentioned]). There are some good youtube creators like 3Blue1Brown which you are probably already aware of, that take advantage of visualizations to make things clear (very useful for linear-algebra).

If there is a specific concept you need clarification on, you can ask about it here (on /navi/ maybe) and I can try to sketch some general intuition or suggest things you look into further.

>>43047
>do you understand the core ideas involved?
>it's just so innate by this point that I don't even know what distinct concepts you'd need to cover.
>well-suggested books... turn out to be complete garbage with no real intuition. Most of my learning has been through identifying concepts that aren't clear, searching online, then finding some obscure forum post that makes the underlying intuition clear
This supports my assessment that math is an esoteric subject you have to be obsessed with to actually understand. The only resources that teach "multivariable calculus grounded in linear algebra" I could find look quite dense. What I'd really like is to learn enough math that it wont confuse and frustrate me when I see it in an ML context, without reading hundreds of pages of boring, pure-math shit.

>There are some good youtube creators like 3Blue1Brown which you are probably already aware of
I'm not. I will look at him too.

Post edited on 27th Oct 2024, 9:47pm

>>43048
> esoteric subject you have to be obsessed with
I agree that it's presented esoterically, but it's not inherently esoteric. The fact that textbook authors seem intent on obscuring basic truths is only a fault of presentation.

>multivariable calculus grounded in linear algebra"
Yes I don't know of a resource as good/basic as the single-var calculus one I previously linked. I think I learned this is sort of 3 phases, first learning MV calc the traditional approach, then linear algebra, then "unlearning" MV calc and viewing it through the lens of linear algebra. Vector Calculus, Linear Algebra and Differential Forms: A Unified Approach is the book I skimmed (the differential forms stuff is a bit abstract and to be honest I don't fully grasp the nuances there, but I think it can be skipped).

>learn enough math that it wont confuse and frustrate me when I see it in an ML context, without reading hundreds of pages of boring, pure-math shit.
But it takes time and exposure to do that. Semester courses in a university are really misleading, most people who do well in those courses already come in with the knowledge before hand. I think it took me about 3-4 years to fully grasp single variable calculus, taught in different ways by different people. Once you have a rock-solid foundation of the concepts (not the computation, that's trivial. But the actual underlying essential truths) then you can move on.

Reading a textbook is usually not going to help unless it's explicitly written in an instruction style like the calc one, it should be used as a reference. Most textbooks are written as a reference rather than being educational. But the process of understanding, either through lecture or self-exploration is the key and it can't be short-circuited. Mostly through self-exploration I guess, I've never found any single lecture to be helpful on its own, aside from introducing a concept that I realize I need to investigate.

It basically is no different from learning a new language: if someone said they became fluent in JP by reading Genki they'd be laughed out of the room. You need to get exposure to the material in several different contexts so you have an intuition of what is meant by the terms e.g. orthonormal basis.

>>43049
>Once you have a rock-solid foundation of the concepts (not the computation, that's trivial. But the actual underlying essential truths) then you can move on.
If it takes years of effort, I am never going to have that. I will never be sufficiently motivated to extensively research a subject I associate with nothing but boredom and anger for multiple years.

>It basically is no different from learning a new language
Except I don't want to learn it for its own sake. It's like if you had to learn ancient Greek and Latin before you could learn Italian.

>>43050
It's more like learning kanji if you want to read manga though. ML (implementation-wise) is basically applied MV calc/linear-algebra at its core. I guess you can explore some parts of ML using frameworks like PyTorch/JAX to hide all the underlying stuff for you (you just wire up your forward pass, choose an optimizer, and it takes care of the rest), and learn the high-level abstract building blocks without worrying too much about the implementation. It really depends on your motivations. (And I guess there's also a 3rd branch of ML research papers which is about formally proving bounds on stuff, e.g. why regularization works in practice which basically requires a shit-ton of advanced theory that's not practically of much use.)

