off-topic-tech

[night girder]

That's going to be fun. You ever been to a sleep center?

[visual tree]

No, only to a store selling beds and mattresses. I am also 100% sure it's sinus issue

Pills only help for 3 weeks

[visual tree]

Although I have to admit I sometimes have restless leg syndrome

It's like my legs are heavy as cement and I can't sleep because of that

[night girder]

Here, they monitor you a full night with sensors and all. But it's a bit surreal.

They want you to relax and act like it's natural. But it's a hospital and you get all these sensors.

So sometimes, you have to come back multiple times. So they can get a reading on you in the most natural habitat possible.

Else you get manipulated/false data.

And you have to fall asleep

Maybe they give you something to fall asleep, but again it depends on the sleeping issue you have.

[visual tree]

Just make me run a marathon and I won't need pills for sleep

[night girder]

😄

Weirdly, I had times that I was so tired, I couldn't fall asleep 🤣

Like, really really tired, but my brain is just ... not calm.

[visual tree]

It's funny how you get less tired if you only sleep for 3 hours assuming you fell asleep really tired

And if you fall asleep for 7 hours while not really felling tired before going to bed, you feel more tired when you wake up

[night girder]

overstimulated brain, that's the word I was looking for.

[night girder]

Nah. That's different

You have these different stages, and depending in which stage you wake up.

You feel tired or fully awake/energized etc.

So it's possible to wake up after 3 hours and not feeling tired. But it's also possible you don't feel good at all.

And there are multiple sleep cycles.

but you see, the 4th cycle doesn't go into deep stage anymore.

During REM stage, your eyes also move.

https://tenor.com/view/nia-nal-dreamer-nicole-maines-sleep-gif-14255164

That's fake, but you get the gist.

[visual tree]

I did measurement with a galaxy watch once and I got this:

[night girder]

eh ... galaxy watch vs scientific machines 🤔

it will be in the ballpark I guess, but the error margin must be higher

[visual tree]

Btw, I am using galaxy watch in watch mode only (using it atm because my classic watch broke down)

At first, I was like, lets see how good those smart watches really are

Gave up after few days because it drove me insane lol

[night girder]

Weirdly enough, I never understood the smart watch.
Or what it had to offer, never convinced me.

[visual tree]

It was like ping...ping...ping... every few hours. Felt like someone micromanaging me

[visual tree]

Don't buy it, it's really annoying imo

[night girder]

Like, I have my phone in my pocket. I get a text message and it shows up on my wrist. woopdydooo.
Or I take the phone out of my pocket and look at the text message.

And the same for the time, why have the time on you, with two different devices?

[visual tree]

Mine was pinging me every time I moved because I was measuring my steps

[night girder]

But what's the advantage?

It just looks like a gimmick.

Or people really that lazy to take their phone out of their pants to check the time, mails or text messages?

[visual tree]

I need a wrist watch of any kind because I am lazy to pull a phone from my pocket

[night girder]

Fuck me.

I am the person who automates everything if I have to do it twice.

But even I think smart watches are a bridge too far 🤣

I rather buy this. It looks ... good.

[visual tree]

I love the mechanics of an automatic watch but also have having to rewing them manually or with movement every 2 weeks. I prefer quartz watc

[visual tree]

You'll need to sell a kidney for that brand 🤣

[night girder]

than this

Like smart watches look like the gum watches I got as a kid.

[visual tree]

Oh, I had those

[night girder]

Yup! And even they had extra purpose than a smartphone.

Candy!

But build quality looks about the same 🤣

[visual tree]

I once used a calculator in elementary school which had a larger display and you could store notes on it

Guess what I used it for 🤣

[night girder]

drugs.

[visual tree]

No lol

[night girder]

oh wait.

we not talking the watches anymore 🤣

[visual tree]

Found the image but it's not the same brand:

@night girder This should give you a hint what I used it for

[night girder]

We did the same trickery

[visual tree]

Math teacher never suspected lol

He thought it was just some calculator with extra features

[night girder]

Ours knew for sure. Eventually, we got a seperate part of the test where we could use the calculator. So 90% was without calculator.
And when done, we had to give that in and get a new paper we can do with calculator.

So that 90% you did without calculator, you couldn't technical cheat with calculator.

[visual tree]

Our math teacher made test questions from a book he held on his desk. Students eventually figured that out and bough the same book lol

And the teacher was like, I am so proud of you

[night girder]

reading AMD patchnotes is scary.
Things I wouldn't never think were possible, seems to be possible. Else why fix.

[sharp matrix]

amd patch notes are becoming terrifying💀

I mean its good to see fixes for these things but now we know more things that can happen unless its fixed....

Which now raises more questions, what else could lead to system crashes with amd drivers, its a scary thought

[night girder]

right?

And Nvidia is probably as bad too.

Also, application freeze when using overlay. It's their own goddam overlay 🤣

[mental oriole]

Mine would be like 30min-2h awake time easy.

[visual tree]

You hibernate during winter?

[mental oriole]

I'm dead during winter

[night girder]

Less sleep more coding!

[mental oriole]

More like

[night girder]

Btw, maxeeks said we need to work together diluted demon.

[mental oriole]

1. Light sleeper
2. Any sound makes it harder for me to fall asleep
3. Idk sounds makes it super easy to wake up
4. My alarm is super quiet yet I wake up to it.

[mental oriole]

[visual tree]

I had to move my NAS to living room because the fans are annoying and I also turned off all led lights I could

[night girder]

Yes, you make engine. I code game

[mental oriole]

I die inside

[visual tree]

I like sleeping in complete dark

[night girder]

My NAS turns of at 0:00 and wakes up at 9:00

[mental oriole]

Now I go sleep.

[night girder]

I whish I could help you with the engine. But I don't think I have the technical skillset for coding an actual engine.

[mental oriole]

I also have the visual snow thingy where things looks pixelated/ grain filter 🙃 It looks very funky in the dark

[night girder]

It's a feature

[mental oriole]

Maybe after new years. My todolist grows for every day 🤣

[night girder]

Yeah I know that feeling.

I add more todo's each day than I tick off.

[mental oriole]

I know that feeling 😄

Now I'm gonna drag that sleep jira back to work in progress

[sharp matrix]

Make a engine from scratch, sounds like somebody is going to coding hell

There is a reason most game studio's use a existing engine, because coding a engine from scratch, as it turns out that shit isnt simple

[night girder]

where was you a few days ago. I was making the exact same point to someone 🤣

[sharp matrix]

I'm sleeping during winter, this is the period I briefly wake up

I have some experience because back in the day I thought it would be fun trying to code a engine from scratch myself, little did I realize the hell that was awaiting me

I mean sure like any problem you can work on it and make progress if your smart and motivated enough, but boy coding a engine from scratch is just its own beast all together, its basically faster and easier 99.9% of the time to just use a existing engine, especially if your wanting to get even a quick game idea prototype off the ground

[night girder]

it just really depends on your goal. For a solo developer, if your goal is writing an engine. Then fine.
But if your goal is writing a game, I wouldn't recommend writing your own engine as a solo developer.
Instead it's more efficient to implement a solution like an engine etc.

For teams of a bigger size, they have more options.

[pure karma]

yea iv been trying unreal engine for myself and the blueprints seem so abysmally awful to use that i already lsot all motivation to do anything before even starting

everything seems so overcomplicated to do just simple things

[sharp matrix]

oh ue blueprints, i took one look at those and gave up

my brain can understand java/c/c++, but its just over with ue blueprints

[pure karma]

yea idk what they were thinking but its just so unintuative

[sharp matrix]

yeah for me coding is much more intuitive, i really just cant understand blueprints

[pure karma]

it seems they tried to make there own version of coding with blocks and then forgot the target audience dosent know what made up bullshit + made up bullshit is equal to

[sharp matrix]

at this point I would rather code a engine from scratch than try to understand ue blueprints

[sharp matrix]

yeah i get that was the idea, but it seems to horrible in implementation

the way they tried to make coding in blocks, while i can get why they might have wanted to give it a try, my brain just look at how horrible it all is, even worse than c++ spaghetti code, because it becomes so horrible to follow and understand past even simple things, that im just like nope that isnt it chief

I think what we have with ue blueprints is what could have been a cool and maybe even a good idea, one that is just terrible in implementation

[languid gulch]

i'm waiting for something in nvidia or radeon to massively break, and it's because they tried to sneak in an unlisted patch or change and it blows up in their faces

[jagged snow]

Well screw google
Locking app management behind the play store and the play store is crashing on my phone 💀

And because it's a "system app" you can't force quit/clear app data etc

(already tried a full device restart, have no OS updates avaliable)

[twin dew]

Some pics of laptop LCD screen.
Electronics on backside, topside, just the actual LCD element, backlight lightguide and backside film, actual single-edge LED element, and the various filter films between lightguide and LCD element.

[bold tapir]

Did you take it apart just of interest

[twin dew]

Just for those pics.
I had taken some apart before.
Took the films after and put them away, rest is going to recycling.

[bold tapir]

What the films? Im not into tech terms

[twin dew]

Those four in the last pic.
Two polarizing filters in middle, light-spreading one closest to the backlight lightguide, and then another closest to LCD.

[bold tapir]

Oh polarizing filters, i know if you get two you can change transparency by rotating them

[bold tapir]

I dont know english good enough to explain this properly

[quick estuary]

Guys is it a good time to buy a pc rn?

[quick estuary]

Asking for a friend

[glossy glacier]

If you already have RAM yes
Otherwise no

[bold tapir]

Always good, just get used for nice price

[quick estuary]

He bought it before the prices went up so ye ig

[dire igloo]

Depending on the RAM, you might still need to buy new

If it's ddr3, basically unusable.
With ddr4, you're either locked to am4 or lga1700, both not really attractive either

With ddr5, you're probably fine but depending on what he bought, he may need to put in some extra work to make it stable

Other than that, I think now is a better time to buy than the first half of 2026

[dire igloo]

Even without RAM, I honestly believe that it's actually gonna get worse until like April or May 2026

[twin dew]

The raw die costs have only about doubled, while the consumer prices are currently tripled.

[pure karma]

yea i checked yesterday and my ram went from 159.99 when i bought it to 589.99 now

[dire igloo]

Could be greed, could be preventive pricing to avoid low stocks

[dire igloo]

You shouldn't compare individual items to get a feel for a general trend

Hence why I use this
https://pcpartpicker.com/trends/price/memory/

[twin dew]

Overcorrection is to be expected when the situation is in flux.
But by that, the consumer prices might go down in future, or not.

[dire igloo]

price trend across all 2x16GB DDR5-6000 kits on US PCPP

[twin dew]

But basically there is pretty high change that DDR5 consumer prices will drop somewhat in next few months.
And that is visible in some of those PCPP charts where they have at least momentary drop down at the end.

[night girder]

Whaaaaat? Companies are leaving consumer market. Amidst a crisis in chips, to opt for dataccenters.

So super normal for prices to skyrocket for us plebs. I said it a week or two ago. Just funny how we still act surprised.

[pure karma]

yea they wont drop there gona disapear instead

[twin dew]

For example DDR5-5600 2x16GB:

[twin dew]

Point wasn't that they would drop back to what they were, but less than current 3x price spike.

