|
lonjil
|
2021-02-26 09:29:04
|
leaking sounds like a good idea
|
|
|
_wb_
|
2021-02-26 09:37:45
|
|
|
2021-02-26 09:38:14
|
What a wonderful passive-aggressive tone
|
|
|
Nova Aurora
|
2021-02-26 09:45:35
|
And standard PR stuff on making everything sound like everyone loves having to pay to know what the 'proper' way of doing things is
|
|
|
|
lonjil
|
2021-02-26 09:46:05
|
So like
|
|
2021-02-26 09:46:25
|
Is literally the only benefit that they handle copyright for you?
|
|
2021-02-26 09:46:49
|
I mean, 100% of the actual work here seems to have been done by people who seem to be on board with actually open standards.
|
|
|
190n
|
2021-02-26 09:49:07
|
yeah iso seems like a leech at least from this conversation. why are they involved with jxl at all?
|
|
|
|
Nova Aurora
|
2021-02-26 09:49:37
|
JPEG got the crew together?
|
|
|
|
lonjil
|
2021-02-26 09:49:56
|
Because the JPEG group asked people to submit proposals for a new JPEG standard.
|
|
|
|
190n
|
2021-02-26 09:50:05
|
ah okay
|
|
|
|
Nova Aurora
|
2021-02-26 09:51:29
|
Although the author's salaries are paid by cloudinary and google.
|
|
2021-02-26 09:51:51
|
Why does ISO get to charge for the copyright?
|
|
|
|
_wb_
|
2021-02-26 09:57:24
|
ISO does have staff, including editors who review specs to make sure things are done properly, e.g. words like "shall" and "must" are used in the correct way, references are done correctly, etc.
|
|
2021-02-26 09:58:23
|
And they do have a process that allows national bodies (ANSI, DIN, etc) to make comments and approve or not
|
|
2021-02-26 09:59:26
|
So that does mean that when approved, it's a standard that does have the weight of not just ISO but also those national bodies behind it.
|
|
2021-02-26 10:00:14
|
So I suppose that is an advantage of the ISO bureaucracy
|
|
2021-02-26 10:01:00
|
But the model of selling copies of specs is kind of silly if you ask me
|
|
2021-02-26 10:01:28
|
They also make money from the member fees they charge to anyone who wants to participate in standardization
|
|
2021-02-26 10:03:59
|
E.g. I have to pay 805 EUR per year to be an ISO member which is needed to be allowed to go to JPEG meetings (of course Cloudinary pays that for me, but still)
|
|
2021-02-26 10:04:55
|
They could pay their staff from such fees instead of charging for the specs
|
|
2021-02-26 10:05:35
|
Or just not be a private company but something that lives from government subsidies
|
|
|
bonnibel
|
2021-02-26 10:07:01
|
So what you're saying is that the real solution here is to abolish capitalism
|
|
2021-02-26 10:07:15
|
_the internationale starts softly playing in the background_
|
|
|
|
_wb_
|
2021-02-26 10:07:50
|
It is quite weird to me that ISO is a private company, based in Geneva, that just happens to have a monopoly position and gets privileged treatment by national and international governments in everything that has to do with standardization
|
|
|
bonnibel
So what you're saying is that the real solution here is to abolish capitalism
|
|
2021-02-26 10:09:04
|
that's usually the real solution, if you ask me 🙂
|
|
|
|
Nova Aurora
|
2021-02-26 10:11:42
|
https://tenor.com/view/elmo-dance-dance-party-soviet-gif-14812397
|
|
|
|
_wb_
|
2021-02-26 10:16:00
|
I am probably the only person in the world who is a member of both ISO and ISA
|
|
|
|
bobadingo
|
2021-02-26 10:16:20
|
Imagine how many security vulnerabilities has been introduced to any kind of software due to the lack of open specs. I'm surprised governments not treating this as a national security issue.
|
|
|
|
_wb_
|
2021-02-26 10:16:22
|
that's the International Standards Organization and the International Socialist Alternative 🙂
|
|
2021-02-26 10:18:26
|
hm, I think security vulnerabilities is more an issue of closed software than of closed specs, since security is mostly a matter of implementation. You could have an open spec with a proprietary implementation that is full of security holes, and a closed spec with an open source implementation that is well scrutinized (or an open source implementation that is also full of security holes, that can happen too of course)
|
|
2021-02-26 10:21:47
|
but of course open specs make alternative implementations easier
|
|
|
|
Deleted User
|
2021-02-26 10:37:03
|
JPEG XL was created from 2 outside-JPEG proposals (Pik and FUIF). How was original 1992 JPEG conceived? Did they have some in-house experts that created the standard? Or was it also an outside proposal?
