off-topic-tech

[night girder]

Then PC calculates the surface.

And voila, you know the surface of your planet for that given moment.

We do it already on small scale. Constantly.

[twin dew]

Still not a snapshot, unless you have Starlink levels of scanning satellites.

But different points will have been scanned at different time, over several days.

[night girder]

I said future no?

And if it’s not satelites, I assume it would be technology in space that will be scanning our planet. Need to have altitude. And we don’t have the technology yet to accurate scan a planet.

[tribal kraken]

You still need to define the resolution of the scan. It will have huge impact on the surface area when including finer detail

[night girder]

That too yeah.

[tribal kraken]

the key point that kade wanted mountains and valleys.

[night girder]

And without ocean!

That seems the hardest to me.

We know more about space than our own deep ocean.

Because PR wise, landing on the moon sounded cooler than spending money on deep sea exploration.

So we got the race to the moon 😂

[twin dew]

Also easier in many ways.

[night girder]

Eh…

I don’t believe that fully.

[edgy hazel]

it's also the frequency. othewise you get this stuf

[twin dew]

Mostly same requirements:
One needs to account for 1 bar pressure difference, getting too hot, and longer timeframes.
One needs to account for insane pressure difference, getting too cold, and shorter timeframes.

[night girder]

I sometimes believe that if we spend all that brain power into other projects, like deep sea, the problems might been fixed. But the focus and money was on space. So ofc it looks harder to go deep sea.

[twin dew]

But of course those shorter timeframes are only for the case where most of the time is spent going from surface to deep and back.
And the time at the depths is pretty limited.

[night girder]

And I know for a fact, astronaut’s train in a bade deep in the sea.

Base*

Because the sea can mimick certain conditions of space.

[twin dew]

Nah?
Just training tanks in specific facilities.

[night girder]

Nop.

There is a deep sea base.

Or a sea base atleast.

Same pressure and time as going to International Space Station.

[twin dew]

Any link?
Because AFAIK all the underwater training of astronauts for ISS outside work is done at couple of very large facilities inside US.

Large enough to have partial model of ISS in the largest one at a time, that they switch around for the specific training.

Ok, 2010 they started NEEMO, but that is separate research thingie.

Not for ISS training.

[night girder]

https://en.m.wikipedia.org/wiki/NEEMO

[twin dew]

And if you read WHAT they are doing?
Not training as such, R&D for future.

[night girder]

…

“for up to three weeks at a time in preparation for future space exploration.[2]”

[twin dew]

Yes, missions that don't exist yet?

[night girder]

What do you think that is!? Training.

[twin dew]

Training is for preparation of specific mission.
R&D is general "how could this work".

[night girder]

Huh? Man I listened to a podcast with an astranaut who has been there. Befere he went to ISS

That’s where I got my information from.

[twin dew]

The training for specific missions is done here:
https://en.wikipedia.org/wiki/Neutral_Buoyancy_Laboratory

For ISS missions.

Same thing, shorter durations, much more controlled environtment.

Same thing as in underwater as weightlessness replacement.

[night girder]

Look at Scot Kelly.

Been there in 2002 and later he commanded ISS.

[twin dew]

Yes? And how is that relevant to the use case for each of the facilities?

That same people use both for different purposes at different times?

[night girder]

Believe what you want. All I heard was they train/prepare there as astranauts since it mimicks space. Just like the wiki says.

I dont understand this argument since wiki seems pretty clear 😒

[twin dew]

That NBL is the ISS mission spacewalk training facility, and NEEMO:s are "short" R&D missions to Aquarius Reef Base rented from Florida International University for R&D missions to get data how future long range space missions might be done.

[night girder]

Sure.

[twin dew]

Well, NOAA and another university before, and switched at one point.

When NOAA lost interest.

[night girder]

Sure.

You think NASA sends astranauts to NEEMO only for research and not to further train prepare them.

Why send astranauts then?

Not saying astronauts don’t conduct research there. But it’s also training/preparations.

[twin dew]

Because they already are on NASA payroll, have the training to do the work, and are the ones who would implement the stuff in space if any such mission ever gets funding.

[night girder]

And that was totally not my point to begin with.

[night girder]

This was my point.

And then NASA neemo mission statement: “Situated sixty feet below the ocean’s surface, the facility’s structure, living arrangements, and logistical operations bear a close resemblance to those that crewmembers encounter aboard the International Space Station.”

[edgy hazel]

I love cowsay

[night girder]

I hate it.

Whatr shell you use herg?

[edgy hazel]

normal bash

[night girder]

oof 😒

[edgy hazel]

oh no are we gatekeeping shells now?

[night girder]

It's not the most efficient shell 😒

no auto completion etc.

I hate using standard bash once I am used to other shells.

[edgy hazel]

ah yes sorry I need so much efficiency on my gaming PC

[twin dew]

Since when hasn't BASH had auto-complete?

[night girder]

you develop too no?

fishshell.com/

[edgy hazel]

I don't no

[night girder]

Oh, well, if you like to test other shells. Fish is pretty dope.

[edgy hazel]

I am absolutely fine with bash and zsh

[night girder]

Ah, you use zsh.

[edgy hazel]

yeah I use a macbook too

[night girder]

fish suggests commands as you type, based on command history, completions, and valid file paths. As you type commands, you will see a suggestion offered after the cursor, in a muted gray color (which can be changed with the fish_color_autosuggestion variable).

It's different.

[twin dew]

So proactive suggestions instead of standard autocomplete.

[night girder]

Tried to remove CPU cooler. But I think I got the cooler + CPU now 😂

[mental oriole]

Gg 😉

[night girder]

it's a A6 3600!

Taking apart a PC someone gave to me, had to recover some data and I could have the hardware. It's all ancient. Still fun to explore.

Anyone know what is beneath this heat sink?

[twin dew]

North or south bridge?

[edgy hazel]

chipset?

[twin dew]

Would need to see the whole board to say which.

[edgy hazel]

we still had north bridges in 2010?

[twin dew]

Ok, that shouldn't have NB anymore with that CPU.

Forgot it had been stated already.

[edgy hazel]

yeah chipset/southbridge

[night girder]

Oh cool.

Cant take it off without damaging it I think.

[twin dew]

Need to get those pushpins back through by compressing the plastic leaf springs.

On the backside.

So AMD A75 chipset.

[night girder]

Yeah, like you say Baldur.

[twin dew]

So the center is going through, and the sides need to be compressed back in with suitable pliers or like.

And then pulled through.

Or pushed.

When done properly will cause just minor or no cosmetic damage.
And will be still functional anyways.

[night girder]

Yeah, came off fine.

[night girder]

I though the chipset would look ... different.

[twin dew]

Just another chip.

[night girder]

though it would be bigger.

it was so smol, made me wonder if it really even needs cooling. But apparently it did.

[twin dew]

Need for cooling is about power density, not chip size.

[night girder]

ok, still didn't expect a thing so small, to draw so much power, it got so hot, it needs cooling.

[twin dew]

Northbridges were usually made on node one or two generations older than CPUs.
And southbridges were made again on node one or two generations older than northbridges.
So the chip is larger and more power hungry compared to what it could be, but cheaper to manufacture.

[night girder]

The chip was tiny. Like really thiny.

Atleast what I saw of it. Didn't take picture of it.

[twin dew]

17.0 x 17.0 mm?

No, that was different thing

Reference MB for the chipset was 17.0 x 17.0 cm Mini-ITX.

[pure karma]

is there a way to stop forced password changes on windows im starting to get tired of it

[pure karma]

they really should have put that in the settings menu and also have it enabled by default

[twin dew]

Has been the default for me so far?

[pure karma]

its not the default on all 3 of my computers after updating to win 11

[rustic panther]

It's the default for me

Do you have win11 home?

[twin dew]

Don't know about Win11 yet, just all the previous versions.

[night girder]

I need to replace a "NZXT FN V2 Zwart, 120mm" any suggestions?

21dB(A) and max 1.200rpm are the specifications.

be quiet! Pure Wings 2, 120mm seems a good candidate.

[pure karma]

uh no i have pro on all 3

[rustic panther]

Arctic P12 for something you can buy 5 of for 30 bucks, but I don't think they fit your specs

[night girder]

120mm, 3 or 2 pin. 12000 rpm and not louder than 21dB.

let me check air pressure

76,46m³/h

[twin dew]

For what use?

And 4-pin PWM is out?

[night girder]

the be quiet I have in mind is around 87m³/h

[night girder]

the connector only has 2 pins.

[twin dew]

So for PSU fan replacement?

[night girder]

No. Case fan.

[stray badger]

Non pwm arctic p12 should work

[night girder]

I take out the case fan. I look at the connector. I see two pins.

the connector of the fan has three pins.

I assume it's a hub? (It's another build I am prepping to give away)

[rustic panther]

Also, why wouldn't you just get another of the same fan?

[night girder]

but if I look inside the connection, the female (I think) I only see two pins.