[pure karma]

im in the 6400 2x16 ship which is always the ignored middle child somehow

[night girder]

I said that too. Super normal in my eyes.

[twin dew]

Closer to 200-250% of previous, than current 300+%.

[night girder]

👆

Micron, Sky Hynix, all left consumer market to focus on data centers.

[dire igloo]

I'm inclined to assume that that dip is from some $1500 kit going OOS

Cuz the rest of the distribution seems to still go upwards

[night girder]

Bad time to pick up hardware as a hobby

[sharp matrix]

My current system has 64gb ram, im good on memory I think for a while, so I can basically ignore the price spikes

[night girder]

And amidst economical wars. Nah, we're fucked for a year or two.

[twin dew]

Point was that we would still be fucked compared to before the spike, but less than right now.

[sharp matrix]

I was prepared for things to be fucked for two years

[night girder]

And I keep hearing this "it will drop back" in this channel while I keep reading news article that claim the opposite.

[dire igloo]

Who's saying that we're acting surprised?
The only people that are still being surprised are the ones that didn't follow the market (which tbf, includes a lot of people)

[night girder]

I don't understand why Baldur thinks it will drop back, knowing he must be reading the same news articles as me.

[twin dew]

I was saying I expect the consumer RAM to stabilize to costing about 2.0-2.5x it was before, while we are currently at 3+x.

[dire igloo]

SK Hynix didn't leave entirely, no?
Latest news I heard was they just have the large majority of their production tied up

[night girder]

Like two of the biggest companies leave the market, that's the opposite of "oh it's calming down now"

[twin dew]

Because the die prices are stabilizing to around 2x they were before.

[sharp matrix]

I see no chance of things really dropping, companies will likely milk for as long as possible, then move to datacenters for highest profits, obviously consumers are screwed for a while that much is clear I believe

[night girder]

👆

[dire igloo]

Assuming the situation doesn't worsen.
Even more datacenters, another water shortage, stronger trade wars, whatever

[sharp matrix]

only real question is how screwed its going to be and for how long, but I would go for 1-2 years of consumer electronics, pcs, ram, etc being royally screwed for some time

[night girder]

That was posted by Hynix.

I don't see how that will help cooldown market.

It only puts more pressure on it, in favor of their own data centers.

Meanwhile the effects are like a ripple effect on water and is spreading to other products too.

[sharp matrix]

it wont, this market is screwed for awhile based on the news im reading

[night girder]

👆

Yeah, at this point, I think it's pretty realistic statement.

[dire igloo]

Baldur's theory, explained simply:

IC makers: "we're booked for 2026"
RAM makers: "oh shit, better increase prices before our stock depletes"
(prices jump to 300%)
later...
RAM makers: "nvm, shortage wasn't so bad, we can turn it down a notch"
(prices decrease to ~250%)

[night girder]

I am still waiting for the Nvidia GPU's to drop back to pre corona prices like people claimed 🤣

[twin dew]

Yeah, but I would call 2x previous pricing with supply problems as pretty fucked too.
Not just 3x previous pricing with supply problems.

[night girder]

oh wait, the prices are still sky high for their GPU's

Yeah, right ... So 🖕 consumers when they have it good. And then begging consumers to buy their crap again when they need us again.

[sharp matrix]

well yeah we all are reading all of these companies basically exiting consumer space to focus on data centers only, consumer market is screwed, there is noway around it

[night girder]

Consumers aren't a light switch you can turn on and off that easily. And once they know they can get away with selling hardware for insane prices.

Why lower the prices? If idiots buy them, more proffit. Look at GPU market pre corona, corona and after corona.

I don't think it's going to happen, sorry. I hope I am wrong.

[night girder]

You know what I find more amusing?

That a few articles after I read "everyone bailing consumers to focus on data centers" - there was this article that said that one data center consumes as much ⚡ as 2 million households.

meanwhile, my car has to go electrical to safe the environment etc etc. I just don't get it how those two rhyme together 🤣

[sharp matrix]

Yeah its crazy the world we live in, save the environment, meanwhile we have to win the ai race in the us, regardless of the electric, water, environment cost, etc

[night girder]

If we have 100% green power, fine, build all the data centers you want, but that aint the case.

[sharp matrix]

I mean if this isnt the definition of a joke I dont know what is

[night girder]

I knew you would understand me Slechtvalk 😄

[sharp matrix]

apu two standards, one for us consumer plebs, and the rest for corporations and all of their previous data centers that is perfectly fine for them to destroy the environment apparently

[night girder]

Also, the datacenters in the Netherlands that Google and Microsoft wants to build, might be in trouble.

Because they are below sealevel, which seems to be a no-go in the Netherlands. If you know the country, you know why.
So high chance they need to find a new place to build them since it's critical infrastructure.

[sharp matrix]

God I hate this world we live in sometimes and the multiple standards that seems to exist for each different group....

[night girder]

Yeah, I don't understand how they expect me to keep up with all these double standards.

Or maybe that's why I don't care anymore 🤣

[sharp matrix]

I wonder if anybody will care even as the sea levels continue to rise🥴

[night girder]

Netherlands is filled with embankment since they always had issues.

[sharp matrix]

oh well guess its back to the regular schedule for today of fuck the environment because of whatever bullshit race we have to win this year because of whatever retarded reason the politicans decided to pull out of their ass apu

[night girder]

So they try to prevent critical infrastructure being wrecked by storms, tsunamis etc. Learning from Fukushima etc.

[sharp matrix]

God all of this shit is just so dumb.....

[night girder]

It just doesn't make sense. Well it does. But it's morbid.

Money > all the rest.

[sharp matrix]

I wonder if the plan of the rich is to leave once they are doing fucking this planet over and have all of the money they could ever wish for

they even made a movie where basically thats exactly what happened i believe, which was kind of funny

only the super rich survived, just to go to another planet and fuck that one up to, repeating the same mistakes

[night girder]

Nah, they don't even think that far a head. They know whatever problems they cause now, it's the problems of the future. They long dead.

[sharp matrix]

Of course makes sense, in a morbid screwed up sort of way

[night girder]

That's why some started to coalmine again etc. You get instant money. And by the time the planet is full of craters, the people making those descisions long gone 🤣

[sharp matrix]

How long until the strip mining starts again to

because fuck the environment lol

[night girder]

"And so what if the sea rises? Not my problem 🤷‍♂️ " - it's a weird way of thinking. But it makes them sleep better.

[night girder]

already start in Germany. They wanted to close nuclear power plants and they just did. Then stuff happened and their gas supplies got cut out. And now they back to brown coal.

[twin dew]

Has never stopped.

For strip mining.

[night girder]

well it's on the rise again. Because of what I stated.

And not only germany.

I wonder what happened in 2020

[dire igloo]

Worse: they even dug away villages and forests and continue doing so

the Monk of Lützerath was literally a protester against brown coal

[night girder]

I heard that story last week on a podcast (again) 🙁

[night girder]

As a result of cost pass-through and portfolio restructuring, we expect average selling prices (ASPs) to also increase next year by 6.9%, revised up from 3.9% in our previous ASP forecast released in September 2025.

Like I said, ripple effect 🤣 Now smartphones will probably be 7% more expensive next year.

counterpointresearch.com/en/insights/2026-smartphone-shipment-forecasts-revised-down-as-memory-shortage-drives-bom-costs-up

[mental oriole]

Everything will be more expensive next year what a surprise

[night girder]

Indie games next year; 80 euro. Tripple A games, 160 euro

[mental oriole]

cant wait for 9999 euro

[night girder]

9 billion million zillion euros 🤣

Time to ditch firefox all.

if you hadn't.

new CEO has big plans. And amazing, out of the box ideas, like adding AI.

source is their own blog: blog.mozilla.org/en/mozilla/leadership/mozillas-next-chapter-anthony-enzor-demeo-new-ceo/

I should start looking into librewolf maybe.

You liking librewolf so far?

[mental oriole]

Been running librewolf as main since I swapped to linux, it works fine.

[night girder]

Do you combine it by any chance with https://github.com/yokoffing/BetterFox ? (if that even is a possibility?)

[mental oriole]

I have not modified anything

[jagged snow]

Yes, actually!

I tweaked a few of the defaults(mostly just whitelisted cookies for a few sites at shutdown so my signins persist for sso)

It's been great otherwise

The one issue I've had is that the auto-updater script uses ahk and it's going to get me anticheat banned one of these days

But no AI features, no tracking, basically just works as a browser should

I really don't like brave/opera and all the bloat they bring, so librewolf is the best option out there at the moment IMO

[bronze jasper]

codeberg.org/librewolf/issues/issues/1919

not a link cause codeberg isn't whitelisted. i found this though. looks like librewolf project owner doesn't care that hard about ai features.

[night girder]

If this is true, that's a dealbreaker to me.

[night girder]

🤨 I read the opposite.

[bronze jasper]

He's right there talking about doing it with preferences

[jagged snow]

I have been following this
My impression was basically that they just won't have it on by default and they won't be scanning your content

[night girder]

Dec 14, 2025, 09:11 AM

[bronze jasper]

ahh he's shifted his position there. will note

[jagged snow]

You can just do manual updates thankfully

[night girder]

Oh, ok. Big fan of manual updates.

[bronze jasper]

i personally find LLMs are extremely useful for browsing. helps cut through so much bullshit if you use them right.

[jagged snow]

Nice
Yeah, they just provide an optional AHK script that will notify you and gives you a one-click option to install the llatest update

[jagged snow]

I think that gets more true the more AI slop there is out there on the web

Browsing with a traditional search engine nowadays is extremely difficult because of the sheer volume of garbage you have to wade through

[bronze jasper]

and i think firefox is moving towards a local model approach in the long term. supporting smaller models that run locally instead of in the cloud

[jagged snow]

Great, let me know once they do that
And let me know when all the ethical issues about training are figured out

[bronze jasper]

the web is already slop imo. we're long past that tipping point. adwords and SEO got us there

[jagged snow]

Agreed on that point.

There are still gems, they're just harder and harder to find

[night girder]

I read the message from the CEO as a big pleaser to investors.
I don't think it's even aimed at the users. It's just pleasing investors by writting that blog post.

[jagged snow]

adding a before:12/2020 filter helps a lot

[night girder]

And the moment the CEO is a investor pleasor, we, the users are fucked.

[bronze jasper]

the ethical issues have been figured out. training a model with content is fair use and considered transformative. courts have ruled on this.

The hiccup is that anthropic and other companies too, used pirated content. Not scraped, but literally a popular archive of literature distrbuted on piracy sites. So anthropic had to pay the largest copyright infringement settlement ever because of that.

[night girder]

Before that Mozilla was one of the few holding on to values, to their userbase etc.

[bronze jasper]

i'm waiting to see how this class action moves forward as they sue other companies for pirating the same archive

[jagged snow]

Wow, that's a wild take actually

[night girder]

But they are slowly turning into a profit big tech company like the rest of the 💩 (Google, Meta, Apple, ...)

[bronze jasper]

that's not my take. that's a judge's ruling

[jagged snow]

You just said that the courts determine morality & ethics, I strongly disagree with that.

Courts determine legality.

[night girder]

I don't understand why you want an AI browser.

[bronze jasper]

i didn't actually

i said something else and that was your hot take about what i said

[night girder]

Even my calculator has AI into it now. Paint has it. Whatsapp has it. AMD has it.