|
|
|
|
_wb_
|
2021-02-26 10:49:25
|
Good question. Those were very different times. The JPEG committee was created in 1986, so it's older than its "parents" — ISO/IEC JTC 1 was created in 1987, SC 29 was created in 1991. I think they were more informal back then.
|
|
2021-02-26 10:53:35
|
The DCT dates back to 1972. Huffman coding dates back to 1952. So the basic concepts were already around. Arithmetic coding was newer (and patented).
|
|
|
|
Deleted User
|
2021-02-26 11:00:07
|
What's amazing is even that simple coding from 1992 was so powerful that it's still the backbone of the Web.
|
|
|
Pieter
|
2021-02-26 11:10:47
|
Wait until you hear when TCP/IP was invented.
|
|
|
|
Deleted User
|
2021-02-26 11:14:37
|
But JPEG could be made better because it's a lossy tool, while you can't improve TCP/IP *that* much because it's just for simple low-level comms
|
|
2021-02-26 11:15:03
|
HTTP could be made better though, see: HTTP/2.
|
|
|
|
Pieter
|
2021-02-26 11:15:23
|
At least the IP part got a major upgrade (and still is...).
|
|
|
|
lonjil
|
2021-02-26 11:15:56
|
something something http/3
|
|
|
|
Deleted User
|
2021-02-26 11:16:06
|
I doubt if we ever get rid of IPv4... <:monkaHmm:654081052108652544>
|
|
|
|
lonjil
|
2021-02-26 11:16:28
|
not as long as certain rich countries have more ipv4 addresses than they know what to do with...
|
|
2021-02-26 11:16:49
|
it's not a co-incidence that developing nations have waaaay more ipv6 adoption than western nations
|
|
|
|
Deleted User
|
|
lonjil
it's not a co-incidence that developing nations have waaaay more ipv6 adoption than western nations
|
|
2021-02-26 11:20:03
|
It's more probably because they got later into the game, so they don't have to support all of that nasty legacy stuff
|
|
|
|
lonjil
|
2021-02-26 11:20:30
|
no, it's because the ones that got in early has way more ipv4 addresses than they need
|
|
2021-02-26 11:20:51
|
when people talk about there being no more ipv4 addresses, it's about the global allocation
|
|
2021-02-26 11:21:10
|
like here in sweden we are only 10 million people but we were one of the first countries to get internet stuff.
|
|
2021-02-26 11:21:33
|
we have so many ipv4 addressses that most ISPs will just give you more addresses for free if you ask
|
|
2021-02-26 11:22:07
|
meanwhile africa totals a billion people or so and most of them still don't have internet, but the global allocation has run out.
|
|
|
|
Deleted User
|
2021-02-26 11:23:03
|
They (Internet adopters) probably overestimated things in the begging so they can't fix their past allocation mistakes... can they?
|
|
|
|
lonjil
|
2021-02-26 11:23:04
|
btw you can also see this in western countries where phones often have ipv6 addresses because the introduction of smartphones added way more stuff to the internet.
|
|
|
They (Internet adopters) probably overestimated things in the begging so they can't fix their past allocation mistakes... can they?
|
|
2021-02-26 11:23:22
|
the orgs that got those allocations in the past don't want to give them up
|
|
|
|
Deleted User
|
|
lonjil
the orgs that got those allocations in the past don't want to give them up
|
|
2021-02-26 11:23:30
|
WHAT
|
|
|
|
lonjil
|
2021-02-26 11:23:31
|
and in any case there aren't enough addresses anyway
|
|
|
|
Deleted User
|
2021-02-26 11:23:35
|
That's insane
|
|
|
|
lonjil
|
2021-02-26 11:24:14
|
Apple owns 1 out of every 255 IPv4 addresses.
|
|
2021-02-26 11:24:29
|
Ford another 1/255 chunk.
|
|
2021-02-26 11:24:38
|
And USPS.
|
|
2021-02-26 11:25:20
|
The US department of defence owns 13/255 addresses.