>>43051
>you can explore some parts of ML using frameworks like PyTorch/JAX to hide all the underlying stuff for you
I want some understanding of how things are implemented. Not phd level, but also not just plugging parameters into a model. What you suggested is spending 3-4 years mastering single variable calculus through independent research. Then somehow learning multi-variable calculus the "right way", which you did by first learning it the "wrong way", learning some other branch of math, then re-learning multi-variable calculus through the context of that branch. No offense, but how is that not being a turbo-autist?

Post edited on 27th Oct 2024, 11:38pm

>>43052
Basic single-layer perceptron example (derivation of backpropagation) probably could be understood and implemented by a high-schooler. But having the intuition and undersatnding for anything beyond that (e.g. backpropagation in a CNN) does require math and exposure to build intuition. An ML class in most universities has a reputation of being amongst the hardest, and that is even among the self-selecting group of students who think they would be comfortable with the math involved.

I kind of don't see your point here: yes it is "turbo-autist" in the sense that people who do ML have had sustained exposure to math. It's not "turbo-autist" in the sense that you have to be a 140IQ genius to do it, if you can do programming then you can probably learn the math involved. And it's in fact the exact same for programming, people who are good at it have been exposed to it for many years, probably since high school or earlier. Very few people come into university knowing 0 CS and end the first year with anything more than a rudimentary understanding of how to write basic control flow logic. ML isn't any different.

What you're asking for is basically the equivalent of wanting to read manga in a few months, and then considering anyone who dedicates years to grinding japanese and kanji to be "turbo-autists".

>>43053
>it's in fact the exact same for programming
No, it's not. Math is uniquely difficult. Not only is it far harder than programming or learning a language(comparing it to Kanji, which can be learned through rote memorization, is absurd), it is useless in isolation, and excruciatingly boring because of how abstract it is.

With programming, you create something. With language, you gain access to a culture. Math is a torturous pursuit that can only be enjoyed by abnormal people. That's my point.

>>43015
I came across this article and I was reminded of your post. https://www.theguardian.com/news/2024/oct/29/acute-psychosis-inner-voices-avatar-therapy-psychiatry

It's probably unrelated to your specific condition unless it involves auditory vocal hallucination, but I really assumed that the above was something most psychologists would try as the first strategy before resorting to drugs. Turns out all you need to do is anthropomorphize the voice, rig up a Live2D avatar, and do something akin to vtuber therapy.

Post edited on 29th Oct 2024, 4:46pm

>>43067
I'd like to know how much of it is the guardian and how much truth. Is there any real article to support this?

>>43070
https://www.ncbi.nlm.nih.gov/books/NBK373172/
although the Cochrane review is less enthusiastic
https://pmc.ncbi.nlm.nih.gov/articles/PMC7387758/

> The theory behind Avatar Therapy is compelling but the practice needs testing in large, long, well‐designed, well‐reported randomised trials undertaken with help from – but not under the direction of – Avatar Therapy pioneers.
But it seems simple enough that surely it should be the first thing to try, it doesn't seem to have much risk and certainly beats pills.

I started to learn touch typing. I figured I'm already a computer addict, so I might be a highly skilled one. I've hit a road block. The bottom left row is fucking aids. How do you type c with your middle finger, without either contorting your hand in some horribly unnatural way, or floating off the keyboard? When I try the latter, I forget everything and start typing my normal way. I bet most people don't touch type, because if they did, the "proper" way wouldn't be so awful. Maybe keyboards would be shaped differently. And I don't want to use some hipster, alternative layout bullshit.

Post edited on 5th Nov 2024, 9:19am

>>43086
I find the traditional ASDF JKL; homerow to be abhorrently unnatural to the shape and range of motion of a human hand on a regular keyboard (perhaps it works better with those staggered-column bowl-shaped split keyboards).
Personally, I just learnt to touch type over time, learning the shape and feel of the keyboards I've used, using the grooves in between keys as my "home row", with my palms and fingers positioned in what they find to be the most lax state, never having bothered to think about the ASDF JKL; homerow.
I figured most would acquire touch typing naturally after their first few years of continuous typing, however it not being utterly widespread makes sense after some contemplation, yet it makes me wonder why that traditional homerow has so much staying power.