We going to have AI bloat for years and years to come.

having AI baked into every application they can, when in my head, only a few would be enough.

[bronze jasper]

I could explain it but i'm sure thats not what you want to hear

[jagged snow]

I said "let me know when the ethical(not legal) issues have been figured out" and your response was that the courts have figured it out. Imo, directly profiting from millions of others hard work without consent and doing so in a manner that diminishes the value of their work is not an ethical business practice.

[night girder]

You can give your opinion, I think I just wont agree with it since I read it before 😉

And the reason I used AI a lot recent months, is so I could experience it and form my own opinion.

I have to agree with what most already told in here, it's dilussional, making up fact bullshit. 10-20% it helps me. 80% of the time it sends me on a wild goose chase where I end up wasting WAY more time than I gained by using AI.

Reading documentation, asking real people and google/duckduckgo get me further. But like I said, there is this 10-20% chance it does kinda help me. But totally not efficient for my workflow.

[bronze jasper]

i wouldn't want llms baked in everywhere. i dont like that aspect of how microsoft is rolling it out, but i don't mind it being an often forgotten feature buried in the edit menu.

browsers just makes sense for it though. I can understand the aprehension in a few different ways personally, but i'm not about those reasons. LLMs help filter content and i'm lazy.

[night girder]

Now if anyone would say; "we going to implement AI as an addblocker in a Browser" ... now that would peak my interest 😄

but investors will raise hell if they hear that 🤣

[mental oriole]

Easy, just block everything that's not a link to the current website domain. 99% of ads blocked, as well a bunch of other functionality.

[night girder]

I understand it, I can only speak for myself and say I don't have a need for it. Doesn't mean that I want to get rid of it. But feels like we going from 0 to 100 and all browsers want AI. And that just rubs me off. Imagine a world where a few browsers said: 🖕 and a few went AI 🤖 and they lived happy and forever next to eachother.

[bronze jasper]

courts often consider ethical reasons in their rulings and also the ethical arguments are shaped by legal precedence and what's considered appropriate. Part of the ethical delima is that court's hadn't figured out where the law fell on this matter.

Fair use has existed for a lot longer than AI has, and it's not a cut and dry situation. Things must be examined on a case by case basis to determine if it is fair use or not. In this case, it was ruled to be fair use.

So now the ethical discussion can proceed around that ruling. If models are fair use, then we're still worried about how content is gathered right? Turns out , just like before AI, Piracy isn't considered ethical. So they were punished for doing that.

there's never a black and white answer to ethics. I have to wonder if those complaining about ethics the loudest have even studied the academics behind ethics at all. There's a lot of "so what you're saying is...." going around which is not an honest conversation.

[night girder]

I don't know.

It's like a mass murder has been commited. And now they going to investigate if a mass murder is commited.

The murders have been commited. They should have been prevented.

If they stole IP without paying the authors to train their AI, it's too late and too little. We can't go back now.

[bronze jasper]

copyright infringement is not mass murder. i abhor that analogy. it sounds more like corporate propaganda from the 90s than anything real.

"you wouldn't download a car would you?"

[night girder]

You are 100% correct. But I am trying to describe it.

The fact remains, the crime has been commited.

Like no way AI/LLM gained that much knowledge AND paid everyone that provided the data. Every author of every book etc.

In the beginning, it was a wild west, and AI companies fully knew what they were doing wasn't ethical.

And now law is trying to catch up ... way too late.

[bronze jasper]

they settled on 1.2 billion dollar settlement for all the authors. something like 10g each author in the pirated archive. and that's just from anthropic. Other cases will move forward against other companies, and they potentially won't be able to settle for the same amount.

[night girder]

It's not correct/fair AI music generators are trained on real muscicians hard work, without them getting paid for it.

[bronze jasper]

When this case moves forward against Meta and Open AI, we'll see bigger settlements for them likely. i doubt that the class action will settle behind closed doors on these matters

[night girder]

it's so wrong ...

So if I steal, a while library, literally all the books from a library. And in a year I get caught.

Then I can just pay for the library and not go to jail? Even thaugh... stealing is illegal?

[bronze jasper]

RIAA is sueing suno right now. That should wash out in their favor

[night girder]

Or if I just copy ALL the books from that library. And I get caught. I can just settle and not go to jail.

That's just wild.

[bronze jasper]

Copying something isn't stealing. It's copyright infringement and has a different legal framework

[night girder]

What do you think the AI is doing?

[bronze jasper]

that 90s corporate propada about pirating tv shows and movies convinced everyone that copyright infringement is stealing. it's not though. It's literally just anti piracy

[night girder]

Well, don't blame me.

Goverments have been telling us* copying is stealing for years.

But nonetheless, it's illegal.

[bronze jasper]

governmetns never have no. law has always treated infringment and theft differently

[night girder]

Well, the adds were literally made by a goverment body here.

I think it was called "brein".

https://nl.wikipedia.org/wiki/BREIN

Ok, seems it's a private organisation. That raises more questions

[bronze jasper]

I don't read dutch but that says "is een private Nederlandse organisatie" which seems suspect

[night girder]

Your AI browser isn't translating? 🤣

[bronze jasper]

im in discord client and i woudn't just have it turned on all the time. its a feature i would want and can see value in. not one that i am proficient at using yet.

you seem to have this fake idea you've decided about how i'm using AI for everything all the time and that's hilarious to you. If that's the tone of this conversation instead of actually treating me like an equal, we've reached the conclusion of it's value.

[night girder]

It was just a little joke 👉 👈

Because I though a tech savy person would use a translator, instead of saying "I can't read that."

Sorry if you felt attacked by a joke.

[fringe pecan]

Would anyone wanna play wit me, I got no friends fr cause adulting life

I play a game similar to this so I got a good understanding of what to do

[jagged snow]

Sorry, wrong channel

#1201555265942724758 is what you want

[sharp matrix]

Reminds me what chatgpt and other models did, stole copyright to train their ai slop with and because too big, they basically got away with it

Its such a amazing world we live in, become big enough and not even copyright rules apply anymore

I sure do love waking up in the morning to the smell of how many double standards now, im starting to lose count

[night girder]

Yeah, agree

[jagged snow]

Older low end system
Would run most modern games on low at 1080p

[twin dew]

Which completely breaks about 50% of the current larger sites.

Sites loading critical javascript from other domains.

[mental oriole]

Thought that was implied by my statement...

[mental oriole]

Ofc this crap cannot be disabled...

[dire igloo]

DRAMless SATA SSD that's suspected to be QLC-swapped. Wouldn't trust it to remain issue-free for too long, will also struggle with sustained writes.
For the rest, nothing that stands out too poorly.

Agree with Fight or Flight here, might even be able to do medium in decently optimized or older titles.
Depends on your fps target of course

[visual tree]

https://www.reddit.com/r/technology/comments/1pmh05h/samsung_to_halt_sata_ssd_production_leaker_warns/

I feel like next 2 years are going to be MUCH worse

[mental oriole]

Just wait until it also includes nvme.

[glossy glacier]

tbf I can see why they would do that. How large is the sata ssd market even anymore?

Probably 99% NVMe by now

[visual tree]

Don't tease the devil

[mental oriole]

Too late

https://tenor.com/view/burn-gif-20530397

[visual tree]

👀

Btw, I just realized car warranty is one huge scam

[mental oriole]

[visual tree]

If your phone breaks down, warranty covers new parts and repair cost and you don't pay anything

If your car breaks down, warranty only covers car parts but you have to pay repair fees which are even higher than car parts

[twin dew]

In Finland the work is also included for actual warranty cases.
But of course the companies try to wiggle out of things being in warranty and not just expected wear parts.

[visual tree]

That's how they steal your money even though car is under waranty

And I always thought warranty covers everthing under the assumption the user didn't damage anything....

[twin dew]

Because the car manufacturers have strict times for each warranty work thing for hours the mechanics can work on the issue and get compensated.
And those times are way too strict basically.
As they determined by how long experts take for the work, after having done it already few times and with everything ready, on new car so everything is easy to remove etc.

So where you live, seems that the companies double dip.
Get work coverage from the manufacturer, and the customer also pays.

[visual tree]

Our car intercooler broke down after only 3 years even though we rarely used the car. We still had to pay €380 for repair costs....

[twin dew]

And like I said, that was the specific repair shop/car shop fucking you up and double dipping.
Charging for things that should have been included in the warranty work.

[visual tree]

It was a licenced Volkswagen repair center

[twin dew]

Yes? So?

Only those can even do warranty work.

Contact Volkswagen direct and ask about that charge.

[visual tree]

Because people here say that licenced repair centers are licenced because they follow the rules and won't rip you off while non-licenced repair centers will 100% rip you off and that's why they call them "non-licenced"

That is a HUGE lie

[twin dew]

Warranty shops trying to put on extra charges to the consumer are long time tradition.
Because the warranty work actually costs them more than the manufacturer covers.

Even when they aren't allowed to do that.

But most people will never actually try to do anything about it.

[visual tree]

At this point, I should sell the car and use public transport, cars only pollute the environment and make a hole in your wallet 🤣

Or use those daily rent-a-car when needed, they are much cheaper

[twin dew]

But the point is, when the warranty shop tries to fuck you up with extra charges, you contact the manufacturers organization in your country and ask WTF.
As THEY have the power to slap the warranty shop for those.

[visual tree]

My mother did that to Seat repair center and they lost licence for that. Glad we didn't get sued lol

They were really pissed

[twin dew]

But it is fully on them.
Either play by the manufacturers rules, and cover the loss on warranty work with the overpriced yearly maintenance visits like expected, try to renegotiate the warranty work coverage, or don't be warranty work shop.
Don't try to put extra charges on the consumer against the OEM agreement you have.

[twin dew]

Well, doesn't really cost.
Usually it will cover the actual salaries of the people doing the work.
Just losing profit as it will not hit the same amount as same amount of hours done and billed at their standing hourly rate.

[obtuse geyser]

Anyone here using the latest lineageOS?
Just upgraded my phone to modern one, and I am off-put by the big ugly wide tiles in the quick settings menu. My old phone had those as small round icons and could fit much more of them in single row, now I have to scroll twice to get to fifth option because it only shows 2 rows (4 tiles) total by default. Does lineageOS have the small tiles option? Can't find definitive answer on the nets.

[twin dew]

Base Android change, so probably not.

[obtuse geyser]

Would appreciate someone with lineage to give a definitive yes or no.
Your "probably" still has more value than gpt yelling "of course, yes" 👍

[twin dew]

Very unlikely to be available for any Android fork based on Android 12-15.
11 and before had the small icons.
16 allows to customize the size between 1x1 and 1x2 for each icon.

So LineageOS 23.

But that screen doesn't match to base Android 16, so you are probably on some LineageOS version between 19.1 and 22.2.

[twin dew]

So which version are you actually on?
Because depending on device it seems to be either 22.2 (Android 15) or 23 (Android 16), for the officially supported phones.

[obtuse geyser]

Oh, customizing between 1x2 and 1x1 would be enough for me. My device is currently on android 14 (stock), the screen is just random photo from internet that shows how it looks on mine currently. Was debating getting lineageOS on it, if it solves the icons problem. supposedly 23 is supported (moto g84). Latest official build available now is a15 with 16 being promised later on.