[rustic panther]

More out of curiosity than anything

[night girder]

Not available anymore. Was a stock fan of the case itself.

[twin dew]

Yes, fan hub thingie for time when MBs didn't have enough fan headers with fan control.

[rustic panther]

Ah, natch

[night girder]

So, unless I am wrong, it's no use going 4 pin if the other part is only 2 pin?

[twin dew]

You can just not use the fan hub?

[night girder]

Or I could just go 3 pin and put it in? Worked for years like that 🤷‍♂️

[twin dew]

Yes, 4-pin PWM would also work, but those are usually less well designed for DC control, if the headers aren't too close together.

[night girder]

I see, I though 4-pin would be bigger since it has 4 pins.

[twin dew]

It is designed so that the fourth PWM pin can go out of the mating connector.
If there isn't anything in a way.

[rustic panther]

Yeah, IIRC those 2 connectors are compatible with each other

[twin dew]

With that caveat that in some old stuff there can be something in the way of the overhang from the 4-pin

[night girder]

one sec. taking picture of it.

[twin dew]

No need, just 3-pin without the RPM signal.

So just 0V and 12V connected, and the fan hub possibly does DC fan control based on what MB gives.
Or might be just passive hub if it doesn't have separate power cable.

Just physical adapter for one to many.

[twin dew]

Ah, inset, so no space for the PWM connector.

[night girder]

I know a 3 pin can fit it.

[twin dew]

And one of the headers probably has the third pin there.
And the RPM signal from fan connected to that specific header is passed onto the MB.

That isn't really a fan hub, just splitter cable.

Same idea, just longer.

[night girder]

it's a bit of a mess, the cables 😓

and I dont' feel like figuring it out atm.

Is therre leeway for air pressure?

[rustic panther]

In what way?

[night girder]

If the old fan is around 76,46m³/h, should I aim for that? Or is 70-90 good? Or doesn't it even really mater that much for a case fan?

[twin dew]

That isn't pressure, that is flow when the fan is in completely free air.

[night girder]

then CFM is pressure?

[twin dew]

No, just same in different unit.

cubic meters per hour, cubic feet per minute.

H2Omm is often the unit used for pressure.
And that is for when there isn't any flow.

Whops, mmH2O

[night girder]

And does it matter to stay close to it?

[twin dew]

Just gives the same performance in theory.

[night girder]

Just want to keep the pressure inside the case the same.

www.linkedin.com/advice/0/what-pros-cons-positive-vs-negative-pressure-pc-cases

www.reddit.com/r/buildapc/comments/2en24g/positive_pressure_vs_negative_and_pull_vs_push/?rdt=60530

[visual tree]

Opinions? (I am not a programmer 😅 ): www.tomshardware.com/tech-industry/artificial-intelligence/jensen-huang-advises-against-learning-to-code-leave-it-up-to-ai

I mean, I used ChatGPT to create a working python script but I guess you still need programmers for complex tasks

[night girder]

Hey, I got an opinion

[dire igloo]

"knowing how to code" doesn't mean being able to write in a specific programming language.
It's identifying how a piece of code works, documenting what your piece of code does (so others understand it should you die unexpectedly) and being able to create efficient processes.

You wouldn't say that cooks are useless because an AI can write recipes now

[night girder]

For now, developers are still needed. (And even if we have AI, AI needs developers, unless AI is going to develop AI). They have a few things to overcome before we can replace developers with AI.

[visual tree]

People in the comments section are really pissed

[night girder]

But I think it's possible in the future, developers are replaced by AI. Are a lot of them are.

Just like how we use to make stuff with our hands, in factories, and we all got replaced by machines.

There was a bit of panick back then (industrial revolution), but in the end. Yes a lot of people lost their jobs, but we still moved forward.

[dire igloo]

The biggest issue with AI code is that you need humans to understand how the code works

[night girder]

Does it matter if it works?

If you want a website, and a AI makes a website, does it matter how it's coded if the end result is "perfect" for the costumer?

[dire igloo]

Security experts are concerned about zero days that are introduced through AI written code that nobody bothered to check

[night girder]

That's a valid point.

But then they will create AI that check for that. Security AI checking developing AI for example.

[dire igloo]

Except you can't adapt to the customers needs should they change because you have absolutely no clue how the website was actually built

[night girder]

You don't change it.

That's the thing, stop thinking as a developer 😉

[dire igloo]

The customer is human

Their wishes will change

[night girder]

Costumer -> AI -> Website (no developer in between).

Costumer -> new feature -> AI -> website.

[dire igloo]

Nope

[visual tree]

Here, people are telling everyone to work in the IT sector in order to earn a lot of money. Sadly, the IT sector is collapsing atm

[dire igloo]

That's the thing that won't work

[night girder]

Costumer changes mind -> tells AI it wants something else -> AI adjust websites. That's the future.

[dire igloo]

Because AI is just text prediction

[night girder]

At the moment.

I am talking future for context. Not what we can do now.

[dire igloo]

It can't adjust it because it'd have to be told the entirety of the website

oh, another thing: Screenreader compatibility and those things

[night girder]

I am sure, that when they make an AI specific for developing websites. They will all patch this out.

[dire igloo]

There's German legislation for digital accessibility, BITV 2.0

[night girder]

https://www.youtube.com/watch?v=cMb-dg_vZmQ

[dire igloo]

At some point in the future, AI might be able to do this

[night girder]

~~People ~~ Experts though AI wouldn't be able to do video for decades.

Now 10 days ago, AI made video.

[dire igloo]

What was the time needed to create those videos?
How many "prompt artists" were employed?

[night girder]

Even scientist don't understand how they did it.

[night girder]

Fromt he podcast I listened too, they don't tell us how they did it. They just wanted to show all the nay sayers it's possible.

[dire igloo]

Because OpenAI are shady and intransparent as shit about it

[dire igloo]

Exactly

[twin dew]

So can just be lot of fakery and human work involved.

[dire igloo]

Fwiw, it could be manmade and a marketing stunt.
Unless we see it actually in action, it's about as trustworthy as some cryptobro telling you that peepeepoopoocoin is gonna be the next ETH

[night girder]

If you want to take that risk as a company to release a fake video saying you "cracked" AI video generation.

Sounds like a big risk, but possible.

[edgy hazel]

I don't think it's manmade

fucking up perspective like that is insanely hard

[night girder]

openai.com/sora

[twin dew]

Each frame can easily be AI generated, and still have lot of fakery and human work involved in generating the video as total.

[night girder]

But to go back to the discussion, can AI replace developers, for sure.

Is this the end times and are all developers going to get fired. No.

[dire igloo]

It can replace some developers at some point in the future

[visual tree]

Guess we will find out once they release Sora to the public (IF they release it)

[dire igloo]

It's the introduction of a new tool

[night girder]

Also, I thinkt he tasks for a developer going to shift more and more.

A developer these days isn't pure code anymore. You have devops and all these cool things.

[dire igloo]

Did the invention of the tractor replace farmers?
Did the invention of a sawmill replace lumberjacks/woodworkers?

[night girder]

A developer needs to be a tester, a analyst, the more they do, the better and cheaper for the company 😄

[jagged snow]

This is no different from automation improvements we've had in the past, except this time it's coming for the white collar workers instead of the unskilled labor

[night girder]

Why would a tractor replace farmers? You think farmers were pulling the plough?

[dire igloo]

Exactly my point.
AI just writes code, that's all.

A recipe-writing AI will replace a cook as much as a code-writing AI will replace a programmer

[night girder]

Tractors replaced farm animals.

[dire igloo]

Then replace tractors with combine harvesters

[visual tree]

Damn, I was planning on studying computer networks....

[dire igloo]

People used to use scythes, now we have combine harvesters

[night girder]

Also, don't forget.

[jagged snow]

No, but the tractor did reduce the number of hands that were required to farm the same amount of food

[edgy hazel]

only took like what? 4000 years?

[night girder]

There will be stuff they will not let AI touch for a while and let humans do 😉

[jagged snow]

By an order of magnitude

[night girder]

Let's think about hospitals for example*. They don't want AI to maintain the code for hospitals. Too critical.

So human developers for sure will be needed there.

[dire igloo]

Any form of critical infrastructure

[edgy hazel]

our society is way too focused on short term profit. There is no way ths shit is done in the next 20-30 years

[dire igloo]

Government too

[night girder]

Developers aren't going anywhere, tasks might change, the scope, some specific type of developers might dissapear.

and AI will need development 😄

So, I would still suggest to the younger generation that has an interest in developing, to do it.

[edgy hazel]

also (THANK YOU PETA) you can't copyright AI generated stuff

[night girder]

Just try to approach development as broad as you like. So learn as much technologies etc you can and are interested in.

My opinion ^^

[dire igloo]

Also, learn to work with the tools you have and to work without them

[visual tree]

Totally agree with this. Happened at the previous companies I worked for.....