|
|
|
|
bonnibel
|
2021-02-26 11:25:53
|
also at least here isps also don't really have much incentive to improve
stuff like ipv6 and rpki is too obscure for most consumers to care about, esp when things appear to work just Fine without them
|
|
|
|
lonjil
|
2021-02-26 11:26:55
|
In some places you get a lot of bs like carrier grade NAT
|
|
|
|
bonnibel
|
2021-02-26 11:27:17
|
and unless you're somewhere with new glassfibre, there's only really 2 options. the company that got the old phone lines to run internet over, and the companies that resell that, and the company that got the old cable tv lines to run internet over (who doesn't have resellers)
|
|
2021-02-26 11:28:58
|
and if you live in an old building, the latter is pretty much always better ime
(sorry, went offtopic. my excuse is its past midnight and im annoyed with the isps here)
|
|
|
|
lonjil
|
2021-02-26 11:29:20
|
one of those 1/255 allocations (a /8) has 16.78 million addresses.
|
|
2021-02-26 11:29:31
|
So Apple has, well, that many addresses
|
|
|
|
_wb_
|
2021-02-26 11:36:09
|
one address for every 8-bit sRGB color
|
|
|
|
Deleted User
|
|
bonnibel
and if you live in an old building, the latter is pretty much always better ime
(sorry, went offtopic. my excuse is its past midnight and im annoyed with the isps here)
|
|
2021-02-26 11:44:04
|
> sorry, went offtopic
That's literally an off-topic channel, lol
|
|
2021-02-26 11:44:21
|
I don't think you can go off-topic on off-topic
|
|
|
|
bonnibel
|
2021-02-26 11:44:31
|
off^2-topic
|
|
|
|
Deleted User
|
2021-02-26 11:44:48
|
Off-topicception
|
|
|
|
bonnibel
|
2021-02-26 11:45:48
|
i'm training to set the world record for how many times i can get distracted and organically change the topic of conversation in 5 minutes
|
|
2021-02-26 11:47:04
|
"bonnie see a therapist you might have adhd or something" you cant make me and also i cant afford it
|
|
|
|
Deleted User
|
2021-02-26 11:47:08
|
ADHD?
|
|
|
bonnibel
"bonnie see a therapist you might have adhd or something" you cant make me and also i cant afford it
|
|
2021-02-26 11:47:20
|
WAIT I WAS ABOUT TO WRITE THIS
|
|
2021-02-26 11:49:02
|
And I also highly suspect ADHD in myself, I'm a diagnosed Aspie and lots of Aspies are also ADHDers, so I've got predisposition to ADHD
|
|
|
bonnibel
"bonnie see a therapist you might have adhd or something" you cant make me and also i cant afford it
|
|
2021-02-26 11:51:45
|
Same thing as you, no one will force me to see any goddamn therapist, psychologist or (the worst option) a psychiatrist *even if it was 100% free*. YIKES.
|
|
|
|
bonnibel
|
2021-02-26 11:52:46
|
i would actually like a good therapist, but for some reason a lot of them seem to be jerks
|
|
|
|
lonjil
|
2021-02-26 11:52:46
|
Last time I tried to get help for depression I got a "well shucks just be positive"
|
|
2021-02-26 11:53:02
|
And something like "do you really want to give up?"
|
|
|
|
Deleted User
|
2021-02-26 11:53:29
|
I'd also refuse to medicate myself, Aspies and autists would laugh at you if you told them you want to cure their ASD, while ADHDers obediently lick Big Pharma's boot...
|
|
|
|
lonjil
|
2021-02-26 11:53:44
|
I mean
|
|
|
|
bonnibel
|
2021-02-26 11:54:14
|
i am staying as far away as possible from anything inpatient though, because literally 100% of the 2-digit amount of people i know who have both voluntarily and involuntarily gone to inpatient stuff have been physically or sexually abused by staff
|
|
|
|
lonjil
|
2021-02-26 11:54:29
|
I know a lot of people who both have ADHD and are autistic, and they all laugh at anyone wanting to cure autism but also like ADHD meds.
|
|
|
|
Deleted User
|
2021-02-26 11:55:21
|
I hate them, you become not really physically addicted, but psychologically dependent on them
|
|
2021-02-26 11:55:32
|
I've got a great analogy. Have you ever played Subway Surfers?
|
|
|
|
lonjil
|
2021-02-26 11:55:44
|
No
|
|
2021-02-26 11:56:39
|
Most people want help to live their lives as they want to. This includes being who they are. They don't want a cure for ADHD it for autism but they want something to help them live as they wish to.
|
|
|
|
Deleted User
|
|
I've got a great analogy. Have you ever played Subway Surfers?