[twin dew]

So yes, Android 16 (and almost certaily LineageOS 23 as it is based on Android 16) supports that icon size customization on per icon basis.

[obtuse geyser]

Cool, thats good enough. Now the question is, do I wait for the official motorola update to 16 or go ahead with lineage23. I'm currently in setup phase of new phone, didn't even start migrating my apps and data (it may have to wait after the december rush). So replacing the os before I start actually using the phone could be economically viable, just mildly scared of bricking it, since I havent touched android installation for years.

[twin dew]

"Nice" take from Acer. Older laptop.
Acer has removed all drivers from their site, and had never pushed them to Windows Update either.

Even when the laptop came with Windows 8 when new.

Needed to scrounge the various "All the drivers" sites to find few, for devices which didn't work with the stuff from WU.

Using touchpad without two-finger scrolling was so annoying...

[mental oriole]

Discord is broken on linux now 🙃

[night girder]

Or is Linux broken with discord 🤣

[glossy glacier]

Why?

[mental oriole]

Have chat box selected.
Then from steam, hover your profile name, and then hover a button (e.g inventory) and try type...
the entire discord client freezes for me 🙃

Also got this sent to me.
https://wccftech.com/after-gobbling-up-dram-nvidia-sk-hynix-plan-to-introduce-an-ai-ssd/

[mental oriole]

And discord is the only app that suffers from this problem so I am fairly certain this is an issue with discord.

[jagged snow]

Random commentor is right, we need pram 😄

[mental oriole]

yee

[night girder]

The population of the Netherlands is approximately 18.4 million as of December 2025.

[jagged snow]

10%

[night girder]

With a premium account.

It (the hack) also includes watch history

[jagged snow]

Oh, so there's about to be a big blackmail wave over there

[night girder]

Well, article says it's unclear. Could be that pornhub is negotiatign to pay. Or not. Nobody knows.

[jagged snow]

Gotcha

[night girder]

But this is an example of the data they have. They say. Haha 🤣

[jagged snow]

Yikes

[night girder]

It even tells you the village. But that's probably where main node of ISP is. Or the server of the ISP.

wonder if it's the only country hit or it's a world wide one.

Phub itself didn't get hacked, but a platform they used called Mixpanel for analyses.

[glossy glacier]

it deselects the text box in discord when i hover over the steam library, but otherwise workks fine

[night girder]

I had discord hardware accelerated issues. Really weird ones.

Like the whole chat, kinda duplicated into one message. So for example, this message here, I could scroll in and read the whole chat of this channel again.
It was crazy. And a few other issues. But once I disabled hardware acceleration for discord it didn't happen anymore.

like inception, but a channel, within a message 🤣

[jagged snow]

Oh, I think I've seen that

My laptop has issues with discord & hw accel where only part of the screen will refresh for minutes at a time

I'll scroll up and down in the client and only the right side of the screen will change

[dire igloo]

good thing NIS2 now introduced supply chain protection

[night girder]

I am trying to find more information on the supply chain protection. Details.

Member States shall also ensure that, when considering which measures referred to in that point are appropriate, entities are required to take into account the results of the coordinated security risk assessments of critical supply chains carried out in accordance with Article 22(1).

critical supply chains

[dire igloo]

www.bsi.bund.de/SharedDocs/Downloads/DE/BSI/NIS-2/nis-2-lieferkette.pdf?__blob=publicationFile&v=3

[night girder]

I am looking for something specific.

[dire igloo]

you won't find that

because it's not regulation, it's a guideline that member states have to turn into national law

[night girder]

1. Is it a law or recommendation.
2. If it's a law, what about countries non EU, for example Mixpanel who got hacked is US based.

So it didn't introduce anything yet then.

[dire igloo]

it did

[night girder]

It's just guidelines, abstract words on paper.

Not laws. If it's laws then I am happy.

[dire igloo]

1. the deadline for when member states had to implement it has already passed
2. there was a directive to apply NIS2 in member states that have not yet made it national law

[night girder]

let me look it up for my own country.

But so far, I don't think NIS2 could have prevented to Phub hack or reduce its risks.
It's US entity that got hacked, they don't have to oblige to our rules.

but good for our EU critical infrastructure. If they actual put it into action and not just words*.

[dire igloo]

and for 2:
That's where the supply chain protection comes into play.
If a European company is classified as critical and wants to do business with Mixpanel, then they have to ensure that Mixpanel complies with the appropriate infosec rules

tho PH is Canadian so idk how NIS2 would've applied to them

[night girder]

I think it's a good thing. Maybe a tiny bit late, but hey, better late than never.

[dire igloo]

it's a whole fuckton of work

but it brought a lot of valuable much needed improvements to infosec legislation

[night girder]

PH is Canadian so idk how NIS2 would've applied to them.

It will, if PH wants EU consumers to pay for their service, PH as to hold up to EU laws.
EU has a standpoint of gatekeeping online services now. If you don't hold to laws, platform is not allowed.
More and more countries starting to do it.

And since the implementation is per member of EU. Each country can decide to not allow PH anymore.

[dire igloo]

generally, yeah, EU does that. idk if it does so for NIS2 specifically

[night girder]

Nah, I don't think PH is critical 🤣

But I really like this idea for critical services. They should not depend on the goodwill of others. It should be inhouse and/or up to our security standards.

[dire igloo]

they are

because of number of customers served and amount of revenue generated

[night girder]

Really happy EU is taking the wheel on this one. And not just sit in the backseat.

The same with cars right? Or airplanes. If they want to fly here, they have to be adjusted to our laws and standards.
I know that. Because Belgium has been waiting on their F 35 for a while because they had to be fully modified so they could legally operate here.

Now we got 4 out of 5 I believe. One broke down already so it couldn't get here 🤣

[dire igloo]

Bundesrechnungshof (German Federal Budget Auditing Agency) recently did an analysis of the BSI (federal infosec agency).
BRH almost always tells agencies they're spending too much money and don't need to expand their staff, usually the opposite.
However, for BSI, they found that they are lacking hundreds of jobs/positions

and number 1 reason is that NIS2 will bring a lot more administrative work for BSI

[night girder]

but for most products, simplified, it's; "our rules, or out"

So not crazy the same goes for internet services.

[dire igloo]

which is good for stuff like supply chain protection
but sucks ass for stuff like mandatory age verification via government ID

[night girder]

privacy vs security. They really want us to believe we have to give up our privacy for security.

Probably because it's most of the time easier, quicker and cheaper, to gain security, when privacy is not a thing.

But a lot harder, to make the internet more secure, without messing up the privacy of everything.

There is a rumor UK wants smartphones, to have a nude filter, so if you take a nude picture. The phone will block it or report it or something.
Not 100% sure on the consequences.

[dire igloo]

except without privacy, what's left to protect?
just a bunch of sentient meatbags

[night girder]

And the reason is, as always (atleast the last few months), ... CP.

arstechnica.com/tech-policy/2025/12/uk-to-encourage-apple-and-google-to-put-nudity-blocking-systems-on-phones/

First non pay wall and english article I could find about it.

And then we have the whole chat control debacle.

Roblox asking for identification of kids. etc.

And every day one or other big database with all our data gets hacked. And they want to collect more data.

Australia banning social media under 16 years old.

[night girder]

There is nothing left to protect. And there is no need for protection.

If we can believe those who are pushing all these privacy invading laws, we will be 100% safe and secure.

[dire igloo]

there kinda is. you could still get shot

[night girder]

Because that's the goal right? Why else trade privacy in if not for secure society?

Yes, you can shot, but in theory, there should be 100% chance of catching the person who shot you?

More dystopian future I am talking about probably 😄

[dire igloo]

my point is, the only things left to protect are the sentient meatbags that are our bodies

[night girder]

I keep thinking of that one movie.

[dire igloo]

because our identities are no longer being protected

[night girder]

Where they monitor everyone and had some spooky ladies in a swimming pool who predicted future crimes.

[dire igloo]

yeah, Minority Report

[night girder]

Doesn't mean we don't have an identy. And there is nothing to protect our identity against.

Because everyone knows everything.

Also, having no privacy, doesn't mean we can't identify ourselves. That depends on what kind of society it is.

It can be possible to have no privacy, but 100% freedom in being who you want to be within a legal framework. Depending on laws.

So the only question is, what is the legal framework (which I don't think we can know)

To me, no privacy != reducing us to just meatbags.

[obtuse geyser]

Wonder how many of those are just VPN adresses that happen to have exit node in nl.

[night girder]

Since it's premium, and they have to pay, it's possible it's connected to your payment adress.

but good point nonetheless.

I also hope that if people used a VPN for privacy reasons, they didn't make an account on a adult website

[obtuse geyser]

Pretty sure there was a free premium offer during covid lockdown.
You can use throwaway emails to register on sites via vpn, as to payment... you have to research your options.
Use vpn to pay for another paid vpn 😁

[cyan crescent]

I dont know if i want to stay on apple or switch back to android.

[obtuse geyser]

An apple may delay a doctor's visit, but becoming a robot makes you no longer need a doctor.

[languid gulch]

if i am ever somehow tempted to go apple, block me because my account's been taken over

[soft bloom]

I want to scream

I... fell for the oldest trick in the book

change of elements in container inside the loop that iterates over that container

aka iterate over list and remove items from it

I was trying to debug it like 4 times

[mental oriole]

Classic

[visual tree]

Some people really live under a rock. Had an email from a customer who only sent his street house number without street name. Asked him for street name and he mocked me the village doesn't have a street name....

Assuming he knows how to read, he should educate himself on the internet that all places have a street name, including his village

It's like some people were living 200 years underground and recently learned about the concept of street names

And such people have a right to vote 🤦‍♂️

[twin dew]

Might be one of the villages with just one road, so no-one cares about the street name.

[dire igloo]

Then the question becomes:
Is the zip code and village name specific enough?

There are some zip codes in Germany where the street name isn't the actual street, it's the name of the village.
That one zip code then includes like a dozen villages which each only have like 20-30 buildings at most

[twin dew]

Which is why there is real chance that asablic is in the wrong in this one.

[night girder]

Even one road gets a name here. Mostly named to where ever city the road is going 🤣

"Antwerpse Steenweg" -> "Antwerp stone road"

[twin dew]

Fireworkers point was in some areas there might be just the village name instead of street name.
And at least in UK at some point there were villages where there weren't any house numbers.
Just person name, village and possibly street.

In Finland, outside of urban areas, mostly we have standardized to ZIP code, street name, and then the "house number" is actually distance from one endpoint of the road in tens of meters.

And in some cases the actual house might be 1+km off in the forest at end of tiny side road.

[visual tree]

The village has only 1 street and the street name is identical to village name

However, cadaster data says the village still has a street name

I'm looking at this from a legal/administrative point of view

[twin dew]

So just forced fill in to provide all mandatory data fields in that cadaster data.

[visual tree]

Since legally speaking, official government data says it has a street name

When a village has only 1 street, street name is usually identical to village name

But you still need a street name for every settlement in the country

Technically, government doesn't recognize an address with only a house number

[dire igloo]

So yeah, miscommunication

Street name = Village name
and because the person already gave the village name and the street isn't named differently, there was no info they could give

[pure karma]

welp im now a only 1 ear working headphone user

cable just noped out

[obtuse geyser]

Old trick - iterate backwards.