[dire igloo]

A good programmer can be fired at any minute without causing any extra trouble for the company - but they aren't because they're good programmers

[visual tree]

I missed the conference 2 weeks ago when one of the ChatGPT creators was in my city (Zack Kass). Not sure if he was actually one of the creators but I would like to hear stories from his perspective

ChatGPT has helped me a lot in automating some menial task at work I don't want to deal with

Could have asked a programmer around but didn't have any cash with me to buy him some beer

And they are probably too busy to waste time on me

[twin dew]

And how confident are you that the solution it gave actually does the right thing for the whole input space?

[visual tree]

Was the question directed at me?

[twin dew]

That it actually functions correctly for the possible inputs to whatever was outputted by ChatGPT.

[edgy hazel]

you... check it...

it's a tool

being used as a tool

I don't hammer a nail in the wall and don't look at it after to check if it's actually in

[visual tree]

Not sure if the question is directed at me but the code ChatGPT made did exactly what I needed in the end. It needed to do some additional tweaks to the code though

[twin dew]

Plenty of people seem to use ChatGPT without checking the output.

So you did do manual pass on the output before use.

[edgy hazel]

yeah and then it fails and they need to change it again

[edgy hazel]

Yes?

As you do with everything?

[twin dew]

There is now at least one case in US courts where lawyer used ChatGPT to create legal brief, and sent it without any checking to the court...
And it contained plenty of hallucinations in it XD.

[jagged snow]

2nd or 3rd iirc

[edgy hazel]

Yes and they failed. It's like not tasting the food when you are a cook

[night girder]

They didn't just develop AI and released it on the public without testing it

[edgy hazel]

ofc you get fired if everything tastes like shit

[night girder]

Develop an AI that can make websites. Test the results of the AI untill you are happy. Release the AI and profit.

The costumers will surely check out their own website if ordered.

[edgy hazel]

Doesn't like all AI say it can contain errors and you need to check it yourself?

[night girder]

And if something is wrong, users will report it. Just like it happens now.

[twin dew]

And fixing a "bug" in AI would mean someone would need to understand how the trained neural networks work.
And that isn't true in most cases.
The NN:s are just black boxes that cannot really be edited, only retrained hoping for it to get into better state.

And mostly for ChatGPT, it has just been adding classic chatbot in front to sanitaze the inputs to the AI part.

The cases where you get canned "cannot answer" reply or like.

Those newer hit the actual AI.

[night girder]

ChatGPT just had a bug a few days ago. They fixed it.

[jagged snow]

That bug was either not with the nural network or they just rolled the network back to an older version

[night girder]

👆

[jagged snow]

Yeah, I know

I read the headlines

There were quite a few theories floating around, and nobody knows what actually caused it

[night girder]

Ah, that gives me so much confidence 😂

[jagged snow]

Exactly my point lol

[twin dew]

That is Neural Networks.
No-one know how each works.

Which makes them impossibly to verify.

[night girder]

Eh... 🤨

[jagged snow]

Well that's true

You can know the inputs, outputs, and training method/criteria

[night girder]

You saying, the scientists don't know how their own neural networks work?

[jagged snow]

Yup

[night girder]

Then why do I keep hearing scientists talking in podcasts about discoveries they make on how their own* neural network works?

[jagged snow]

They know how they built the network, and they know how networks in general work, but they don't know how each individual network actually does what it does

[night girder]

I know fully well about the input and output. But: input is 1 -> blackbox -> output is 5. input is 5 -> blackbox -> output is 9. One can safely assume eventually, that what happens in the blackbox, is +4. You reverse engineer it. That's what I heard they do with their own neural networks.

They keep testing inputs/outputs to understand what happens in the middle. And they are making discoveries.

[jagged snow]

That works with simple networks, but it becomes extremely impractical on the scale of a large social media platform, for example

With something like the youtube algo, they just monitor it to see if it's having the results they want

[night girder]

And they think* they know how the neural network works 😄

[jagged snow]

If it's not, they try something else
if it is going in a direction they like, they do more of what they already were

[jagged snow]

There's a difference between knowing what it's doing and knowing how it's doing that

That's the key distinction

[twin dew]

And even for the trivial one, if you don't test ALL the space, there might be some corner case that doesn't follow the simple rule you seemed to have.

[night girder]

They don't care about the how atm.

[twin dew]

Where giving it 213712317232131321 as input for example would suddenly do -4 and not +4.

[night girder]

They want to know what. And if that is 99% stable, they don't care about the how.

Aslong as chatGPT is doing what they want it to do. Without too many weird edge cases. They seem to be content.

[jagged snow]

I know
I was just responding to your comment where you used the word how#off-topic-tech message

[night girder]

Ah, gotcha 🙂

I just don't believe they are totally in the dark about what's going on.

I think the truth is probably somewhere in the middle, some ideas (we know how it works), and sometimes no f'ing clue 😂

Let's not forget it's still fairly new technology.

[jagged snow]

Yeah, they can't be totally in the dark or they wouldn't be able to make improvements
The difference is that you can't just look at a neural network and figure out exactly how it will behave in any scenario without having to test it like you can with any other pice of code

[night girder]

Nah, that's not possible.

I think also that's why they released it so quickly, since they realised, only testing it would lead to further knowledge what's going on.

[twin dew]

Mostly the improvements are done by reinforcement training.
Feeding the model more specifically selected data and hoping it "learns" right things from that and changes the output to more suitable direction.

[night girder]

and what's better than a whole world testing your product? 😄

[twin dew]

And for ChatGPT, often by that front simple chatbot filter to not just handle the wrong cases.

[night girder]

It's not 100% guesswork and magic 😉

I actually believe, there are very smart scientists who know quiet a lot about the neural network of ChatGPT for example.

[twin dew]

They know about the base algorithms, not the trained network.

[night girder]

But yeah, it's like understanding a brain perfectly.

[twin dew]

And about how to select the training data to try to steer the NN development.

[night girder]

for chatGPT, I know it's trained by the employees.

And they have guidelines on how to train it. Be polite etc.

Try to be neutral (this one made me chuckle when I heard it).

because it's so hard to be neutral, if everyone in this chat would ask the same question to chatGPT and read the answer. And then has to fill in a survey about the quality of the answer, I am sure we are not all on the same page. Because that's normal, we are opinionated. So how is chatGPT even able to train their AI in a neutral way without any bias 😒

Once it was properly trained by the employees etc. They released it on the world. So now it should be getting better feedback and be more neutral.

[twin dew]

That doesn't seem to match training really.
That is testing to see if the trained network version is something they can release or not.

As after each retraining round, they have no idea if the result is better or worse before testing.

If the stuff the employees did was actually part of the training, it would take ages to get any changes.

As in YEARS.

[visual tree]

Guess there is still time to become an electrician. There is a huge shortage of electricians here or handymen in general. Good luck trying to automate that AI!

There is a huge demand during summer for HVAC technicians. You will usually have to wait for months before they can visit your home for repairs

Since I am afraid of heights, I don't mind AI replacing HVAC technicians but that probably won't happen anytime soon

[night girder]

AI and automation go hand in hand 😉 More AI developments benefit robotics developments.

[pure karma]

AI is great when used for pushing things forwards

like isent nanite for UE5 entirely AI bassed for the most part

[night girder]

Saw some gifs on reddit of robots who lay down a floor now!

[pure karma]

construction companies are getting desperate i see

[night girder]

https://www.youtube.com/watch?v=akP2dKBd2HE (this is old video, but can't find it back)

And then you got another robot that can do the tiling!

[visual tree]

Hope there are no robots who can climb ladders and install AC outdoor unit

I have serious fear of height and would faint while climbing lol

[pure karma]

i mean wouldent that just be a automated crane

[night girder]

I know it's important to try to pick the best job or education. But you can't predict every outcome of the future.

[night girder]

https://www.youtube.com/shorts/zTD6r_f8yD8

[pure karma]

id love to see a robot trying to fight a lobster

[visual tree]

I don't understand why accountants laugh when I tell them accounting has a high risk of automation. Guess they do something that can't be automated, who knows 🤷‍♂️

[visual tree]

Yeah, government here was advertising IT industry as a dream job and everyone wanted to study computer science

Until everything fell apart and people aren't interested in IT anymore

[edgy hazel]

Isn't like programmer the only job getting fucked in the ""near"" future? All other fields are fine

[visual tree]

Probably. Here, people be like: "You don't need a college degree when you can take programming courses and become an expert in only 6 months"

Until they realised it's not that simple

[edgy hazel]

I don't see anyone willing to give out their networking stuff to an AI anytime soon. Too many variables and too high risk. Infosec too. And anything with hardware needs a guy aswell

[visual tree]

Imagine AI creating so many security holes that your network starts to look like a swiss cheese lol

[edgy hazel]

"bing just blocked our production network, we need to go into the datacenter to fix it manually"

[night girder]

Tax fraud 😂 You can't make an AI that does that haha 😄 (legally)

[visual tree]

Guess I won't be mentioning that to the people from the accounting department unless I want to end up on their blacklist lol

[night girder]

Haha 😄

You know the best way of learning is by making mistakes. I just wonder the severity of mistakes they will make in the future when it comes to AI.