|
|
2021-02-26 11:58:34
|
There's my favorite hoverboard that has "Speed Up" (obvious one) and "Smooth drift" (you fall so slowly you can almost fly). That shit is addictive... I've been observing myself constantly using this hoverboard (or ones with similar capabilities) and hating when times out and go back to "normal" speed and gravity.
|
|
2021-02-26 11:59:35
|
I've been constantly using those hoverboards because I became dependent on them so much that I hate the game without them.
|
|
2021-02-27 12:01:00
|
I'm extremely afraid that the same thing would happen with me on meds: I'd so much hate non-medicated life that I'll take meds even on weekends and holidays. Like a point of no return...
|
|
|
|
bonnibel
|
2021-02-27 12:01:44
|
i mean i wouldnt call people getting meds that genuinely help them and staying on them for that reason an addiction. like, we don't frame people who wear glasses or contacts all day as addicted to lenses
|
|
|
|
Deleted User
|
2021-02-27 12:03:33
|
Instead of wearing lenses I'd make a surgery. And I don't think that you can compare those two, because meds change your way of thinking. *It's literally designed to affect your brain*.
|
|
|
|
lonjil
|
2021-02-27 12:04:24
|
I'm trying to get on a life long medication that alters the brain 😎
|
|
|
|
Deleted User
|
2021-02-27 12:04:25
|
And even if you decide to wear glasses, you get them once (or maybe a bit more than once for lens corrections). Meds? You have to buy them constantly
|
|
|
lonjil
I'm trying to get on a life long medication that alters the brain 😎
|
|
2021-02-27 12:04:40
|
I wouldn't be happy about it
|
|
2021-02-27 12:04:55
|
I don't want to become a lifelong pharma junkie!
|
|
|
|
lonjil
|
2021-02-27 12:05:05
|
As soon as someone figures out permanent estrogen surgery I'm down.
|
|
|
|
Deleted User
|
2021-02-27 12:05:12
|
Estrogen?
|
|
|
|
lonjil
|
2021-02-27 12:05:26
|
Female sex hormone
|
|
|
|
bonnibel
|
2021-02-27 12:05:45
|
🏳️⚧️✨
|
|
|
|
lonjil
|
2021-02-27 12:06:17
|
<:TransHeart:677836078564245505>
|
|
|
|
Deleted User
|
2021-02-27 12:06:54
|
Ah. That's why I was so confused after seeing `lonjil` GitHub profile...
|
|
|
|
lonjil
|
2021-02-27 12:07:27
|
Haha that's a photo from when I was like 12
|
|
|
|
Deleted User
|
|
lonjil
Haha that's a photo from when I was like 12
|
|
2021-02-27 12:08:34
|
Wow
|
|
2021-02-27 12:11:17
|
<@!167023260574154752> BTW who drew you avatar? Or is that from the interwebs?
|
|
|
|
lonjil
|
2021-02-27 12:11:37
|
Uhh it's from a picrew
|
|
|
|
bonnibel
|
2021-02-27 12:11:39
|
it's a picrew
|
|
|
|
Deleted User
|
2021-02-27 12:13:49
|
I recognise the Twitter one (they have their ubiquitous look), but Discord? I doubt it's from picrew too...
|
|
|
|
lonjil
|
2021-02-27 12:15:51
|
It's a different picrew
|
|
2021-02-27 12:16:00
|
Idr which one
|
|
|
|
Deleted User
|
2021-02-27 12:16:55
|
Ok, I don't recognize this picrew, the one from Twitter is almost everywhere lmao
|
|
|
|
lonjil
|
2021-02-27 12:18:38
|
<@456226577798135808> here you go https://picrew.me/image_maker/257476
|
|
|
|
Deleted User
|
2021-02-27 12:19:51
|
I don't want to use it, I was simply curious, but thanks 🙂
|
|
|
spider-mario
|
2021-02-27 09:19:24
|
having some nostalgia fun
|
|
2021-02-27 09:19:40
|
could we run JXL on this, I wonder 🤔
|
|
|
|
_wb_
|
2021-02-27 09:21:14
|
interesting challenge
|
|
2021-02-27 09:22:00
|
why not, I guess?
|
|
|
|
spider-mario
|
2021-02-27 09:25:00
|
is there even a C++11 compiler that can still target it?
|
|
2021-02-27 09:25:30
|
> Beginning with Visual C++ 2008, Visual C++ does not support targeting Windows 95, Windows 98, Windows ME, or Windows NT.