[mental oriole]

Better trick, move all elements to be deleted to back and then delete them all with backwards iteration.

[obtuse geyser]

nah, that requires extra operations for moving them. Assuming we are talking about linked list, you can remove the items from the end and it auto collapses. If you delete going forwards, the collapse changes index so you skip elements then go outside boundary. A common interview question.

[mental oriole]

Not talking about linked lists.

[mental oriole]

Tell me its just a normal dynamic array.

[mental oriole]

Technically ypu dont even iterate, you just pop back until you're done deleting everything you wanted to delete.

Also linked lists are bad for performance either way, don't use them unless your usecase really needs it.

But Im doubting we're talking performance.

[delicate anchor]

Any way to fix this? https://questions.satisfactorygame.com/post/66e07817772a987f4a8a9b7f

[jagged snow]

To be fair, they're only bad for performance if you need arbitrary acces. If you only need 0 or n and you're storing data of sufficient size that the pointer/dereference overhead is minimal it's a perfectly reasonably approach.

[soft bloom]

Well, pretty much. Python list

[sharp oasis]

Ptm7950 seemed to fail after awhile

So to LM it is

[bronze jasper]

that's just engineering linked lists to act like vectors. so just use vectors

[jagged snow]

I work mostly in embedded, so for me vectors aren't often avaliable 😄

[twin dew]

Doesn't seem to have failed?
Still full coverage with no sign of material change.
Just unevenness from some having stuck to the cooler instead.

Contact pressure from the cooler has been very bad, but that is very common with laptop coolers.

Leading to excess PTM to not get pushed out enough and leaving it too thick.

[manic cipher]

Apparently the blue port on the audio panel on the motherboard allows one to hook up a radio to it.

Can streaming software ignore that port when recording?

[twin dew]

Analog Audio In.
And what inputs etc. are recorded is always selectable.

[manic cipher]

I guess when one is using the blue port they can listen to their radio while streaming and not have the audio cross over to the desktop audio unlike when one uses internet streaming.

[twin dew]

Depends what you are recording.
If you are using "what you hear" or like, it will be there.

Which that "Desktop audio" would be.

[manic cipher]

I would have been thinking stuff like game audio or microphone but ignoring certain inputs if that is possible.

[twin dew]

Which you would need to select specifically to be recorded, which then would ignore browser sounds when not selected.

[manic cipher]

I discovered Line In within the last hour and tried it.

[dire igloo]

What are you using for streaming?
With OBS, you have super fine control over all scenes, so it's absolutely doable to select what people do and don't hear

[twin dew]

Seems to be just using the default of recording whatever he hears, which would also include that line in if set to be heard.
So solves nothing.
And with the real solution line-in isn't needed and that internet music would not be recorded.

[dire igloo]

Yup, exactly

With a good OBS setup, every single item in the scene serves one and only one purpose

I wonder if there could be another effect going on where the radio signals are just artifacts from the mic cable accidentally functioning as an antenna

[twin dew]

Connecting actual physical radio receivers headphone or line out to line in.

[dire igloo]

I mean, if you deliberately hooked up a radio to it, then yeah, it's just scene sanitation in the streaming software.
I was moreso referring to the kind of tech that phones utilize when they use a connected 3.5mm cable as a radio antenna

[twin dew]

That is just FM-radio receiver in the phone, and the cable is just the antenna, because the jack is wired for that.

[dire igloo]

Ah, yeah, makes sense

Tho I wouldn't put it past old MB makers to include an FM receiver that uses the blue port for an antenna

[twin dew]

Never been done AFAIK.
And that blue is the standardized color for line in.

[dire igloo]

Yeah, didn't think it was realistic either. Possible, not probable

[twin dew]

And even if it was done, then it would need drivers and control software on the PC.

[dire igloo]

Fair

Tho that might've just been part of an install package

Anyways, @manic cipher if you need help setting up your scenes, feel free to share screenshots of your streaming software and what you have for inputs in your scenes

[mental oriole]

More crap...

[twin dew]

But any disabling of out great features must have been in error!

[night girder]

Diluted, I need help.

What is crap because I see a lot of crap in that snippet 🤣

[twin dew]

Lot of crap, but then there is that "Please turn out crap back on!"

[twin dew]

"Nice" that HP and Dell are starting to disabling HEVC/H.265 decoding HW support on IGPs in some of their laptop models, because the licensing costs are increasing from $0.20 to $0.24...

I had missed the prices... Otherwise this is now old news.

[visual tree]

"You need to work for Big 4, they said. You will learn a lot, they said...."

I am wasting my time explaining to Deloitte auditors how SAP works because they are clueless and think they know everything....

[twin dew]

And when SAP is almost always customized to hell for each specific deployment.

[visual tree]

A friend of mine also has a problem with them and mentioned Big 4 hires fresh college graduates with zero work experience who have no clue about finances or accounting besides what they learned in books

What you learned from books is MUCH different from what you will encounter IRL

[twin dew]

Because those are the "cheapest" employees, when you look only at the salary alone.

[sharp oasis]

I'm having pretty severe thermal throttle issues currently I wonder if that's more attributed to heat density of running Pcore only

[twin dew]

During turbo (157W), or base power (55W)?

Most late Intel laptop CPUs have been set to turbo way above what their cooling can handle until the turbo time ends.

[sharp oasis]

During modified turbo at 118w

It handles a base of 95w fine

I scaled back the turbo wattage a ton as it just cant sustain 157w

For longer than like 4 seconds

[twin dew]

But like I said, the phase change material hasn't flattened down to anywhere near normal.
Which should leave almost none of it between the CPU and coldplate.

So might have needed more thermal cycles to set properly.
Or just that the coolers mounting doesn't provide enough pressure.

[sharp oasis]

I wonder if it's combo of GPU load and CPU. Cus like the one time I got this thing to stay cool with was a fucking leaf blower to it's heatsink

They use the same heat pipes and heatsink

But even 107C is a bit hot for even my liking

Throttle point is 105c

@twin dew it could also be heat density from disabling ecores. So my thermal load is really dense and restricted to only the pcores. But it should have a vapor chamber to help with this at least slightly.

[twin dew]

Basically there should be spots that seem to be down to nothing in between with bare die visible once PTM7950 has set properly, as it should thin up to tens of nanometer range for the thinnest parts by the extra getting pumped out.
The difference to normal pastes is just that when it cools, it pulls back somewhat so it doesn't pump out completely over time.

Also do note that there are lot of cheaper stuff being sold as PTM7950 on Ebay, chinese sites etc.

Or selling mixed stuff, where some get real stuff, some get whatever that looks close to same but isn't the same.

[sharp oasis]

I'm thinking of running conductonaut

And doing the work to isolate SMD's and put a gel around the cooler to catch any stray droplets

[pure karma]

I put paste between my heat pipes and it improved temps at lower power draws quite a bit

[languid gulch]

waiting for nvidia to announce that customers will have to source and solder their own vram

[soft bloom]

https://www.youtube.com/watch?v=NIk_0AW5hFU
oh no, another 'many worlds' explanation...
how do they end up completely ommiting that any explanation sould match experimental results? with previously explaining how hidden variable won't work

also, great shoutout to all the engineers who worked on making these experiments happen.
it wasn't just a single one - it evolved through time, with some new ones, that rule out local variable happening to this day

[night girder]

There is something faster than light ... what a title.

is it quantum entaglement? Which in theory can be "faster than light". It doesn't travel faster than light though.

[soft bloom]

that's what Einstein suggested as explanation. And Bell proved him wrong on that.

[night girder]

Einstein said nothing can travel faster than light. But I guess that's why the title left away the word "travel" which has room for interpretation maybe...

And Einstein and Bell can be both right at the same time?

Because I don't think there is any traveling involved with quantun entaglement. Afaik.

[jagged snow]

Basically there are two interpretations but I do tend to agree with Einstein and the hidden local variable concept

Even if I'm not at all qualified to have an opinion

[night girder]

There is just a connection, no matter what the distance is, ...

[jagged snow]

Sorta
Doesn't mean matter or energy or even information is transmitted

[night girder]

and theoritical, distance doesn't matter I think, if one particle changes state and at the other end, one observes another particle (that is entangled*), the state is correlated to first particle. That's the entanglement/magic.

Still weird to wrap my head around.

[jagged snow]

Yep! But Einstein's explanation is that at the moment they become entangled the state that they will enter upon being observed is fixed but not realized until observed

So it's not so much that observing one chsnges something about the other

[night girder]

Because my brain keeps telling me, there has to be some sort of communication between the two, IT terms; there has to be data that keep the two in sync.

[night girder]

Yeah I just know that observing locks the state, that's what the cat experiment was about.

[jagged snow]

Not necessarily!

[night girder]

Schrodingers Box?

Well not locks.

They are all states at once.

But if we observe, we see a state ...

If we don't observe, the particles are at all states again. If remember correctly.

[jagged snow]

So, in it terms
When they become constant they're essentially two seperate objects with a private member variable that determines what state it will be when observed... Observing is like a getter function that allows you to see that state

[night girder]

but what we observe... isn't really the state or am I wrong?

Ok, I think I am wrong.

Once one has measured the system, one knows its current state; and this prevents it from being in one of its other states ⁠— it has apparently decohered from them without prospects of future strong quantum interference.[3][4][5] This means that the type of measurement one performs on the system affects the end-state of the system.

So we do kind of lock in it seems.

This is some wild stuff.

[soft bloom]

heh, then don't try to read many experiments that were advancing in ruling out any possibility for hidden variable

[soft bloom]

nothing can travel, including information. entanglement suggests that there's information exchange between 2 particles, regardless of distance between them, instantaneous

[jagged snow]

Again, far from expert
But it's either FTL relay of information(notably not matter or energy) or something akin to the hidden variable

[soft bloom]

... was a demonstration of how quantum logic doesn't apply well to macro objects. since we obviously know that cats don't just behave in superposition instantly collapsing to one state or the other upon opening the box...
it's the old question of does tree leafes fall in a forest that nobody observes?

[night girder]

I just watched double slit experiment again. And I have to say. It feels like we're missing something.
Especially the part of the atoms acting like a wavelength when going through double slit. And when observed they act like particles. How. https://www.youtube.com/watch?v=A9tKncAdlHQ

Or is my interpretation of the video wrong? I mean he explained it pretty well. But didn't explain why.

When disabling the observation equipment, they behaved like a wavelength again.

[soft bloom]

I will just say that it's fascinating to read resaearch papers about experiments that evolved(!) to study this question. it is still in debate

[soft bloom]

I will recommend watching series done by Looking Glass Universe (she was invited to be producer for the video I linked rom Veritasium today)
she was researching this topic to make experiment of her own at home, and she brings up that problem of particle vs waves interpretation
it is... not something that has a universally agreed upon answer

at some point, you start asking yourself, if you are mistaking math model for truth

[night girder]

yeah I can understand why it's not agreed upon ... because there is a gap of knowledge.

like there is a missing puzzle piece.

[soft bloom]

that's the problem - so far we only have tools that measure particle, and not do anything with a wave.
but even in this wave, we sense that it cannot be just particle

[night girder]

Most I heard is; "we observed that this happens. but we don't know why."