If they are smart, they would step on the brake and slow down to be safe.

[visual tree]

There are enough of examples on the internet already about AI chatbots doing "unexpected" things

[night girder]

Yes, but unexpected.

Or an AI that opens the flood gates of a reservoir or two different things 😂

If we implement AI in our everyday life and in our society, it can cause a lot of harm if something goes wrong.

Just like we had to have a nuclear reactor blow up first, before we learned from it. (Chernobyl)

Now they build nuclear reactors closer to coastlines for several reasons etc etc. That sort of thing. How big is the mistake going to be, before we develop AI with clear and calm heads.

[visual tree]

Hope I will be alive by the time fusion reactor becomes commercially viable

[night girder]

I know we going to f'up in the future. That's humanity. But just wonder how bad. I hope it's not too bad like Chernobyl.

[night girder]

https://www.youtube.com/watch?v=NqabT21d8VM Have you seen this video?

[visual tree]

We were already on the brink of extinction https://en.wikipedia.org/wiki/1983_Soviet_nuclear_false_alarm_incident

If that's true, Petrov deserves a medal of honor

[night girder]

One of the potentional outcomes of this research could be, that we can mimick the nuclear reaction of the sun and have a sustainable energy source. (but that's very very whishful thinking).

And in the end, all we do is put pressure to materials 😄

[night girder]

I heard this story was very true 🙂

But history is written by the winner.

[visual tree]

I hate how the longer you live, the higher the chance something like this could happen again

[night girder]

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

[visual tree]

It's only a matter of time until something goes terribly wrong

[night girder]

The more humans we get, the more chaos.

If we zoom out in a macro level, space is getting more and more chaotic and random too.

[tough owl]

im kinda shocked as to how many apps have arm binaries

[night girder]

science.howstuffworks.com/entropy.htm

[night girder]

When I heard this term explained by Neil deGrasse Tyson, I also realised that this applies to humanity. And it's pretty logical. Since it also involves a bit of chance calculation I think. The chances you have 1 crazy person if there is only 2 humans is lower than you have 1 crazy person if there is 1 billion people for example

[visual tree]

The real question is: will we learn our lesson before we destroy ourselves?

[night girder]

This is my opinion, not based on facts. But if I look at history, I don't think so.

One would assume we don't have war anymore after all these wars throughout history, but no, we continue.

One would assume we would be more carefull with new technology, after all the disasters we had through history, but no.

We still had someone built a submarine and test it out to check the Titanic.

While the Titanic ITSELF should tell the tale of not underestimating nature, but no, in the water they went with their submarine and console controller.

[visual tree]

Btw, I was never good in physics but this keeps me awake at night lol

[night girder]

Why?

[visual tree]

Not sure if there is some explanation for this

[night girder]

That subtext feels a bit misleading.

"as if the electron knows it is being observed" would piss off so many scientists 😄

I encountered the double slit experiment when I was trying to understand Quantum entanglement.

[visual tree]

@inner shale Can we borrow your cat for Schrödinger's cat experiment?

[night girder]

Here is another video on the subject asablic: https://www.youtube.com/watch?v=vyP80vLos60

Sometimes I whish I still was at school so I could ask a prof what is bullshit and what isn't 😂

I am still pondering over the screenshot you send.

[visual tree]

If I was conducting that experiment, there would be no patterns

Because I only did bowling once and 90% of balls ended up in the canal

[jagged snow]

Typical bowling experience

[visual tree]

Definitely not doing bowling again lol

[night girder]

You didn't put up the side barriers for kids?

[visual tree]

There are no barriers

[night girder]

We can put up side barriers at most bowling alleys, so kids who bowl, don't end up in the canal/gutter?

it's all automatic. Based on the profile of the person who's turn it is.

So they go up and down depending on who is about to throw 😄

[night girder]

Shit, I though this was common for bowling 😄

[visual tree]

I had a traffic accident and my coordination is terrible because of it

[night girder]

Oh, sorry to hear 😒

[visual tree]

Can't even catch a ball when someone throws it at me lol

[night girder]

On the bright side, you won't ever be drafted 😄

[visual tree]

But that could be because I got hit in the head by a medicine ball in P.E. and don't like balls in general

[jagged snow]

Oh lol

[visual tree]

Felt like someone punched me really good

[jagged snow]

You've had some interesting experiences

You ever been hit by a fast thrown baseball?

[night girder]

My mum had a seizure once. Changed a lot for her too. Mostly her balance.

The brain is finicky.

[visual tree]

God, no lol

[jagged snow]

Not a fun experience

[visual tree]

My worst game is dodgeball since bullies love throwing balls at you really hard

[jagged snow]

Although also far from the most pain I've experienced in the name of fun

[jagged snow]

I'm a bully 🤚
I take great enjoyment in that

[night girder]

You ever played basketball and got a pass from a teammate straight in the face? 😂

[jagged snow]

Never played much basketball to be honest

[visual tree]

Sorry to hear that

[jagged snow]

I've shot hoops and played a few pickup games but not much else

[night girder]

It's nothing compared that she is still alive, if you know what I mean? In the context of things, it could have been far worse 🙂

[night girder]

Sport is just pain.

https://tenor.com/view/dodgewrench-dodgeball-gif-7601197640735809486

Oh, I loved that movie. Dodgeball.

[visual tree]

I liked using gym during P.E. but that was rare because the P.E. teacher forced us to run and play football only....

*and sometimes basketball

[night girder]

Yeah, from all the sports that I played, it's my favorite.

[visual tree]

Also, Nintendo Ring Fit actually helped me lose some weight although it took some time

Lost interest once I beat all levels

[night girder]

Oh, I forgot. Chess is a sports no? 😄

[visual tree]

I think I burned like 20k calories

They also consider mental games as sport too afaik

Going back in time, don't know how I pulled this off tbh 😅

[night girder]

Dedication 😄

[visual tree]

I think it was more because I wanted to know what happens if you beat levels

If you forced me to go to a gym and do the exact same exercises without Nintendo, there is no way I would do it lol

[edgy hazel]

i play ring fit to see the sexy dragon

[night girder]

@visual tree Look, if you disable the observer. You get the wave pattern back 😄

[edgy hazel]

very cool heres a cat at night

                        . .

[night girder]

Pattern when observer is enabled.

[visual tree]

Wonder if someone conducted a test by disabling a detector first and then enabling it later

But I assume the most logical result would be 2 patterns

[night girder]

Ok, so this just confirms it doesn't it?

We are in the Maxtrix.

[visual tree]

Guess we will never know

[night girder]

"It's kinda of like how in a video game, the environment and object only load when the player focuses on or interacts with them". - from the YT you watched.

From what I have heard of it, I don't really believe in the explanation of probability.

Even if a particle has a probability of being observed at a certain location in space, that doesn't explain how the probability changes when it is observed.

[maiden coyote]

What's off?

Why does it just drop next to land o.o

[night girder]

probably slowly erroded by the water.

[visual tree]

My brain hurts from thinking

[night girder]

Something is off with this experiment.

Like, there is a missing puzzle piece.

[wanton orchid]

YouTube is one of the worst possible source to learn about that stuff almost any single time I see a video talking about that kind of thing it's straight up bullshit

there are hundreds of bullshit ing specialists out there and that kind of poorly unknown field is a very nice target for them

[pure karma]

great youtube patched ad blockers again

lets see how long that lasts for

[wanton orchid]

great ublock origin still work around it on firefox

[edgy hazel]

firefox gaaaaang

[night girder]

https://tenor.com/view/doggo-meme-firefox-gif-14524252

[wanton orchid]

btw atomic stuff is all about statistics simply because you are manipulating extremely high speed extremely high quantity/density of elements interactions
so all we can/want to focus on is how it generally works
then baby step corner cases from that because you can create tools that generally isolate some elements interactions
and then work/learn about that subset
almost every body forgetting this is all a big statistic models
and most confuse aspects of it, because of different scale and technics of measurements
which even lead to some scammer going inside these interpretation breach to say it's "mysterious behavior"
when it's just the model that switchs

[pure karma]

yea im also on firefox now

mostly because it was more compatible for customisations

[wanton orchid]

fuck you "manifest 3"
more like Google "many feast 3"

[pure karma]

the only place i still main chrome is on my school laptop because its 4x more resource efficient when working whit 4gb of ram

[wanton orchid]

it's not even correctly documented nor implemented

it's a scam basically

(I built manifest 3 extension on Google chrome, and worker basically stop registering listened events as soon as the worker process sleeping, literally fails being capable)

[night girder]

Like I said, something is off in the theory. "Unexplainable" doesn't cut it. Either the experiment is wrong. We don't know something. Or looking at it from the wrong angle.

[wanton orchid]

when using background page as in manifest 2, all good

[night girder]

https://en.wikipedia.org/wiki/Double-slit_experiment This I all believe.