|
|
|
|
Pieter
|
2021-02-27 09:29:32
|
i wouldn't be surprised if mingw32-w64 can (which despite the name supports both 32-bit and 64-bit windows)
|
|
|
|
Deleted User
|
2021-02-27 09:47:41
|
Does Win98 have WIC or does it use different mechanism for adding another image codec?
|
|
|
|
_wb_
|
2021-02-27 09:56:33
|
Oh, that part I wouldn't bother with. I meant trying to get cjxl/djxl executables that run on old Windows
|
|
2021-02-27 09:56:41
|
Or on DOS even, maybe
|
|
|
|
spider-mario
|
2021-02-27 11:14:27
|
wow, this still works
|
|
2021-02-27 11:16:14
|
ah, this is from the time when web browsers still had a progress bar
|
|
2021-02-27 11:16:16
|
I miss this
|
|
2021-02-27 11:58:58
|
> The other point is that mingw-w64 does not support Windows OSes below XP officially. The mingw-w64 stuff might work well for Windows 2000, but this isn't guaranteed by us.
|
|
2021-02-27 11:59:01
|
> Use old versions of mingw.org, at least 2-3 years ago to be safe (gcc-3.4.5/gcc-4.2.1 era).
|
|
2021-02-27 11:59:02
|
:/
|
|
2021-02-27 12:00:02
|
no JXL for Windows 98, then, I guess
|
|
|
|
_wb_
|
2021-02-27 12:05:16
|
but for nuclear submarines running XP it should be fine
|
|
2021-02-27 12:05:16
|
https://www.popularmechanics.com/military/weapons/a19061/britains-doomsday-subs-run-windows-xp/
|
|
|
Crixis
|
|
spider-mario
ah, this is from the time when web browsers still had a progress bar
|
|
2021-02-27 12:49:13
|
Vivaldi has the progress bar
|
|
|
|
lonjil
|
2021-03-01 04:30:51
|
> mozjpeg/jpegtran.c: In function 'main':
> mozjpeg/jpegtran.c:772: error: unterminated #if
> 772 | #if JPEG_LIB_VERSION >= 80 || defined(MEM_SRCDST_SUPPORTED)
how the hell do you manage to screw up so badly that your #if doesn't terminate...
|
|
|
diskorduser
|
2021-03-01 05:45:22
|
Is there any offline tool or web site to compare images like imgsli?
|
|
|
|
Nova Aurora
|
2021-03-01 05:49:31
|
Squoosh works but that is only for encoding comparison
|
|
2021-03-01 05:50:44
|
slow.pics works, but it doesn't have the slide functionality
|
|
|
|
diskorduser
|
|
lonjil
<@456226577798135808> here you go https://picrew.me/image_maker/257476
|
|
2021-03-01 05:57:22
|
How do you remove the licensing warning ?
|
|
|
Fox Wizard
|
2021-03-01 02:35:39
|
Kwaliteit
|
|
|
|
lonjil
|
|
diskorduser
How do you remove the licensing warning ?
|
|
2021-03-01 06:11:50
|
the what
|
|
|
|
bonnibel
|
2021-03-01 06:36:38
|
you click on the brush icon, black is the 1st page of license, grey the 2nd, & white is the watermark
|
|
|
lonjil
the what
|
|
2021-03-01 06:38:56
|
they changed it so it starts off with an overlay telling people not to use it for commercial purposes, which you can switch out for a nontransparent watermark placed over the ear
|
|
2021-03-01 06:39:50
|
|
|
2021-03-01 06:39:58
|
|
|
|
|
_wb_
|
2021-03-01 07:19:13
|
|
|
2021-03-01 08:07:25
|
The watermark is too annoying to make usable avatars with this
|
|
|
|
lonjil
|
|
bonnibel
they changed it so it starts off with an overlay telling people not to use it for commercial purposes, which you can switch out for a nontransparent watermark placed over the ear
|
|
2021-03-01 09:19:37
|
Ah
|
|
|
|
Deleted User
|
2021-03-02 04:05:45
|
Always take a careful look at the licensing.. you’ll realize many works have been given a license that is irrevocable by the author… some may try to “”revoke”” it even though this is not possible unless the licensee has violated the license agreement
ex. CC, also technically this applies if the license holder says something too vague
|
|
|
Master Of Zen
|
2021-03-02 06:50:41
|
<@!794205442175402004> can you set yourself JPEG XL developer status? please
|
|
|
|
Nova Aurora
|
|
Master Of Zen
<@!794205442175402004> can you set yourself JPEG XL developer status? please
|
|
2021-03-02 06:53:10
|
How does it feel staring at the surface of the sun?