>>43087
I mostly type without looking, with my hands free floating, but glance down to orient myself and for special characters. Otherwise I'll frequently misinput characters close to what I want.

>>43087
Yes, absolutely. I think people promoting this rigid 1:1 mapping between fingers and keys is doing a disservice. Maybe it works at the start, but you will quickly get RSI if you keep following that. I've been touch typing since a child, and I only recently consciously realized what I was doing a few months back, and when I had the conscious realization that what I was doing was nothing close to what "orthodox" touch typing on a qwerty layout was supposed to be, it made some mental block that made it hard for me touch type for a couple of weeks. (That's when I made those posts on /an/ about realizing what the girls were going through when they suffered Yips).

Then I researched more and found that all the pro touch typings say that they don't follow any of the textbook advice and after some time things went back to normal for me.

I still can't consciously describe what I'm doing, but basically the gist is that people who intuitively touch type (like the really fast people) don't have any fixed resting positions for their fingers. They don't keep "returning to home row keys" and instead directly move from the letter they last typed to the (possibly predicted) next letter they'll next type. There are also no fixed finger-key mappings: hands are around to avoid bigram conflicts. E.g., RT typed with two different fingers (middle and index usually), BE is typed with right+left hand, and OB is typed with right+left hand: thus B is typed with the left in BE, but with the right in OB.

This is the reason why QWERTY actually has staying power. It's because that once you give up the rigid mapping, and approach a more dynamic layout, there is no longer any RSI or speed issue. It must be the biggest hidden secret, because if you follow the "traditional" approach you will probably find it very hard to surpass 80wpm or so, so every comfortably skilled typist _must_ be doing this dynamic approach.

It is a very complex process, impossible to consciously describe. The mapping depends on so many factors: which keys I typed last, what keyboard I'm using, how my hands are oriented next to my desk, probably even just on whether I'm feeling lazy that day.

Well I mostly type without looking, mainly because for the past (how many exactly?) years I've been talking to people exclusively on the internet so I guess it was inevitable. But I didn't really learn the thing, it just came naturally which explains why I suck so much at it, especially when I am tired.

With few exceptions, 3rd party GPUs are incredibly ugly. Pic rel is an an extreme example, but the majority of them include the same superfluous, gaudy elements. Pointless ridges, fins and creases. Red streaks. Loud logos on the fan hubs. Strange geometric shapes.

The gaymer aesthetics exists in other computer parts, yes, but it's much easier to avoid with them. For every rgb-laced component, there are many that look normal. GPUs choices on the other hand are a lot more grim.

It's like manufacturers think gamers are mentally challenged and wont buy a computer part unless it looks like a transformer. NVIDIA's founder edition cards looks nice and sell well though so I don't know why they have that impression. AMD is in a worse spot aesthetically.

Post edited on 5th Nov 2024, 2:00pm

>>43091
What about the following aesthetic?
https://www.pcworld.com/article/2202972/we-need-more-anime-girl-graphics-cards.html

>>43092
I skimmed through a couple paragraphs of that article and almost got cancer.

>>43092
I wouldn't buy it myself. I'm generally not a fan of applying one thing's aesthetic onto another. Also, partially covering up the fan blades can't benefit performance.

Post edited on 5th Nov 2024, 3:04pm

>>43086
I don't know man, I've naturally "learned" touch typing without specifically studying it. Maybe I'm pressing the keys the "wrong" way, I don't know, but it works for me. I only have trouble with numbers.

>>43086
you're right that the shape of keyboards is horrible, and it is bad for your hands.
There are more ergonomic keyboards, but they're expensive:
http://xahlee.info/kbd/ergonomic_keyboards_index.html

I learned touch typing in one week with a linux app called KTouch about 20 years ago.

I recommend getting the vimium browser extension, which adds a lot of easy to reach shortcuts to the browser.