[wanton orchid]

if you see someone focusing very hard on making sure to not sort out something
then it's basically the opposite of science

[night girder]

I just can't understand/believe that placing an observer would change the behavior magically.

[wanton orchid]

that is the basic not like the YouTube screens
it behave in one way
but have 2 mixed explanations possible

[night girder]

Here is the observer effect https://en.wikipedia.org/wiki/Observer_effect_(physics)

but as you can see, even the page is slim and uncertain.

Read this: "A notable example of the observer effect occurs in quantum mechanics, as demonstrated by the double-slit experiment. Physicists have found that observation of quantum phenomena by a detector or an instrument can change the measured results of this experiment. Despite the "observer effect" in the double-slit experiment being caused by the presence of an electronic detector, the experiment's results have been interpreted by some to suggest that a conscious mind can directly affect reality."

[wanton orchid]

observer effect is a basic too
but not relating to observing a wave

btw don't you get how stupid that is ?
"if you don't observe the result then you can see observe that the result changed"

[night girder]

"observer effect is a basic too" wdym?

It seems flawed to me.

[wanton orchid]

the tool used to observe the experiment is directly part of the system

when you account for a result you must consider the whole processing chain

[night girder]

"However, the need for the "observer" to be conscious (versus merely existent, as in a unicellular microorganism) is not supported by scientific research, and has been pointed out as a misconception rooted in a poor understanding of the quantum wave function ψ and the quantum measurement process." 👈

[wanton orchid]

that is so basic that trying to imply "weird things" from that is pointless

[night girder]

I just think we don't know what is going on honestly.

[wanton orchid]

be aware that lot of people wants to get attention by building on confusion
and that is one for most people not familiar with not directly observable phenomenon

[night girder]

I wonder how many times the double-slit + observer experiment has been done. And how legit the results are.

it's not something we can just test out in our room to verify 😂

[wanton orchid]

it's been done so many times on tv (irony)

[wanton orchid]

except you mostly can

[night girder]

double-slit + observer changing behavior? with atoms? you can measure that in your room?

I am not doubting the double-slit experiment itself.

[wanton orchid]

you will just observe no change in behavior

[night girder]

It's the + observer part that I don't believe.

[night girder]

👆

That's what I am talking about.

[wanton orchid]

btw do you know how old tv works ?

[night girder]

not really

[wanton orchid]

an electronic linear accelerator with reactive phosphorus coated glass that then emit light upon receiving electrons

[night girder]

yeah, I don't have that around me atm.

[wanton orchid]

all to say that almost any family had the basic equipment for that experiment at their home already
but remember

it is MYSTERIOUS
AND only Einstein and the tv commentary friend could dar be able to do it
especially since Einstein is dead and can't call them scammers (anymore, because he actually did at the time, but sssshhh)

there is an USA episode where some lawyer tried to change the value of pi as by the law
(it thankfully failed)

but that is if you believed too much in society

[serene charm]

Late night realisation that firefox is the android of browsers

Goodnight

[jagged snow]

Firefox is also the best android browser

[maiden coyote]

It sucks on iphone... oh.. that's just the iphone

I can't really tell if chromium or Firefox is better on Linux.

[tough owl]

both are fine

I just refuse to chromium based browsers out of principle

[wanton orchid]

I would arg that it's usually easy to get ungoogled chromium on Linux now
but it means no syncing
I hope someday there will be an actually third party syncing service standard support in browsers

[twin dew]

As on iOS, all the browsers are just reskins over Safari. For those who didn't know.

But seems to be now changing?
iOS 17.4 finally allows for alternative backends because of EU ruling.

But only for EU users.

[wanton orchid]

the operating system that forces you to use whatever it wants you to use

[tough owl]

we love webkit

[edgy hazel]

safari the only browser not killing my macbook battery 😦

[tough owl]

intel moment

[twin dew]

Have you tried Edge yet?

MS has AFAIK made lot of work on their Chromium fork to get it not be battery hog.

[tough owl]

edge 🤢

[night girder]

From 17.4 Apple no longer supports webapps in EU.

Because of the digital market act. Now EU is opening investigation into Apple because of this.

[mental oriole]

At this rate it's about time to ban them from EU market.

[night girder]

I am tiny bit confused with why they are stopping with PWA in the EU

[twin dew]

Because EU forced them to open other parts of their ecosystem, and are now retaliating?

[night girder]

I read that the act forces apples that other vendors (Firefox, Chrome, …) are allowed to install PWA. And Apple doesnt want that since there is no profit in it.

Also apple states that nobody uses PWA and the impact is neglectable.

This last part is what the EU needs to investigate

Gotta love being stuck between a tech giant and a giant goverment

[white kraken]

I think the official reason is, that they cannot guarantee the security if non-Safari Browsers are providing WebApps.

[dire igloo]

The official reasoning is always security bogus

You can just put it behind a warning and maybe some menu hiding shenanigans

Make it difficult and obnoxious to do, don't just prevent it entirely

[white kraken]

Yep.
bUt SeCuRiTy!

[wanton orchid]

note that you still can install some apps some other way on mac
it's worse security wise
also any apps have to provide the part Apple does not provide
and it's as tedious to guarantee that part is secure in itself
than it would for any other app specific part such as internal web engine
as long as the operating system itself can't be compromised which is easy to check
example :
Apple will not check that a remote discord friend cannot delete your messages in discord

it is internal to discord
likewise it should not arg that it should check for browser bookmarks to load the correct web ressource
it's the browser app responsibility

So saying it's for security is bullshit
because that's the browser responsibility
browsers that can't consistently endorse that responsibility since they are not allowed to use whatever secure engine they want

[night girder]

That's the thing, they don't need to guarantee it if the user uses 3rd party software.

If I use a product of google on my iPhone, and there is a bug in that product that leads to security risks, it's the responsibility of Google and not Apple I believe. Making this a bit hard to believe 😄

[wanton orchid]

demonstration of what you just said

[night girder]

Apple just doesn't like to play along with the other kids 😄

It's a very closed eco system, always been. And they only open up this system with some sort of internal or external pressure.

10 years ago, their main excuse for this was that if you use apple hardware + software, it all just works because apple works with apple. Now their main excuse seems to be security and not compatibility anymore.

I am pretty happy with this pressure. Apple is a bit too closed for my comfort 😄

[dire igloo]

People will still say "I bought Apple for security and still got hacked, Apple should've prevented this."

That said, there are still vulnerabilities in Apple's stuff

[dire igloo]

Because compatibility has improved to the point of that reasoning being easily exposed as a cheap excuse, so they had to find smth else

[night girder]

Also, the others just catched up. And times changed. For example, I had an iPod and it worked really well with iTunes. But iPods are from the past now and replaced by cloud streaming.

[dire igloo]

Remember the times when you bought an iPod Touch for gaming because Android phones didn't have the performance?

[night girder]

The first iphones had almost no competition when it comes to UX. The touchscreen felt so good.

[wanton orchid]

especially Safari which is lagging behind in number of security features at ressource access level and at TLS level

[wanton orchid]

it still is somewhat true in that regard
because Linux rendering api is still cluttered
(but improving with Wayland etc)

[night girder]

Why is BIOS* so poorly documented? What am I doing wrong?

[wanton orchid]

because manufacturers sucks at documenting even the directly user interface things

[night girder]

All I want to know, wth "nitro control line" means. And it's just a pain in the butt to find this information on the internet.

So how am I suspose to know what setting it's suppose to be on?

[wanton orchid]

where did you find that setting

[night girder]

it's not in their manual.

[dire igloo]

Any context?

[night girder]

It's a setting in the BIOS* - corrected it. Not motherboard itself.

[wanton orchid]

cpu performance ?
pch ?
power management ?
thunderbolt setting?

[twin dew]

https://old.reddit.com/r/Amd/comments/155mddw/new_agesa_1007b_performance_update_6200cl30_vs/jt1s5x3/?context=3

[night girder]

LOOK!

[dire igloo]

What is it supposed to do?

[night girder]

3 results if I add bios haha 😄

Wtf?

[twin dew]

AGESA memory settings that got added in that AGESA.
Just another latency setting that you normally don't touch.

[wanton orchid]

you still didn't answer where you found that setting

[twin dew]

That Reddit thread I posted has the AMD OC person answering that those new NITRO settings do for some parts.

[night girder]

but I want official documentation what all these settings do, not reddit and hope these people know what they are talking about.

[twin dew]

Including the RX, TX and Control lines.

[night girder]

How is a user suppose to know the default settings and what is what without a manual?

[twin dew]

Doesn't exist for almost any platform unfortunately publicly.

[night girder]

Every single electronic device, with a UI and settings, documents it. Except the BIOS for some reason? 😒

[twin dew]

Neither AMD or Intel.

Yes, it is shit situation, but it is what it is.

[night girder]

I don't understand it. Cars have more settings than BIOS and every single setting has 1-5 pages written about it.

but for something more complex, they don't add it?