|
|
|
|
Master Of Zen
|
|
Nova Aurora
How does it feel staring at the surface of the sun?
|
|
2021-03-02 06:53:47
|
Well, to begin with, I don't sit in dark room
|
|
2021-03-02 06:54:05
|
|
|
|
|
Deleted User
|
|
Master Of Zen
<@!794205442175402004> can you set yourself JPEG XL developer status? please
|
|
2021-03-02 06:57:15
|
https://tenor.com/view/spongebob-squarepants-my-eyes-burn-burning-eyes-gif-4415974
|
|
|
|
Nova Aurora
|
|
Master Of Zen
<@!794205442175402004> can you set yourself JPEG XL developer status? please
|
|
2021-03-02 06:57:27
|
he did it
|
|
|
|
Master Of Zen
|
2021-03-02 06:58:52
|
YEY
|
|
2021-03-02 06:59:04
|
|
|
|
|
_wb_
|
2021-03-02 06:59:11
|
How do you get an icon there?
|
|
|
|
Master Of Zen
|
2021-03-02 06:59:12
|
https://tenor.com/view/dance-kid-club-gif-9152583
|
|
|
_wb_
How do you get an icon there?
|
|
2021-03-02 06:59:30
|
There
|
|
|
|
Nova Aurora
|
|
_wb_
How do you get an icon there?
|
|
2021-03-02 06:59:32
|
Using an emote
|
|
|
|
Master Of Zen
|
2021-03-02 06:59:54
|
Sadly, you can set custom emotes only if you have "turbo"
|
|
2021-03-02 07:00:04
|
by default it's only generic emojis
|
|
|
|
Nova Aurora
|
2021-03-02 07:00:26
|
like the bed in my status
|
|
|
|
_wb_
|
2021-03-02 07:07:03
|
I see
|
|
|
|
Master Of Zen
|
2021-03-02 07:14:41
|
Good choice)
|
|
|
|
Nova Aurora
|
2021-03-02 07:15:58
|
It's a weird emote, but it checks out
|
|
|
|
_wb_
|
|
|
Master Of Zen
|
2021-03-02 07:33:14
|
<@!794205442175402004> Roles a little bit wild)
|
|
|
|
Nova Aurora
|
|
Master Of Zen
<@!794205442175402004> Roles a little bit wild)
|
|
2021-03-02 07:35:01
|
Jon's still a noob? <:kekw:808717074305122316>
|
|
|
|
Fox Wizard
|
2021-03-02 08:17:25
|
Light theme
|
|
2021-03-02 08:17:28
|
<a:ban:714294878418501693>
|
|
|
|
_wb_
|
2021-03-02 08:21:55
|
I made the roles accumulate, otherwise the bot starts demoting people
|
|
2021-03-02 08:22:35
|
but yes, every expert is also still a noob. gotta stay humble! 😆
|
|
|
|
Master Of Zen
|
2021-03-02 03:38:49
|
<@!532010383041363969> oh, nice to see devs with avatars
|
|
2021-03-02 05:17:23
|
https://bellard.org/nncp/
|
|
2021-03-02 05:17:44
|
^ Lossless compression with NNCP'
|
|
|
|
_wb_
|
2021-03-02 05:19:00
|
ooo Bellard taking a shot at the Hutter prize now?