For about 6 months, I've been at my first job out of college. It's working with mainframes, which use COBOL. COBOL itself, while being alien, cumbersome, hard to grok, and reliant on other languages used nowhere else like JCL, isn't the main problem. It's the tooling. My god is the tooling a hideously excruciating, byzantine nightmare.

To interact with the mainframe, you need a specialized terminal emulator, with wonderful features like not being able to adjust the viewport size. Forget about SSH. There is no command line, instead there's an abomination called ISPF. Imagine if you operated your computer entirely through nested speadsheets, that are keyboard navigated. To do something as simple as copying a file, you have to drill down to the "copy file spreadsheet", type the full paths of the old and new file into the correct cells, and run the macro. Everything is like this. Everything.

How about code editing? Forget about vim. Forget about vi. It has its own thing with its own weird conventions. For example, when it was invented, there were no arrow keys, so up, down, left and right, are f7, f8, f10, and f11 respectively. IBM did make an Eclipse fork to work with files on a mainframe, but you can only use that to view code, not edit it(at least at my job).

Version control? Forget about Git. To move code into production, there's like a 50 step process with some done in ISPF, and others in Jira, which is web app used for agile shit. Changes are documented by going to the "diff spreadsheet", running that on every file you edited to create a file with the comparison, then using a utility the terminal emulator provides to send those files from the mainframe to your machine. At my job, these are placed in a Windows shared drive.

So yeah, it's shit. At least I don't have to do anything most of the time.

Post edited on 20th Nov 2024, 7:38pm

>>43179
I'm still in college and never worked on a tech job but that doesn't sound so bad. I'd probably get used to these quirks and maybe even find a way to automate some of the tasks. And if it was remote, that'd be a no brainer.

>>43180
Yeah it's great that it's remote and I don't have to do much. It's only a year long contract though and I don't really know what comes after. It just doesn't seem sustainable. The technology itself is terrible as I've already described, and the only reason things are chugging along is because of one guy in his 60s, who started working at the company before I was even born and learned COBOL in college. Ostensibly I'm supposed to be able to replace him if need be, but it took him years to learn the system through hands on experience, which I haven't been getting a lot of.

After trying it, I have to say, Hyrdus's UI is fucking horrible. Oh my god is it bad. Some of the worst garbage design I've seen in my life; it makes me want to claw my eyes out. You have all of these stupid sub-menus and buried options, but nothing resembling the standard interface danbooru established 200 years ago. When you first start, what should you see? What are users used to? A big, centered search bar with a button next to it that says "search", and a nice message saying "you have no files added. Do x, y and z to add some". And why does it call tabs "pages" "Oh read the documentation" fuck the documentation. To find out how to import previously deleted files, which I deleted because I was confused, I had to find some tumblr page explaining that somewhere buried in the options, there's a setting that prevents you from re-importing deleted files which is turned on by default, and there's no option to disable that setting in the import window. Holy shit.

This is how you add tags to a file from a booru. You can't do just this while importing a file for some reason I can't even begin to comprehend(or that option is not nearly visible enough)
(I don't even know if this documentation is official. It's on some other site from the homepage of the project, which has its own docs.)
https://wiki.hydrus.network/books/hydrus-manual/page/file-look-up

>1. Select the files you want tags for
>2. Right-click, share>copy>hash/hashes>md5 to copy the hash values of the selected files into your clipboard.
>3. Paste in your newly copied hash values into Notepad++
>4. Add "md5:" as a prefix at the start of each line. You can do this by Replacing (By pressing Ctrl+H) ^ with md5: or md5= Replacing ^ will add whatever you replace with at the start of each line.
>5. Copy all the lines with the md5: prefix
>6. Paste your clipboard in a gallery download tab of your choosing (booru needs to support md5 searching) by pressing cog button>paste multiple queries merged (or something like that) on the booru download tab. After v336 of Hydrus, this should all the placed into a single queue.

Garbage.