So all I can do is ask on reddit, hey anyone "backwards engineered" this setting?

"And also, is my settings in the bios correct then?"

[wanton orchid]

neither asrock nor Asus nor gigabyte nor msi comments most motherboard settings
it sucks so much especially when half is named after fancy stupid marketing and the other half is wrongly translated

[twin dew]

Because only the BIOS & AGESA coders know what the stuff does, and haven't documented the stuff even for internal use.
So no-one lower in the pole can write documentation for public.

[night girder]

Sorry for this. I was going through my photos and I took a picture of some bios settings as a reminder to look up what it meant.

and after some searching I found nothing, annoying,.

[night girder]

This really irks me, ngl. and we pay big money for a PC. So they have no excuse.

[wanton orchid]

it did for anyone that finally got to actually touch the settings and not just rely on Asus marketing about how customizable their board are

[night girder]

Even reading this makes little sense to me. "then setting >6000MT/s DDR5 Nitro Mode kicks in automatically, but you can try more conservative settings (max out Nitro TX/RX/Control values) if 6200/6400 1:1 doesn't work with the defaults. Some parts will max out at 6000 1:1, but with tuned timings most workloads are not going to suffer at all."

[wanton orchid]

got burned by Asus on that
not even taking marketing seriously but even basic settings such as fan ramp up was not working nor documented properly

[twin dew]

Because those are internal settings that got explosed for extreme OC.
Up to 6000MT/s the AGESA uses one set of settings, for over that it uses different set, by default.

Like UCLK:MCLK is 1:1 below and 1:2 above.
Nitro Training (more extensive, but longer taking memory training) isn't used by default below, is used by default above.

And the three memory controller internal timings get loosened by one step above by default.

So those three got exposed so you can loosen them even more when trying to get RAM OC stable.

[night girder]

So I just have to assume Ntro RX Data "2" and Nitro TX Data "3" and Nitro Control line "1" works better than Nitro RX Data "Auto" and Nitro TX Data "Auto" and Nitro Control line "Auto"

[twin dew]

No, Auto means 1, 2, 0 with MCLK 6000 or below, and 2, 3, 1 on MCLK 6001 and above.

[night girder]

😒

[twin dew]

No difference in setting them manually.

[night girder]

Then why the fuck is it on 2 3 1?

[wanton orchid]

? not on auto ?

[twin dew]

ASRock?

[night girder]

no!

[twin dew]

BIOS writer as MB maker can override the AMD defaults.

[twin dew]

S Jackson speaks about the AGESA defaults.

Not what ASRock might have set their BIOS to do.

[night girder]

And that's all "stock" / default.

[twin dew]

So ASRock is setting the AMD recommended "Over 6000MT/s RAM" settings at all times by default.

[wanton orchid]

that's some latency buffer timings ?

[night girder]

I don't know.

[twin dew]

Yes, inside the memory controller.

[night girder]

I think this motherboard is putting wrong values as a default maybe.

[wanton orchid]

bigger means more stable/buffering ? (for all three settings)

[night girder]

since I read other people putting it on auto or have it on auto.

[wanton orchid]

looks fine then

[twin dew]

Match to the "Over 6000" settings Jackson talks about, with Nitro Training enabled and the specific "latencies".

When 6000 and under on Auto would mean Nitro Training disabled and the latencies one step lower each.

[wanton orchid]

it's on the safe side

[night girder]

Eh ... I hope so. Today my new CPU arrives. So, would suck to find out the crashes are caused by a wrong setting 🙈

[twin dew]

Causes longer time for memory training, and little bit worse performance.
By settings meant for extreme RAM OC.

[wanton orchid]

I expect safer settings to allow for lower voltage at lower speed (<6000)

because it sounds like more buffering of lanes

[night girder]

But, if Asrock wants this.

Then why not put it on Auto and then do 2 - 3 - 1?

[wanton orchid]

because then it's still 2 3 1 hard training at lower speed too

[night girder]

RX -> AUTO (but really, Asrock decides it needs to be 2). TX -> AUTO (but really, Asrock decides it needs to be 3) etc.

Isn't the point of Auto it adjusts (automatically) depending on other settings/hardware?

[wanton orchid]

Auto does not mean most reliable

[night girder]

No. But it means automatically?

So I would assume it would scale depending on context.

[twin dew]

The automatic behavior/rule is - [ IF DDR5 >6000MT/s THEN RX=2; TX=3; CTRL=1, ELSE RX=1; TX=2; CTRL=0 ]

The inside AGESA coding for those when on Auto.

[night girder]

but DDR5 isn't 6000MT atm, afaik.

[twin dew]

Yes, ASRock has forced them.

[night girder]

So, Asrock forces a setting for 6000MT, even if the RAM isn't clocked for that?

Just to be safe?

[twin dew]

For 6200+

6000 would by still use the other set by default if AGESA was left to do what AMD coded it to do.

At least Asus leaves all the AMD AGESA three menus at Auto for everything...

ASRock and Gigabyte both set stuff in there for some stupid reason...

Not sure about MSI.

[wanton orchid]

would not be surprised if msi leaves everything default too

[twin dew]

If ASUS want to "override" something, they have the same setting under their own menus set to whatever.

[night girder]

RAM is set to DDR5-4800 atm.

[wanton orchid]

current amd default is quite performance focused

[charred pewter]

on intels, almost everything in MSI bios is 'auto'

[twin dew]

And it is ASRock BIOS coders doing stupidity, and forcing the 6200+ settings at all times by default.

Same with them forcing 1:1 UCLK:MCLK, 2000 FCLK and so on.

[night girder]

In theory, this could cause instabilities?

[wanton orchid]

why the fuck forcing 2000 fclk by default

[twin dew]

🤷

[wanton orchid]

I mean most stable ram performance tuning ok
but 2000 fclk is neither most safe nor most stable

[twin dew]

And the expectation is that Letterdiefs CPU damage is from ASRock force-setting VSOC 1.35V by default in the BIOS version he used first for a moment.

Before flashing newer version with system booted.

[night girder]

With the shitty part being, that it hasn't crashed for a week now 😒

[twin dew]

When it should have been Auto, which would have set 1.00-1.05V

[wanton orchid]

I think my current vsoc is 1.015
and I'm on not smaller node (5000 series)
still blows my mind to push 1.35 to soc

[twin dew]

That was what all the various MB makers BIOSes set as default if you set the RAM to 6000MT/s.

But most didn't set that until you enabled higher RAM frequency manually.

[wanton orchid]

ddr4 OC V
on the soc lmao

[twin dew]

Up to 1.40V was supposed to be "safe".

[wanton orchid]

yea I know
but I still find it a bit stupid to go all the way up to 40% more tension

for supposed hardwiring expert manufacturers

sounds almost like tying to push 7V on the 5V line

[twin dew]

Also to the AGESA settings not being documented, the stupid thing is that the AGESA contains way to insert that documentation into it and Asus shows those direct in the BIOS when the setting is selected.
Usually just that what is written there is completely useless or just missing.

Some are very well documented, most are crap, some are just empty.

[wanton orchid]

power management utility [enable|disable]
detail :
Power Management Utility Enabling

yay

[night girder]

Look what just arrived.

[wary olive]

Mine works like a fucking beast

[pure karma]

does that mean im the odd one out now

[night girder]

I only bought it because of the current CPU maybe dying

else I would have sticked with my current CPU

[pure karma]

i daily 4 diffrent intel cpu's but i havent welcomed a amd cpu yet... i was considering it but ryzen 7000 on a 120mm rad sounds like a fairly awful idea

[dire igloo]

Why's it dying?

[pure karma]

ha finally hit the big 50

i guess its true that fixing a amd athlon a day keeps the intel doctor away

[valid eagle]

@twin dew finally made my choice and just ordered my pc. Will be : 14700K on a MSI MAG Z790 tomahawk max with 64go ddr 5 (6000) and finally a Nvme M2 pci 5.0 SSD 🙂
Thanks for the time you took, it actually helped me and gave me some things to dig into 😉

[night girder]

ask Baldur for the best answer 😉

[twin dew]

Max FLCK it can do stable seems to be going down with time.
Somewhere between 2000 and 1733MHz at this time.
And that matches what my 7800X3D did when VSOC damaged.
And that 7700X was run with too old BIOS for a moment, and ASRock has been setting some other settings to "OC" levels by default, so the original BIOS might have force-set VSOC to that 1.35V when MB was new.

The CPU is doing reboots of a type that are consistent with CPU firmware detecting uncorrected ECC error.
And isn't giving any MCEs for corrected errors, so not CPU caches.
And no ECC RAM, so not RAM.
So most likely IF errors.

[mental oriole]

Rookie numbers: #off-topic-tech message

[night girder]

Anyone know if this is normal for CPU fan Min Duty Cycle?

100 as a min for the CPU seems a bit… drastic?

[twin dew]

For PWM, that 20% original would have been completely normal lower limit.