|
|
2021-03-02 05:19:21
|
this is something for <#794206087879852106> btw 🙂
|
|
2021-03-02 05:20:46
|
https://tenor.com/view/world-of-chaldea-chaldea-reorg-calla-impressed-gif-19463485
|
|
2021-03-02 05:20:52
|
https://tenor.com/view/kent-brockman-neural/-overlords-simpsons-reporter-welcome-gif-16530360
|
|
2021-03-02 05:21:42
|
`s/insect/neural compression/`
|
|
|
|
Master Of Zen
|
2021-03-02 07:20:10
|
https://www.thetimes.co.uk/article/deepfake-videos-of-tom-cruise-watched-by-millions-tr8lkmfdk?utm_medium=Social&utm_source=Twitter#Echobox=1614620214
|
|
2021-03-03 03:12:55
|
|
|
|
|
190n
|
2021-03-03 06:06:01
|
"if you didn't actually need $1200, you could have gambled it"
|
|
|
Dr. Taco
|
2021-03-03 06:32:12
|
"I gambled money and need to tell others that the risk I took was worth it to justify my self worth"
|
|
|
|
bonnibel
|
2021-03-03 07:06:23
|
You know, every time I buy a sandwich or pay my rent, I think "what this really needs is for this to take like 50 years of household energy"
|
|
2021-03-03 07:08:11
|
every time i look at my paycheck i think "dang, i wish the value of this fluctuated enormously"
|
|
|
|
Nova Aurora
|
2021-03-03 07:13:07
|
dammit, paid 10,000 dollars for a sandwich again
|
|
|
|
spider-mario
|
2021-03-03 07:17:54
|
I found this article interesting for anyone curious: https://web.archive.org/web/20210115174304/https://www.singlelunch.com/2020/10/21/badeconomics-putting-400m-of-bitcoin-on-your-company-balance-sheet/
|
|
|
|
bonnibel
|
2021-03-03 07:19:15
|
its not a real sandwich unless the transaction fee is in at least the double-digit range of dollars and the energy cost is at least 1 gwh
|
|
|
|
Nova Aurora
|
|
bonnibel
its not a real sandwich unless the transaction fee is in at least the double-digit range of dollars and the energy cost is at least 1 gwh
|
|
2021-03-03 07:20:21
|
Didn't steam drop bitcoin after the value of the average transaction was exceeded by the transaction fees?
|
|
|
Scope
|
2021-03-03 07:22:22
|
https://i.imgur.com/Wf0dBtA.png
https://foundation.app/probzzzz/dogsdogsdogsdogs-1920
|
|
|
|
bonnibel
|
|
Nova Aurora
Didn't steam drop bitcoin after the value of the average transaction was exceeded by the transaction fees?
|
|
2021-03-03 07:23:33
|
basically, in combination with the immense price instability
|
|
|
Scope
https://i.imgur.com/Wf0dBtA.png
https://foundation.app/probzzzz/dogsdogsdogsdogs-1920
|
|
2021-03-03 07:44:33
|
energy/co2 cost of that: http://cryptoart.wtf/#address=0x3b3ee1931dc30c1957379fac9aba94d1c48a5405
|
|
2021-03-03 07:45:07
|
|
|
2021-03-03 07:45:14
|
😬
|
|
2021-03-03 07:45:36
|
a fun price for the world's least effective watermark
|
|
|
BlueSwordM
|
2021-03-03 07:46:19
|
AV1 CPU-0: "We can encode a video the least energy efficient way possible."
Bitcoin: **Let's see about that**. <:kekw:808717074305122316>
|
|
|
|
Nova Aurora
|
2021-03-03 07:46:26
|
at one point I had a dollar of ethereum
|
|
2021-03-03 07:46:39
|
I wonder how much that is now
|
|
|
BlueSwordM
AV1 CPU-0: "We can encode a video the least energy efficient way possible."
Bitcoin: **Let's see about that**. <:kekw:808717074305122316>
|
|
2021-03-03 07:47:07
|
can't you use the blockchain for storing binary data?
|
|
|
|
BlueSwordM
|
2021-03-03 07:47:28
|
Yes, but you need to compress the data first to then store the raw values.
|
|
2021-03-03 07:47:47
|
Otherwise, it would be quite inefficient storing that much raw video. 😛
|
|
|
|
Nova Aurora
|
2021-03-03 07:48:16
|
And you'd waste millions of miner's storage space
|
|
|
|
BlueSwordM
|
2021-03-03 07:48:54
|
I'm actually surprised we don't have a cryptocurrency as a form of compression.
|
|
|
|
_wb_
|
2021-03-03 07:49:04
|
<@799692065771749416> are there any proposals for a cryptocurrency that is not based on wasting energy?
|
|
|
|
Pieter
|
2021-03-03 07:49:28
|
Plenty of dumb proposals.
|
|
2021-03-03 07:49:39
|
I think chia's proof of space and time is the most promising.
|
|
|
|
BlueSwordM
|
|
BlueSwordM
I'm actually surprised we don't have a cryptocurrency as a form of compression.
|
|
2021-03-03 07:50:03
|
Essentially, we'd have distributed encoding and compression as a form of cryptocurrency. 😎
|
|
|
|
Pieter
|
2021-03-03 07:50:13
|
Where they use "storage of random data for periods of time" as resource being expended, rather than energy.