[soft bloom]

sort of like, if we couldn't see ghosts, but had evidence of spooky interactions, that coubldn't be explained by living breathable rigid humans

[night girder]

we know that a apple falls down, but we don't know the concept of gravity (yet)?

[soft bloom]

we know how it behaves, we think...
except we still have 'dark matter' and can't really understand how galaxies formed, why everything didn't end up in single black hole, how planet systems form, quantum gravity is a whole beast on itself...

[dire igloo]

Wdym? We got interference patterns

[jagged snow]

Absolutely agreed, sorry if it came across like I was trying to make an authoritive statement
Was just trying to explain the Einstein interpretation

[soft bloom]

in just little over 10 years since Elite Dangerous released, that was built on modern (at the time) understanding of planet system formation and distribution of matter in our galaxy, models changed quite significantly, and you can read some discovery in cosmology almost every week

[soft bloom]

which is a result of particle reacting with matter

[dire igloo]

Wha

Explain that to me

[soft bloom]

sort of like the difference between looking into the packet stream to printer, that is digital, and looking at a printed out paper

can you decipher printer protocol by looking at the paper?

[dire igloo]

You're not making sense.
Interference patterns are like THE method of doing science with waves

[soft bloom]

exactly

I am saying that we can't measure waves directly, we always have a medium that collapses wave into particle, and then particles interact and we read result

kind of like walking through the forest and seen footprints of bigfoot, but never seeing it directly. does it exist? or just someone playing pranks on you? many will say that it's just the way footprints erode over time...
giving example to point at the problem

[night girder]

well, we know there are two interference patterns and a correlation with observation.

Well, maybe I should say one interference pattern (wave) and the other normal expected pattern from particles.

[soft bloom]

that's why Bell's experiment is so powerful - it uses math and instrument we have to point out that there's a difference in expeceted results from different interpretations. but it still won't "show" us entanglement. and many won't ever beleieve without "seeing" for themselves.

interference pattern exist ONLY as result of MANY particles interacting with matter.
we didn't ever record interference pattern of a single particle, because the only thing we can get is 1 'dot' on a screen.
(even though there are allegedly experiments where particles interferes with itself, interference suggests that there 2 waves)
that's why I say it's so hard to prove that waves exist - we don't measure them. we only observe result of particle interaction, and that particle behaves like a wave

[night girder]

We send one particle at a time, through the double slit experiment.

and we get a wavelength interference pattern. Just like with light wavelength.

We observe, and send one particle at a time through the double slit experiment and we get particle pattern.

That's all I understood from it.

So there aren't "many particles" that interact with matter at the same time. Just one by one.

[soft bloom]

and 1 particle usually stands either for actual particle like electron, or a photon, which blends the lnie between wave v particle even further

yeah, but shooting exactly 1 particle gives you just 1 dot on a screen

you measure probabilities of where particle ends up

and then use wave model to describe results

...
math behing oscilations in mechanical system, electrical engineering and hunter-prey populations is fairly similar
but using formula doesn't tell you what exactly happens there.
formula is ust a math model, and nothing more. a way to predict, but not an explanation

heck, universe might be running on giant supercomputer with every particle being controlled by an all knowing process that just decided where it will go and according to which probability distribution

btw, this is a big hole in quantum mechanics - instead of classical stuff where you have rather clear answer, you get... probabilities. that is a big sign that understanding is missing. and it could be simply result of dogmas that prevent asking right questions

[solemn forge]

nvidia will switch tracks and do memory instead to meet demand

that memory is ofcourse so high grade only openAI can use it

[languid gulch]

delete that before some AI scraper sees that and someone at nvidia decides it's a good idea

[bronze jasper]

Nvidia doesn't do their own fab. They design the chips and have actual fabrication places build them.

If they switched to memory they'd still have to buy capacity at a factory

[soft bloom]

btw, this exists
https://github.com/vvstubbs/removewindowsai

I forgot where I saw it
and I don't like how it uses URL to download script
but, that's something

[twin dew]

They already stopped including it with the die and the card makers now need to source their own.
Where for long time Nvidia didn't allow them to do that even if they wanted to.

[solemn forge]

oh right TSMC

[twin dew]

And memory fabs and general silicon chip fabs are mostly completely separate, too different processes.

memory and flash lines are somewhat transferable between each other with some work.

[soft bloom]

I started using va fhd monitor evven when power is off wnd it is just so much fucking better then tn screen in my laptop ...

[soft bloom]

so, J Blow anounced his new game
and visited StandUp (with primeagen, teej and Casey Muratory)
and ... I realised I miss listening to someone how actually knows their stuff,
and sort of like meta overlook of the problems in game design and software engineering

which reminds me that I should give another go to figuring out how to solve problems, learn stuff and have fun
last time I tried lookin those topics up, it was half confirming what I already know, half confirming what I already do, and half almost esoteric woodoo "maybe helps, maybe not, but we will do it anyway" afair

[night girder]

So I just looked this up. Double Slit experiment is a real experiment and not a mathematical prediction of the pattern.

"you measure probabilities of where particle ends up
and then use wave model to describe results"

So this is bullshit. This is not how I read the experiment was conducted. It's not just theoritical and mathematical predictions.
It's real world observations in experiments.

[soft bloom]

indeed is
you even can use fog to see interference pattern in space, instead of flat screen (as video on Looking Glass Universe showed)
but, notheless, that only shows you result of how individual particles interact with matter
and still leaves out the question of is there some medium in which wave of that particle propagates and behaves like a wave
||oh, wait, people in early 1900s century did explore ethyr, but...||
the 'wave' is math, pattern on the screen - is observation. math predicts that observation, but is not The truth of how it happens to be.

you can predict changes in hunter-prey population over time with same math used for oscialting circuit in electronics, but it doesn't mean that fox and rabbits are coil and capacitor.

[night girder]

No?

The wave isn't math. It is observed.

[soft bloom]

having observations that match prediction doesn't make model ... not predicting.
Newtons gravity is a model. it predicts. it matches observations in a lot of cases... but cannot describe Mercury - that's where we needed relativistic gravity.

[night girder]

I didn't say that.

All I am saying, and you seem to deny, is that they OBSERVED that a particle behaves differently when observed in the double slit experiment.
Either it has wave like behavior. And when observed, it shows particle like behavior

And you cannot deny, that scientists see different patterns. The experiment can be wrong. The results can be wrong. But they seen over and over again, the same result.

[soft bloom]

https://www.youtube.com/watch?v=W3Egv6iO3dI
https://youtu.be/sAm7iVdAvTA
I suggest watching both of these to understand the nuance I am pointing at

[night girder]

The wave pattern for electrons passing through a double slit, one-at-a-time. If you measure “which slit” the electron goes through, you destroy the quantum interference pattern shown here. However, the wave-like behavior remains so long as the electrons have a de Broglie wavelength that's smaller than the size of the slit they're passing through.

[soft bloom]

I wasn't even talking about wether or not it changes behavior when one slit has particle detector

[night girder]

Electrons exhibit wave properties as well as particle properties, and can be used to construct images or probe particle sizes just as well as light can. Here, you can see the results of an experiment where electrons are fired one-at-a-time through a double-slit. Once enough electrons are fired, the interference pattern can clearly be seen.

If you measure which slit an electron goes through when performing a one-at-a-time double slit experiment, you don't get an interference pattern on the screen behind it. Instead, the electrons behave not as waves, but as classical particles.

That sums it up for me. So I don't understand the whole "Oh, but that's because the particle interacts with matter." - argument either.

Well IN BOTH cases (wave or no wave) the particles hit the optical screen (interacting with matter).

The gist is just that observing changes behavior which leads to a few ??? in science.

[soft bloom]

I am saying that, to my knowledge, there's no experiment that would take a SINGLE particle (electron, photon or larger) and would succeed in measuring it's probability wave distribution.
all there is is shooting many particles, individually or not, and measuring result of that.
Looking Glass Universe also shows you how it behaves with water analogy, but we can SEE waves in water when we create a SINGLE wave. We have direct observation of wave in a water. We don't have that for photons or electrons.

[soft bloom]

reminder

[soft bloom]

Once enough electrons are fired, the interference pattern can clearly be seen.
enough

btw, do you see how its formulated?

wave-like properties
we DON'T KNOW if it's a wave
we only know that it matches model for wave

Imagine you are training a shy horse
you reward her for crossing an obstacle of a wall with 2 narrow slits for her to cross
on the other side there's a row of apple buskets, plenty, with no meaningful difference between then so your horse can choose them practically at random.

At first, you closely observe how she crosses the obstacle
and for some reason, she always chooses basket that she can see directly from the start, running through either slit in direct line, ending up at essentially one of two baskets.
all simple so far

then, you refill the baslets, and decide to read newspaper, while she is doing excercise over and over again.
after some time, you come to count how many apples are left in each basket, and see that... missing apples match interference pattern.
DId your horse secretly entered wave form and collapsed at different baskets one at a time? did it decide to have fun and just ran around in weird trajectories? or did it somehow broke your training program and decide to eat apples at random without actually running across the bstacle?

You won't know, you didn't record it on camera.
and next day you search closely - she does the normal thing as before. You put a motion sensor on one of the slits - she bahves normally.
You don't know what to do.

[bronze jasper]

the thing with the double slit experiment is describing it with macro sized objects is not an honest account.

"you put a motion sensor" obviously doesn't affect the horse and that's why it's ocnfusing that it would affect a horse. But in the quantum sized world, "put a motion sensor" is an obstacle that the photon must interact with, and when it interacts with the sensor, it's sensed. But also since it interacted, it changed where the particle is going.

[soft bloom]

yeah, the analogy far from perfect, and using macro object is indeed biasing towards particles behaving like a wave, instead of ... just wave.
yet that is an unsettled debate, of what it actually is.

the last paragraph is optional and more of a joke there.

[sharp matrix]

right and that goes into the uncertainty principle, because measuring inherently changes the behavior of what you measuring, leading to less accuracy depending on the property your trying to measure, because of that interaction as you say

[soft bloom]

and that's why I linked those videos from LGU. iirc she did one as response to Veritasiums video about photons spreading in every direction, sort of, before collapsing? it's a fun story

when talking about less studied topics it's less about "what actually happens" since we don't really know yet, it's more about what nuances we observed and are aware about. and sadly, this field didn't have any breakthrough in a loooong time.

[night girder]

Yes, because I try to formulate in the same way.

wave "like" behavior.

[sharp matrix]

well no surprise, particle physics isnt exactly a easy field to make break throughs, will we ever known what the double split experiment is really telling us, maybe one day, but probably going to be a long time before we fully understand the science if ever, but i would imagine it would be around the same time when we finally have a unified field theory

i have a feeling that unified field theory is going to be some ways away from being realized, could take some more decades or longer, its a tough problem

[bronze jasper]

The future circular collider will probably unveil a lot theory on particle physics. I'm surprised we were ever able to use the LHC to detect the higgs

construction will begin sometime next decade

[dire igloo]

if it walks like a duck, looks like a duck, quacks like a duck and flies like duck...

[night girder]

in python it's a duck 🤣

Summer is right though. Scientists are careful with calling it a wave.