For DC, 50-60% without checking what your fans can do, and it is setting to 60&.

[night girder]

It set CPU minumum from 20 to 100, that normal?

This means my CPU fans are always 100%

[twin dew]

No, like I said, the original 20% min would have been normal, not 100%.

[night girder]

Fuck, then something is wrong.

Oh btw. Remever our discussion about disabling fans?

[twin dew]

And that Chassis Fan 1 60 -> 31 would also be weird.

[twin dew]

Ok, so there was specific toggle for that in ASRock UEFIs.

[night girder]

This is my old asus

[night girder]

No clue what the CPU fan is doing. It’s a splitter into the CPU_FAN header on the mobo.

The chasis fans is normal. Minimum is lower but the curves look ok.

[night girder]

is it done often that people connect chasis fan straight to PSU?

[edgy hazel]

Sometimes, not often

If you ain't got a header or something

[night girder]

Yeah, there was header available, it was just not convenient to get to it.

Fan 1 was connected CHA_1. Fan 2 was connected to PSU. Fan 2 was unable to get to CHA_2 (too short cable). I connected fan 1 to CHA_2 and then Fan 2 to CHA_1.

and Q-fan is probably confused with my CPU fans because of PWM/DC. I ignored the 100% suggestion and put it on Turbo for now.

[visual tree]

Ok, there is definitely something bad going on with gmail atm...

About a month ago, gmail started labelling a lot of legit mails as spam

Including mails sent by google

Not to mention I've already received those mails before and never reported them as spam

[sand saddle]

Lmao, that's the only mail that's in my spam folder rn

But I have to agree, some of those google mails feel like spam

[visual tree]

I suspect google tinkered with the anti-spam filter and did something wrong

Perhaps made it too aggresive and now It thinks everything is spam lol

I had 24 hours to redirect my package from GLS but missed the deadline because I found out the email went in my spam folder and 24 hours have already passed by then....

Couldn't redirect the package to a packstation near my house and had to drive all the way to the depot. Thanks Google

[gilded helm]

https://www.youtube.com/watch?v=Wiw4fhFM7rQ

Interesting in retrospect. It feels like this is around the point where GPU pricing was fine again. RX 6600 for $230 ($185 now), 6700 XT for $360 ($330 now), and 6800 XT for $560 ($460 now).

[charred pewter]

im fine having spent $3k for my 5yo gpu ...

😉

[languid gulch]

if i ever got hardware like that, it'd either be in a giveaway or supplied by work 🤣

[languid gulch]

https://arstechnica.com/space/2024/02/it-turns-out-that-odysseus-landed-on-the-moon-without-any-altimetry-data/

from the "holy shit give that team a bonus"

[jagged snow]

Yeah, that's pretty wild

[tough owl]

is there an easy way to undervolt a 14600k on windows with a gui on a b series mobo?

[languid gulch]

300W PSU?

[tough owl]

no

[languid gulch]

200W psu

[tough owl]

friend is planning on doing a sff pc build with it

and cant exactly fit a peerless assassin in a small case

[languid gulch]

i feel like even a B series board should have that capability, but then again it's intel

i can see an H series board not having that

[tough owl]

well its in the bios

but xtu doesnt wanna play nice

[languid gulch]

looks like it's time for a hunt thru reviews

[gilded helm]

Why sff?

[languid gulch]

i want to go to intel for my next platform, but man the stuff that's randomly locked down with them vs amd can just be plain stupid

[twin dew]

Gmail and lot of others made DMARC mandatory or the mail will be labeled spam starting 2024-02-01.

As in the sender has to have valid DMARC DNS record for their mail.

Which had been warned about for over 6 months...

You can check any received mails headers to see why it was marked as spam.

SPF has been mandatory for over a year.

[tough owl]

he wants it small, idrk

[languid gulch]

is there a way for gpu manufacturers to test each core?

or do they just check clusters

[night girder]

I didn't know outlook added paid adds as mails in the mail list 😒

Just found out by helping someone.

[night girder]

www.bloomberg.com/news/articles/2024-02-27/apple-cancels-work-on-electric-car-shifts-team-to-generative-ai

Apple and their car plans, what a disaster. And just shifting teams from making a car to AI is also funny

[wanton orchid]

I mean if it's marketing team ..

[pure karma]

probably trying to make a portable pc

like a backpack pc or something

[pure karma]

i was wondering why i had a 20mb file of straight text and i think i got my answer

[mental oriole]

U-boot ethernet can go to hell...

[mental oriole]

Nice.

[tough owl]

its so cursed

[mental oriole]

Eyy don't give me flashback to uboot Ethernet hell

[cobalt ivy]

You could need a uptade :D‘

[pure karma]

oh boy..... i dont miss windows XP thats for sure

[river mountain]

INFO FOR EPIC USERS

https://vxtwitter.com/PC_Focus_/status/1762769993597411460?s=20

https://vxtwitter.com/PC_Focus_/status/1762822079991152933?s=20

[jagged snow]

oof

Alright, I at least don't have any payment methods added to my account

[twin dew]

So changed the password, might need to redo the 2-factor...

[jagged snow]

I don't have 2-factor and I think that password is unique

[twin dew]

Password was random unique 20+ char one.

And their pages are shit...
Logging all devices out sent email second auth code, and I didn't have email set up for that.
And removing app 2nd factor didn't ask for code...

[jagged snow]

By the way, I'm thinking of switching to bitwarden for easier cross-platform password management

Anyone have any experience with them?

[tough owl]

yea

I quite like their service

[jagged snow]

Alright, that's enough to convince me

[tough owl]

🤣

[jagged snow]

As far as I can tell, they haven't had a single security issue in the past

[tough owl]

what are you currently using

also @edgy hazel uses bitwarden too iirc

[jagged snow]

keepass / google's password manager

[edgy hazel]

Bitwarden best

[tough owl]

oh yea bitwarden way better than those

[edgy hazel]

[tough owl]

though if you want to generate otp passwords in the app or use hardware keys you'll need to pay $10 a year

[jagged snow]

they charge for 2fa 💀

[tough owl]

2fa server compute aint cheap 🤣

[jagged snow]

Yubikey doesn't require any compute on their end

[river mountain]

my old work did bitwarden for almost everything... except for one manager who wanted to try and switch us over to keeper

[tough owl]

its all stored on device

and the apps do the compute for it

its so good my school has an instance for staff 🤣

[jagged snow]

Why did they specifically single out Duo? That's the only 2fa app I use

[tough owl]

duo is a paid product no?

[jagged snow]

Nope

[tough owl]

for businesses it is

so much so that my school doesnt let just anyone use 2fa with duo, only staff get it

[jagged snow]

Oh wow
Premium is only 10/yr

I might just do that
I guess I'll wait and see, but I'd really like to use my yubikey

[tough owl]

I use both

[jagged snow]

For now, I just set it up as a standard 2fa key in duo instead of the more convenient one tap auth

[charred pewter]

only freebies for me, so they never even had payment info from me even on an old order 😉

[tough owl]

Privacy.com

[tough owl]

<@&387163995947270144>

thank you

[edgy hazel]

ordered a 3d printer

dwayne the duck johnson and the rocktopus will be soon mine

[tough owl]

what one did you buy

[edgy hazel]

creality ender 3 neo v2

[tough owl]

its only 230 in the states

[edgy hazel]

yup

[tough owl]

aint half bad

[edgy hazel]

enough for starting out. Will be used for DnD minis 😄

[visual tree]

Not sure how email from Google ended up in the spam folder though 😅

[night girder]

Funny, by coincidence I changed the epic account password of my niece today.

[charred pewter]

ive gotten some google notices about 'upcoming changes that can affect your business account' ... end up in spam

[fringe ivy]

I prefer Authy; I can use it on multiple devices.

[visual tree]

Is it from Google or a scam?

Mine was definitely from Google

[fringe ivy]

On Duo and certain others I had to remove the token to setup a new device.

[visual tree]

But Google declared itself spam 🤣

[night girder]

Did you check filters and blocked addresses in gmail? (web client itself)

because [e-mail] is one of the filters in gmail that I have, that e-mail is blocked and send to spam and I can't remember that I made that filter.

[visual tree]

Just checked, it's not on the list

I believe Google might have messed up the anti-spam filter somehow

[jagged snow]

Very good
I am one step closer to being de-googled

[night girder]

can you check the from header?

[tough owl]

I only keep gmail around for legacy accounts that use gmail

[jagged snow]

They no longer have my passwords or my browsing data, I'm working on moving away from google sso for everything as well

[visual tree]

Looks legit to me

[night girder]

From a browser, open Gmail.
Open the email you want to check the headers for.
Next to Reply , click More More and then Show original.
In a new window, the full header shows.
Click Copy to clipboard.

ooh, wait the bard ones. I think they also ended up in my spam filter.

Yeah, I think google forgot to whitelist their own [e-mail] for example. But if you are unsure, you can always check the official headers.

Hackers can send e-mail that looks they are from "[e-mail]" but when you check headers, you see it's from different source.