Post edited on 16th Dec 2024, 11:56am

In addition to animal repellents placed on the lawn to annoy anyone who walks past with blinking and high-pitch noise (that's probably going to cause hearing damage since the stuff is certainly not regulated), the latest annoyance I'm seeing in my neighborhood is laser Christmas lights.

Apparently people are too lazy to put up lights but they still want to feel "festive" so the have the bright idea to buy a gadget which projects red/green dots all over the place. Which seems like a decent idea at first, until you realize that they're use lasers to do, and trusting the Chinese shadow company manufacturing this to actually make sure these things are safe. I don't care if they want to burn their own eyes out, but the light leaks from the edges and ends up assaulting me down the street. And I certainly don't want to go blind due to their incompetence.

It's also frustrating to find any information about the safety of these devices whatsoever which makes me more skeptical of them. From what I can gather piecing through AliExpress listings they use a < 100mW red/green laser (or so they claim) through a diffraction grating which should theoretically split the power for any individual beam down low enough to be "safe." Based on some worst-case calculation, even if a single split beam is 10mW, it should be safe for a 2-second exposure at 100 feet. Which is about the distance it hit me at, so I think I should be fine although my eyes do feel tired but I'm not sure if that's related or because I'm developing a fever. (In practice given that diffraction grating doesn't split evenly but makes the outer beams weaker than the central beams I'd assume the edge beams would be < 2mW).

I hate that I can't even walk outside without having to dodge all this stuff like a minefield.

>>43346
Don't worry and happy new year!

>>43347
Thanks, although it's hard to avoid overthinking these things and becoming paranoid.

And I guess another thing I found is that the cheap green lasers are apparently diode-pumped and are converted from infrared. So if these sidebands are not properly filtered then that quoted "100 mW green laser" could also include 100 mW of infrared, so to add a margin of safety in those calculations the total source power need to be doubled from the quoted spec. I guess luckily the eyes are more sensitive to green anyway so in practice the green output tends to be lower than red by about half in order to maintain uniformity of perceptual brightness, so the original calculation should still hold, and I likely haven't been hurt in any way.

Post edited on 31st Dec 2024, 3:53pm

>it's hard to avoid overthinking these things and becoming paranoid
I know. Your rant is a mirror of my own anxieties.

>>43348
Fwiw the eye pain went away, although now I have a headache. So probably just body weirdness. I also did more compulsive research [1] and found that if there was any damage that isn't immediately obvious it'd have to be associated with injuries to peripheral vision. Both an amsler grid as well as the cooler field test [2] came out good (aside from the natural blind spot everyone of course), so it was probably just anxiety. I still think my paranoia is warranted and those shouldn't be legal though because trusting the chinese for safety has never ended well.

[1] https://www.laserpointersafety.com/treatment/treatment.html
[2] https://avtanski.net/projects/blindspot/

>>43363
I'm glad for you! I remember going trough the same anxiety when my kidneys hurt, I'd spend three days researching void to come up with no conclusion at all.

I think I am seriously addicted to listening or watching to Youtube videos. This seems to be a common thing among Hikkineets too, at least among those who are not addicted to video games or something else entirely, but I've heard that a lot of times. I constantly justify to myself why I need to have a Youtube video running in the background, when in reality it's only distracting. Additionally, I associate watching Youtube with a soothing feeling, whenever I'm having bad thoughts.

To be honest, I don't want to stop entirely, because this would mean I couldn't open links I see in chatrooms or so, if I were to block it in my hosts file. I wish I could use it responsibly like a normal person and don't have to deny it myself completely.

>>43368
i actually have a similar problem, though in my case, i just end up spending a while scrolling looking for something to watch because i actually don't like most youtube videos.

a friend of mine suggested listening to albums in the background instead. maybe try that?

>>42674
Did you ever end up solving this? I've noticed that I sleep worse in winter and maybe that's because I leave the windows closed to prevent the cold. Maybe I should just open it a crack and use extra blankets.