[soft bloom]

and that's how earth centric model was born 🙂

joking

[night girder]

They try to describe it like stated. "the behavior is similar to a wave" or the pattern we see, is the same we see with light waves etc.

but I don't think any scientist dared to say, yeah it's a wave for sure. Because they haven't been able to observe it (because observation collapse wave).

[dire igloo]

at some point, it's more about definitions.
if a non-wave matches all the criteria for a wave, you either have to change the definition of wave or you have to accept that that thing is a wave

[soft bloom]

actually, once again back to oscilations. many processes are described with same math. but doesn't maen that capacitors and coils in your laptop are actually tiny foxes and rabbits

[dire igloo]

except geocentric model didn't match reality

or rather, reality didn't match the model

[soft bloom]

it was 'perfectly' viable before new observations
it;s easy to llok back and be like "oh, they just didn't account for X"

[night girder]

1. The pattern (the end result) is similar to a pattern that we see when we do double slit experiment with for example light. Aka. Cancelling waves and ampliying waves. The interference pattern.
2. So scientist think it's a wave and wanted to observe, but once they observed, the pattern changes to what they expected from a particle (no interference pattern).
3. Scientist don't know if it's an actual wave. They only know the end result, on the wall, seems to be same pattern they see when doing double slit with light.

[dire igloo]

i mean, yeah. that's how models work
you either change them to match reality or you accept the mismatch

[night girder]

All we know it quacks like a duck.

But can still be some different thing totally, that just ... seems to also quack 🤣

I can quack for example, doesn't make me a duck.

[night girder]

or accept a gap in our knowledge? If the answer was simple. We would have it by now. That's why it's so fascinating.

Still possible something else is going on we just don't know about particles (yet).

[dire igloo]

depends on what you define as "duck"
if there are five criteria to match (quacks, flies, looks, walks, swims) and you fulfill only one (quacks), that means you are not a duck.
but if all an animal needs to be counted as a duck is to quack like one, then yeah, you're a duck

[night girder]

We only know it quacks like a duck.

[dire igloo]

yeah, that's the part where you change the model to match reality

[night girder]

When we use our binoculars to see if it flies like a duck.

The animal changes into a cow!

[dire igloo]

not really

because the measurement is the same, you just add one more thing

[night girder]

Ok, explain to me what other criteria, besides inteference pattern, makes you believe it's a wave?

Because the moment we observe, the pattern is different and the behavior changes.

[sharp matrix]

Probably, we still don't have a complete theory of everything, gravity for instance remains a unsolved problem for a unified field theory, so clearly our understanding of how the universe works still has much more to progress

[dire igloo]

so it's more like it quacks like a duck but when you use a microphone to record it, it moos like a cow

[soft bloom]

yes.

I was talking about different aspect of it, but yes.

it's kind of like ...

• skies are blue
• no, roses are red!
• yes, but skies are blue!

[sharp matrix]

Its clear the double split experiment is telling us something, we just don't have a complete understanding yet to know what that is

[dire igloo]

it's the moment when we measure the slit that the particle passes through
and another example: quantum tunneling

[soft bloom]

we just don't have any model that would explain why the heck particle behaves like a wave, any model that would gain track and give insight into what to expect, with ability to make new experiments and verify if that's true

[dire igloo]

we can send a singular electron towards a double slit and after enough electrons pass through, we see an interference pattern
but if we change the setup and put a measurement device on each slit to know where the electron passed through, it no longer produces an interference patern

[night girder]

Man... I hate science cliffhangers 🤣

[dire igloo]

the easiest solution to that: we have no fucking clue what a particle even is

[soft bloom]

JWT is like biggest producer of them

[night girder]

That's what this feels like. Scientist telling "Hey we saw this, no freaking clue wtf is going on, but amazing right?" 🤣

[night girder]

That's why we are doing experiments. Attempts to understand it more.

[sharp matrix]

Right, thats why our models are still incomplete, because a complete model would tell us why the particles are behaving in the way they are, this shows that there is definitely either holes in our current models or our particle physics models need to be reworked

[night girder]

But so far, it raised more questions it seems.

For a moment my silly head just though, what if some* particles just collide and bounce out of trajectory (creating interference pattern, without being a wave). But then they did the one by one particle ... so out of the window went my simple peasant theory 🤣

[sharp matrix]

it seems like a our particle physics models do need more work, because even those models dont accurately explain what is happening in these experiments either

[soft bloom]

well, we have standard model, but it doesn't give more insight into photons or electrons

[dire igloo]

there are two definitions for mass:
F = m*a
E = mc²
basically: mass is either an object's resistance to changes in motion or it's the equivalent of the object's combined energy

thing is: these two definitions always match. Why? nobody knows

[soft bloom]

btw, double slit experiment was done with heavier stuff
I believe last thing was some atom as 'particle' they shot?

^projectile

[dire igloo]

there's a great book about all those things, We Have No Idea by Jorge Cham and Daniel Whiteson

[night girder]

In 2013, a quantum interference experiment (using diffraction gratings, rather than two slits) was successfully performed with molecules that each comprised 810 atoms (whose total mass was over 10,000 daltons).[4][5] The record was raised to 2000 atoms (25,000 amu) in 2019.[17]

[soft bloom]

well, I don't thin problem is necessarily in the model...
we have math to describe double slit experiment. we don't have understanding of it. which is a different part of modeling, that is a bit more... esoteric, to say.

but, for example, we don't even ask question why some electrons will go left, and others right. why some photons will reflect of a mirror, and other's don't.

classical mechanics alternative would be to say that we have 50% chance that cannonball will fly in parabolic trajectory, and 50% it will go in a straight line... we just have to make enough shots to average out chances and achieve 50% hit rate on the target...

[night girder]

Seems funny. But frustrating.

A book answering nothing.

[soft bloom]

I need to read that, no clue how they made difraction grating thin enough to pass molecules through it

[night girder]

Also, "who is shooting deadly particles at the earth?" 🤣 Who? Really?

[dire igloo]

it really is

[night girder]

Tempted to buy it.

It will explain, why we don't know, what we don't know?

[sharp matrix]

So basically 366 explaining how we understand nothing really at all

[night girder]

Also, stupid question.

[dire igloo]

kinda

[night girder]

How come we have quantum computers.

But don't understand quantum mechanics fully?

[dire igloo]

but we often don't know why we don't know what we don't know

[night girder]

Or is quantum computer != quantum mechanics in general?

[dire igloo]

a lot of text assisted by some graphics

[soft bloom]

combined energy
woah, hold on a second.
define 'combined energy' there.
you can change electron's kinetic and potential energy without affecting mass, clearly.

[dire igloo]

quantum computers use quantum mechanics, but it's like saying that a CPU is the same as electronics

[night girder]

we know how electricity works when we made computers.

[sharp matrix]

so more graphics explaining how little we really understand

[soft bloom]

that's the beauty of modern science paradoxes
commerce is eager to exploit anything we know how to model
while science is... reluctant to ask bold questions
so there exists a grey area of "we have no clue how but it works"

[night girder]

we are making quantum computers, without understanding quantum fully. Seems weird

Like making a car, but nobody really understand why the engine works like it works (explossions, valves etc).

[soft bloom]

we don't really know electricity... we built a lot around simplest stuff that involves it, but if we KNEW we wouldn't have that debate about speed of light in the cables

[night girder]

That's not a debate. That's tested.

[soft bloom]

more like riding a horse without knowing it's biomechanics, yeah. it's barely tamed

[dire igloo]

kinetic and potential energies are tiny compared to internal energies
stuff like bonding energy within a molecule (which makes Nitrogen a great explosive) or inherent energy of the subatomic particles

[night girder]

https://www.youtube.com/watch?v=2Vrhk5OjBP8

[sharp matrix]

Isnt the problem with science though we can ask the bold questions, but us trying to figure out the answers is another question and now we are making a even larger collidor in hopes of trying to understand how the universe works better

[twin dew]

No, but lower speed depending on the insulation used.

[soft bloom]

yeah, but it happened BECAUSE debate arised. and it revealed how many people have shallow or wrong understanding of something what should be basic

[dire igloo]

you don't need to know what a duck actually is if all you need is something that quacks

[sharp matrix]

I wonder how big do we need to make these colliders again, surely we will eventually understand how particles work if we just smash them together with high enough energy levels right....right

[night girder]

Well you do need to know if you want to build a duck.

Quantum computers don't grow on trees.

[twin dew]

So I have no idea where that "debate" comes from...
When the answer has been known for long time already and has to be taken into account in antennas etc.
And why microwave was used for latency sensitive things and not copper.

[night girder]

It's just something that doesn't click in my head, we make something, we don't understand fully, so how do we even know it's performing like it should etc.
Wouldn't it rather be logical to first have a better foundation of knowledge, science and understand before starting to put it into practical applications like computers.

[soft bloom]

I will tease you all and say that collider's biggest achievement was discarded as error in observations

[sharp matrix]

science is taking such a brute force approach, why do we always resort to smashing things together to try to undersand how things work that way lol

[dire igloo]

but if you're hunting ducks and need a flute to attract them, you don't care if that flute is an actual duck, it just has to sound like one

[night girder]

I can build a car, just on sheer luck, but if I don't understand how the engine works. And I need to "debug" my car not working.
It's insane.

[twin dew]

https://en.wikipedia.org/wiki/Speed_of_electricity

[night girder]

Yeah, ok. So then we have a flute, that sounds like a duck. That flute, doesn't mean we understand ducks at all. It's just a gimmick?

[sharp matrix]

see commercial doesn't care if it works, as long as it generates profit regardless

[dire igloo]

smashing things together is an easy way to find out what it's made of

[sharp matrix]

just throw quantum into the name, who cares if it works, as long as the profit comes

[night girder]

Quantum AI

What can go wrong right? Two things we don't understand. Or have a fully grasp on.

[dire igloo]

if the flute helps you hunt duck meat, do you care about the taxonomy of water birds?

[night girder]

Well the goal is understanding the duck. That's why we hunting ducks. Not the meat. Science wants to understand quantum and particles. Not just the result we can achieve with it.

[soft bloom]

it comes from ... wrong/simplified understanding
debate in media space, which reflects what normies think, on average
and still, it was between educated people...

[night girder]

To put something to use and benefit from it, we should understand it first.

[sharp matrix]

Nothing is going to be wrong, it'll all work flawlessly

[soft bloom]

because.. that's something you can get big funds for

[dire igloo]

yeah, that's quantum mechanics.
the quantum computer is the duck flute

[night girder]

I can give a monkey a knife, but it probably would help the monkey if I show what it can do with the knife. Sharpening sticks. Open fruit. Clean your nails. Cut your beard 🤣

Meaning, if we don't understand the underlaying mechanics, it makes it just harder.

We know about sharpness, how things cut, why they cut, why something can cut another material etc.

[soft bloom]

if it was like this we wouldn't have ozon problem, radioactive glass, or the whole oil based energy resources with plastic as main polutant we can't get rid of now...

[night girder]

But then again, in the beginning. We did use sharp objects without knowing the why. So I guess you do make a good point.

[dire igloo]

good luck teaching a monkey how to use a knife

[soft bloom]

and on the other side of spectrum of understanding vs implementation we have nuclear power, which is stagnated by security precautions that are insanely over the top

[night girder]

My point was; I think understand the WHY benefits.