Just so you know it's possible (if you didn't already know ofc)

[night girder]

DMARC, SPF and DKIM all pass. Seems all legit.

Was curious thank you 😄

[visual tree]

That's what confuses me. Not only Google but a lot of other legit mails from other sources ended up in my spam folder recently

[night girder]

I know it's possible to spoof SPF/DKIM/DMARC, but then the chances of hackers getting that done with google, is almost non existing.

[visual tree]

Not to mention missing GLS package forwarding deadline because the email ended up in spam and I had to drive all the way to depot to collect the package

[night girder]

Yes, that's why I asked. Maybe the filter saw something we didn't and was actually trying to protect you.

Does anyone know if google has a whitelist principal?

If (e-mail is not in contacts of user && whitelist_feature == 1) -> (put it in spam)

I quickly looked around, but didn't find anything. Weird 🤔 And you are sure it's gmail doing it, not a client you are using?

[visual tree]

I am 100% sure it's gmail

[night girder]

support.google.com/a/answer/7663057?hl=en

Sometimes Gmail incorrectly marks messages coming from senders outside your domain as spam. As an administrator, try the steps in this article to fix the problem, and prevent it from happening in the future.

[visual tree]

Haven't touched gmail settings for years

[night girder]

Note: If Gmail identifies a message as potentially suspicious, it can be rejected or sent to spam, even if the sender is in your allowlist. 😂

[charred pewter]

yup heh

[night girder]

I hope they tested that spam filter really well 😄

[charred pewter]

nope heh

[visual tree]

Yeah, atm I am manually marking emails as "not spam" so they don't end up in the spam folder until the issue is fixed

[night girder]

Ooh.

I might have found the culprit.

Gmail uses machine learning to better understand your preferences, and to recognize spam. When you manually mark a message as not spam, Gmail learns how to manage messages addressed to you.

[tough owl]

YAY I LOVE GOOGLE READING MY EMAILS

[night girder]

You probably teached the ML to put those mails in spam. You gotta train it better 😉

maybe you once marked a @ gmail.com as spam, and the ML is like, ok blocking domain gmail.com 😂

[visual tree]

Haven't done it....yet.... 😅

[night girder]

For the person who is @ gmail, sorry for the ping. Shouldn't have named yourself like that. This was bound to happen 🤷‍♂️

My hunch asablic, the ML for you is confused and you gotta retrain it. Gotta love ML.

[visual tree]

Well, people sometimes tell me I am unpredictable 🤣

[night girder]

Even quantum computes with the most powerfull AI can't figure out which mails have to go to spam and which don't go to spam for you 🤣

[visual tree]

Which can be a good thing in some ways lol

[night girder]

This case is pretty interesting to see what happens when ML/AI doesn't do what we expect.

And while it sucks for you to miss an e-mail, it could be worse 🤷‍♂️

[visual tree]

I can ever trick those stupid personality tests from HR with ease and make them think I am a good candidate

[night girder]

I hate those tests 😒

[visual tree]

Once you do a lot of those tests, you will understand how useless they are and how you can trick the HR people

[night girder]

I know. It makes no sense.

And those tests looking for specific traits in people, what if I have a positive trait, that the test can't find? 😒

[visual tree]

For example, if you have a rating scale from 1 to 5, don't always pick 1 and 5

And avoid choosing 3 as much as possible

[night girder]

I mean, it's a good tool to help people decide to hire someone, but it's only guidance. Not a hard rule, imo. But some companies see it has a hard rule.

That the test is 100% correct and the results too.

[night girder]

So that leaves what? 2 and 4 and 0 ? 😄

[visual tree]

I don't see much difference between tossing a coin and personality test tbh

[night girder]

Oh, from 1 to 5. So 2 and 4. hehe

[visual tree]

There are only 2 and 4

[night girder]

my feelings exactly 😂

[night girder]

See? I would have failed already 😄

[visual tree]

If you always pick 1 and 5, everyone will know you tried to pick favourable answers

And be careful with questions that include "never, sometimes, always"

They can be also formulated in different ways but have the same meaning

[night girder]

But that's the thing, people are trying to figure out, what the test is trying to figure out, they are aware the test is testing them, so they try to figure out what the best answer is.

[visual tree]

And it's perfectly normal to answer you sometimes lie. If you say you never lied, that is a huge red flag

[night girder]

Super simple example from me: "Do you like harming other people?" ... Most will think, "If I write yes, they going to think I psychopath , so it's better to answer no". 😂

[visual tree]

or you sometimes do things that are not considered good

[night girder]

and I know questions are reformed to double check. But still, people aren't stupid.

Oh, and those tests that are ""not"" an IQ test but in reality are one, shove em up your behind.

[visual tree]

Yeah, you must always answer no unless you are being hired by mafia

[night girder]

Either don't lie to me, and call it an IQ test. Or don't do them. But don't rename it and think it's ethical.

[visual tree]

I hate questions like "Have you ever stolen something, even if it's something that doesn't have much value".

[night girder]

Yes, one piece of candy, the size of my nail, when I was a kid.

[visual tree]

If you say yes, HR people will think you are a criminal but completely forget almost everyone at least stole a pen in school

[night girder]

but in my defense, when I grew up, I went to the store, and I placed a single piece of candy right back. So it's all good 👍 🤣

[night girder]

Joking aside, I agree with you.

[jagged snow]

Man, some security features really just suck

[visual tree]

I stole an eraser in elementary school (yes, I regret doing that) and if you are going to judge me for the rest of my life because of that, then everyone is a hardcore criminal according to that logic

[night girder]

Oh, convenience vs security discussion? 😄

[jagged snow]

Meta won't let me change the phone number associated with my account for the next several days because I switched browsers

[jagged snow]

I'm fine with inconvenience, this is an actual security issue

I don't have that phone number anymore, so by not allowing me to change to my new number that's a pretty big issue

[night girder]

My first instinct: contact helpdesk. But then I realized. You might aswel just wait the few days, will be quicker than contacting meta and getting actual help.

[jagged snow]

That's what I figured too

[night girder]

what if you only use the old browser to change phone number?

[jagged snow]

So I guess I'll just wait

[night girder]

and then delete it again?

Or did they hard block all attempts (even the ones that should be leggit, but because security kicked it, is now also blocked) 😒

[jagged snow]

apparently their system is really borked

I just went to try it through chrome and now it's displaying that the change already went through

[night girder]

🥳

[jagged snow]

Alright
All my important accounts now have new unique 28 char passwords

That took way too long

I also went through and deleted about 50 old accounts I didn't use anymore

Little bit of digital footrpint sanitization done

Feels funny to be switching back to Mozilla

But they've come a very very long ways
Thunderbird is the best email client I've been able to find, and Firefox has enough features that chrome doesn't that it works very well for me

[visual tree]

I am planning on buying yubikey bio and using 1password to improve security

Until then, I am waiting for my paycheck 😅

[night girder]

theoretical it should be possible

[jagged snow]

I'm okay with still existing on the internet, that's not something that I'll be able to get away from considering that I'm active on social media

But I do think that I've done a fairly good job of making sure nobody can link this discord account to my identity

[night girder]

No, I wonder how much left over data (not exposed, but still on the server) remaims when we delete accounts.

That's what I meant 😄

[jagged snow]

In theory, it shoudl be none based on the laws in place

[night girder]

In theory, it should be deleted. But I know for fact it doesn't happen everywhere.

[jagged snow]

Most laws word it as "all associated information", but that rarely happens in practice

That doesn't bother me nearly as much as the fact that so many places require you to send an email in order for them to delete your account

That's like the gyms that require you to send a letter to cancel your membership

[night girder]

I care about it to be honest.

If I ask to delete my data, I want it deleted. Not exposing it to the internet/public != deleting.

But, most companies have backups etc. Start deleting single users from it. Too much hazzle they say.

[jagged snow]

I can see an argument for a company only deleting the current version, and then allowing it to be purged with the next backup

[night girder]

aslong as they don't get hacked should be fine

[edgy hazel]

they need to contact you and let you know that they can't remove your data from backups. Companies can balance it to avoid having to touch incremental backups and such.

[sharp oasis]

I say 2% chance this works and 99% chance my board fucking explodes

[jagged snow]

Get a fire extinguisher

And 2% is way too generous

[sharp oasis]

0.2 more like @jagged snow

I did manage 6400

1.9v

Is the kit dead

Surprisingly no

Is my is dead

Yeah

[jagged snow]

It's on its way

Remember that they can experience degradation just like the silicon in your cpu

[sharp oasis]

Oh ofc

Its probably degraded beyond doing 4000 anymorw

[jagged snow]

The new triplicate

[tough owl]

Still on windows 😔

[jagged snow]

I have ubuntu too

I'd probably run it full time but neither Adobe nor Autodesk supports it and I use both of their products fairly frequently

[jagged snow]

And to be fair windows kinda just works™️ for everything I do

Lots of work, and gaming