Most of my problems and inconveniences in life were just solved by me not appearing somewhere ever again, not talking to someone ever again, or just going away. I don't want to go to school? Just not go there and wait in the local library until school is over! Don't want to go to the afternoon care after school? Just not go there and go straight home instead! I don't want to be in a hospital and leaving it would mean a lot of bureaucracy, that I don't want to deal with? Just wait until they allow me to go home and never return again! Have problems with a certain person? Just disappear for long enough, until they forget about me! The list goes on ad in­fi­ni­tum, but I think you get the point. I feel like this thought pattern of just disappearing whenever I don't want something anymore, has made me decisively a coward and shaped me as a person. You probably think reading this, why nobody ever bothered to set bounds to this behavior. The answer probably is, that when I was a child, and people actually cared, I just disappeared for long enough until they didn't, and now as a more or less adult person, the people just don't care anymore enough to prevent me from doing it.

As good as this might have worked for me, I am still ashamed of myself and I wish to better myself. I don't want to be a coward, but in a sense, the life as a Hikki-NEET is also just being a coward, everything about my existence is and I probably reached a point in life, where nobody cares about me anymore (in any sense of word) and therefor I don't have to run away anymore. I am free now, but at the cost of being a coward forever, because I have no opportunity to change. Of course, I could just make myself the opportunity, like getting into education again, like doing something hard, like meeting new people, but I have fantasized about those things for years and I never made it happen, so I doubt I change my ways any time soon.

I am posting this on this board, because I know how to resolve this problem (by just using another software), but I wanted to complain about it regardless. So, Httrack, this website mirroring tool, just ignores my settings. I set "NO external sites" in the menu, double check if it's actually set, it is set, good, then I continue to download the page and it starts, however, when I check what it's doing 10 hours later, I realize it proceeded to download Wikipedia and some other junk, that I did not ask for. I try to abort it, it doesn't do anything, I try 5 times again and then I can finally abort the procedure. Why on earth does it ignore my settings? I set it that way, I closed it properly and proceeded normally. Why does it do this? I don't know. This is my third attempt at downloading some website and it just won't stop downloading external websites. You might think it's only downloading what's link, but it doesn't, it just crawls the the internet from that starting point and never stops. Speaking of which, why is this function the default? When I want to mirror a website, I cannot imagine a use case, where I want to crawl every link of a link of a link of a link of a link and this without some ending. You would think this is a some unused feature, nobody needs, but it's the default and I can not disable it proper, for whatever reason it is.

Screw Httrack. Long live Wget.

Post edited on 16th Jan 2025, 2:21pm

I bought an HP Z2 G5 sff off ebay for $270, and have been in the process of setting it up and upgrading it. Thankfully, I don't intend to use this system myself, I'm giving it to someone who currently only has an old laptop.

There's a lot of annoying shit. The PSU it came with is only 260 W, so it's not enough to put a GPU in, even the 70 W one I have. That's an extra $50. It's some proprietary type with non-standard dimensions and connectors, so you can't use a 3rd party one without a lot of hackery.

The CPU cooler is not very good, and besides the PSU's, it's the only fan in the whole system. The mounting hole placement is standard for LGA1200, so it can work with a 3rd party one, but the case has these raised bumps the stock cooler screws into instead of the motherboard back plate. You need to either flatten these bumps, or use some screws meant for AM4 coolers, and widen the holes on the intel cooler so they fit.

The motherboard has another 4pin fan header, but for some reason, they don't work well with Noctua fans. They run waaay too slow, and you can't change individual fan speeds in the bios. So I need to get Notcua's fan controller.

The ram does not work at full speed. The CPU supports up to 2933MHz, but their speed is capped at 2400MHz. This is probably because XMP cannot be turned on in the BIOS, and Intel's overclocking utility doesn't work except on specific CPU and motherboard combinations. If you want full speeds, you have to buy HP's approved ram models. This is apparently even a problem on HP's "gaming" Omen line.

Unless you want the cheapest possible option, avoid these OEM systems and build it yourself.

>>43420
I was in a somewhat, but not really, similar situation >>/pic/7439

He's an asshole and you should move on. Get another hobby. Within a month or two, you wont even miss it.