[soft bloom]

it does

[night girder]

The cave humans used to sharp rocks to cut stuff without understand it. They just knew it does.

[soft bloom]

but using it can™ help understand it

[night girder]

So they hunted etc. But the more we learned about the how and why. The more benefits it had.

Now we have lasers that can cut stuff. Far far more advanced than smashing two rocks together to create sharp edges.

[dire igloo]

yup. but the understanding had to come from the doing

[night girder]

And that's why we have quantum computers.

Mhm. Now the puzzle fits in my head 🤣

[dire igloo]

plus, a knife is still a useful tool even if you don't know how sharpness works

I don't care why my knives are sharp, I want them to cut my food

[night girder]

Like I said, the cave man used the knifes. Without knowing why. To hunt. Cut fur etc.

But the more we learned. The more application it has. And the more advanced it got.

Diamond cutting for example.

So it has purpose. But could be the tip of the iceberg when it comes to usefullness.

[dire igloo]

yup, but we would've never found that out if cavemen never started using sharp rocks to cut stuff

[night girder]

True, but then we were monkeys.

[soft bloom]

heh, I almost thought it tries to say that it runs at 3m/s...

[night girder]

So I don't know. I guess I expected more from scientist.

Then just "We don't understand it, let's build it, and ... we see from there?"

[soft bloom]

and give money 🙂

[night girder]

And like I said, how do they even know the quantum computer works like ... expected.

or fully performance.

[twin dew]

Depends on things, for current antenna cables etc. usually around 2/3rd light speed IIRC.

[dire igloo]

the science is "let's have a sharp rock first so we can cut stuff, then we can figure out more about how cutting and sharpness work"

[night girder]

But we had already computers. I really struggle with the benefits of quantum computers.

[dire igloo]

yup

but quantum computers work differently

[soft bloom]

the section about speed in good conductors is a bit fonfusing and easyto misinterprit.
it's something that should be taught better imho.

[night girder]

Let me put it like this; wouldn't scientist better do other experiments to understand quantum mechanics better?

[twin dew]

https://en.wikipedia.org/wiki/Velocity_factor

[soft bloom]

and btw in the whole eletricity thing...
what is electromagnetic field? I mean, aside from math construct that any point in space has 2 components and particle act accordingly

[dire igloo]

https://youtu.be/RQWpF2Gb-gU
https://youtu.be/Dlsa9EBKDGI

[soft bloom]

another roof made better explanatino imho

[dire igloo]

quantum mechanics is a field of particle physics
quantum computing is a field of computer science that uses physics

[soft bloom]

it's different, but touches more... misunderstood bits

[dire igloo]

never seen it, I just know that these work as a good explainer

[soft bloom]

they want to eat. so they sell ideas to business

he has whole series
https://www.youtube.com/watch?v=9S9a21u3ZnQ
that mention quantum furier transform, which really is a missing piece of puzzle to all the explanations I seen on YT

it's the only one that explain how the F they measure something from qubit

[bronze jasper]

the big reason that Quantum computing isn't going to replace binary computing is because you can't copy the data at all. As soon as you read out a qubit, you destroy the qubit.

they'll do good in research applications but binary is hear for the long haul

[soft bloom]

btw. talking of using something we understand...
https://www.youtube.com/watch?v=D8GOeCFFby4
we don't understand LLMs anymore, but this video explores one that we still can understand

[languid gulch]

so what you're saying is whoever actually figures that out is going to be an instant trillionaire

[bronze jasper]

how to figure out measuring quantum states without destroying them? That'll be kept secret and used for war and conquest before it shows up in the public

[languid gulch]

exactly

[bronze jasper]

that'd be a good super villain in a bond movie

[night girder]

on the bright side, it does add a lot of security.

https://en.wikipedia.org/wiki/Quantum_cryptography

I wonder if we eventually will get a mix between the two. Quantum for security side and 0/1 for practical side.

[languid gulch]

would be neat

[coral timber]

anyone know what the hell this means?

[languid gulch]

that almost looks more like a brain fart than an actual failure

[coral timber]

lol

[bronze jasper]

the page file is your system memory when it's copied into a file on your storage. it's size is probably still set to auto, but if it's set to auto and is too small, that probably means your storage is full to the brim

[soft bloom]

I think PyCharm is... delusional, to say the least

[coral timber]

Storage isn't full I can confirm that

[twin dew]

For C:?
Then might just been too fast RAM requirement rise and the auto-sizing couldn't keep up.

But basically means that both RAM and SWAP ran out.

And by default Windows has automatically sizing swap file on C:\

[coral timber]

Damn

Alr, thanks yall

[twin dew]

Setting big enough static size helps if there is rapid swing where that automatic size cannot keep up.
And can be also set to use other partitions.

Or can in theory be from unstable CPU or RAM.

[bold tapir]

ai powers getting wild

[solemn forge]

intelligence ❌ Artificial intelligence ✅

[cyan crescent]

Whats a good 8TB m.2 ssd? Looking for reliability and longevity

[dire igloo]

aren't 8tb m.2 SSDs like double the price per TB compared to 4TB?

[cyan crescent]

Idk. Its going in a laptop with only one m.2 slot

[dire igloo]

other than that, the least expensive option that isn't ass. which one that is depends on your market, what country you're buying in

[cyan crescent]

Usa

[dire igloo]

and 4TB is insufficient?

[cyan crescent]

Price isnt too much of a concern. Yes. I have filled 4tb

Im on a ship for months on end with little to no internet. I download as much as i can before going out

Space is a concern hence the smallish laptop

[dire igloo]

4TB:
Samsung 990 Evo Plus, $328
Biwin NV7400, $355

8TB:
Lexar NM790, $730

[dire igloo]

which also makes cooling an issue

a hothead SSD will crumble more easily in a laptop

[cyan crescent]

Its an asus zephyrus g14

2022

[dire igloo]

doesn't really change much about the rec tbh
NM790 known configs all include very power efficient controllers, so heat won't be an issue, same for NV7400
and NM790 just so happens to be the cheapest 8TB m.2-2280 drive on US PCPP, so no reason to pick anything else aside from downgrading to 4TB to save money

[cyan crescent]

Money isn’t a concern. I have plenty saved up

[dire igloo]

still, NM790 is the recommendation

highend performance if you need it (assuming HMB is supported) and stays cool even under load

plus, not a lot of config variance

the low price is just a bonus at that point

[cyan crescent]

whats HMB?

[dire igloo]

host memory buffer
it's a tech for drives that don't have built-in DRAM-cache that allows them to use a small amount of system RAM

[cyan crescent]

Just gonna slot that in then move the laptop to bazzite.

[languid gulch]

would be hilarious to put 5x sabrent 8TB's in my board

like double the frickin price of my system

[soft bloom]

Merhaps better to opt for external drive? Cable speed will be enaugh fro whatever quality you choose for small laptop screen, 1080p likely more then enough for 14"
Prices will be better for that format, but post importantly more storage to have and ability to simply swap between 2 storages if there's a need for quick update of content

[languid gulch]

i do think that Paul's Hardware got it right. there's an industry conspiracy to not enlarge NVMe capacities except for commercial/enterprise

[twin dew]

PCB space limits in M.2 2280 form factor.

[languid gulch]

a lot of boards have 22110 now

[twin dew]

And ok, seems that they aren't doing double-sided for some reason either anymore.

So 12TB would be doable with 4 flash dies on the backside, controller and 2 dies on the front.

[languid gulch]

are current 8tb NVMe all double sided?

[twin dew]

Looked at 4TB

[twin dew]

At least the Samsung 9100 8TB was double-sided with 2 dies on each side.

[languid gulch]

is the density not increasing anymore?

[twin dew]

Every few years.

That is how we moved from 2TB to 4TB consumer M.2 SSDs.

[languid gulch]

it's also annoying that there's no 2.5" sata high capacity drives

16tb is the highest those go

[twin dew]

But with FLASH, the density is mostly only increasing with more layers.
We have hit the minimum node size for FLASH, any smaller and there wouldn't be enough space to store any real charge.

[languid gulch]

ah

i mean, that makes sense

so now the option would be to start stacking more?

[twin dew]

That has been going on for long time.
We are at about 200 layers right now AFAIK.

Where each layer means full flash die I mean.

Sorry, bad simplification.

[languid gulch]

does it mean we'll be moving on to a new socket/standard soon?

[twin dew]

Very unlikely for consumer standard to get changed.

Ok, Hynix went to 321 layer NAND FLASH dies in 2023.

[languid gulch]

jeez

[twin dew]

So if I got it correctly from some articles about problems in stacking more layers, then that is 321 layers is 160 layers with FLASH cells on them.

And rest are insulator layers around those "active" layers.

And you then usually have 4 or 8 of such multi-layer dies stacked on top of each other.

In each flash chip.

[languid gulch]

i'd love to see that manufacturing in person

[night girder]

Damn, Lexar has 8TB NM790, only 799 euro. Or 4TB for 299 euro. Seems about right ...

Seems 4TB is best bang for buck.

(third columns is price per TB).

[pure karma]

i mean i could see 4TB start going down since thats the next step up

rn 2TB is the norm for mid to high end where as next its gonna be 4TB

id still rather have a HDD instead at that price but if its nice to see the option become more available

[night girder]

I would buy all hardware now for the next 10-20 years to come 🤣 Hamster it.

Speaking of... should look into my NAS HD's to check their health.

[visual tree]

Guess I can finally sell my old car soon. City is buying electrical buses and implementing free public transport

My car is nothing more than a hole in a wallet

[night girder]

However, Sanders tells us that's unlikely to happen this time around. "I think we're looking at the peak in 2026," he said, adding that even then he only expects DRAM prices to settle in 2027 before rising again in 2028.

So it will go worse. Then better. Then worse.

• Every **even **year = don't buy DRAM.
• Every **uneven **year = buy DRAM.

But everything is totally fine with the market. Invisible hand clearly at work here.

https://en.wikipedia.org/wiki/Invisible_hand

[twin dew]

There has been two-year boom, two year bust, cycle in DRAM for decades.
Which is part of the reason why the DRAM manufacturers aren't planning to start putting up new production lines now.

Besides not trusting the datacenter boom to last long enough for those new fabs to get into production anyways.

[bronze jasper]

intel has 3d xpoint memory and that might start getting more popular as flash reaches it's limits. it's a 3d lattice of memory cells

[twin dew]

Completely different tech on cell level.
And the fabs are gone.

[bronze jasper]

Yeah the biggest upset in the market right now is the panic buying that happened after open AI bought a bunch of factory capacity. Once the panic washes out, i think the market will more or less look the same. New factories will be built. Supply will once again meet demand.

Prices at the consumer end will be fucked for a year though.

[bronze jasper]

it's still storage at the end of the day. It just wasn't economical enough to compete against flash for the time being. Nothing stops intel from pulling it off the shelf

[twin dew]

Only the "the fabs are gone" part.

[bronze jasper]

When they first developed the tech, the fabs had never existed before at all

[twin dew]

So any restart, even using exact same node, would take years.
And it is out of date in many ways already, so getting more expensive for same amount as the flash gets cheaper.

And the failure was Intel trying to segment it to sell their own CPUs, and misrepresenting it in many ways in the marketing because of that goal.

[bronze jasper]

why is it out of date? its faster and the memory density is potentially higher.