|
|
2021-03-03 07:52:38
|
(and, of course, Bitcoin proponents would want me to point out that the energy spent on securing PoW isn't "wasted")
|
|
|
|
Nova Aurora
|
2021-03-03 07:52:48
|
mmmhmm
|
|
|
|
Scope
|
2021-03-03 07:53:03
|
https://en.wikipedia.org/wiki/Proof_of_stake
|
|
|
|
Pieter
|
2021-03-03 07:53:10
|
Proof of stake is a joke.
|
|
|
|
Nova Aurora
|
2021-03-03 07:53:14
|
bitcoin is one of the worst cryptocurrencies ever designed
|
|
|
|
Pieter
|
2021-03-03 07:53:22
|
It's circular reasoning all the way down.
|
|
|
|
Nova Aurora
|
2021-03-03 07:53:41
|
good as a proof of concept, bad for being a currency
|
|
|
|
Pieter
|
2021-03-03 07:53:42
|
Or rather, PoS as a consensus mechanism on itself is.
|
|
2021-03-03 07:54:04
|
I'd say it's the only (somewhat) functional cryptocurrency so far.
|
|
2021-03-03 07:54:28
|
Note that I'm most definitely biased about this.
|
|
|
|
Nova Aurora
|
2021-03-03 07:55:38
|
and the lightning network is complete nonsense to me
|
|
|
|
bonnibel
|
|
Pieter
(and, of course, Bitcoin proponents would want me to point out that the energy spent on securing PoW isn't "wasted")
|
|
2021-03-03 07:55:42
|
~~i mean it's literally using more energy than most countries in exchange for... a pyramid scheme you can use to buy a vpn and drugs~~
|
|
|
|
Pieter
|
2021-03-03 07:55:43
|
It works great.
|
|
|
|
Nova Aurora
|
|
Pieter
It works great.
|
|
2021-03-03 07:58:35
|
The lightning network? Last time I looked at it it looked like a hack to make a broken system work slightly more efficently
|
|
|
|
Pieter
|
2021-03-03 07:59:11
|
If that's your opinion I don't think we can have a useful discussion.
|
|
|
|
Nova Aurora
|
2021-03-03 08:01:07
|
maybe you'd be better at describing it
|
|
|
|
Pieter
|
2021-03-03 08:01:36
|
LN makes a lot of sense to me. A global consensus system is just fundamentally more appropriate as a fair, but slow and expensive, settlement system. And payment channel based systems that work between participants directly are more appropriate as payment systems, but needs a settlement system as a backend.
|
|
2021-03-03 08:02:21
|
LN has lots of interesting challenges still, so I don't want to present it as a silver bullet. But at least in practice, it works very well.
|
|
2021-03-03 08:02:56
|
In our office we had a candy dispenser that was LN-operated; it just worked, instantly, and cost basically nothing (after setting up channels, of course).
|
|
|
|
_wb_
|
2021-03-03 08:13:08
|
I wish there was a way to do a PoW-based settlement syst where the work itself is also useful in itself - now the work is basically just "something that is hard to do and easy to prove, by design", but it is not something that is also useful (from an external pov).
|
|
2021-03-03 08:14:25
|
Storing random data doesn't seem useful, but storing arbitrary (but not random) data is useful.
|
|
2021-03-03 08:15:32
|
Being a backup of wikipedia/wikimedia, or something like that
|
|
|
|
Pieter
|
2021-03-03 08:15:39
|
I think generally that both very hard (because useful work is generally useful to someone, and if there is a party deciding what is useful, you lose the decentralized aspect) and messes with the incentives.
|
|
2021-03-03 08:16:03
|
Because if someone would be willing to pay for it, you don't get the property anymore that mining needs to bet on the right chain, or risk losing income.
|
|
2021-03-03 08:16:36
|
If the "value" of the work is negligible compared to normal mining income that's not a problem of course, but then it's arguably not a solution to the problem you're looking for.
|
|
2021-03-03 08:18:10
|
Another issue is that you need work that commits to the chosen chain - you can't have something that's usable for multiple chains simultaneously, or you lose that property too.
|
|
2021-03-03 08:18:38
|
In PoW, it's hashes of the tip block's header. In proof of space & time, it's output of a stream cipher seeded by the block data.
|
|
|
|
_wb_
|
2021-03-03 08:22:12
|
Maybe it needs to be based on an entirely different concept, not mining and chains but something else completely, that has maybe different kinds of incentive mechanisms to make it work
|
|
|
|
Pieter
|
2021-03-03 08:23:08
|
Yes, that's my belief too. If something wants to do "better", it'll be a completely different concept, not a small tweak of bitcoin.
|
|
2021-03-03 08:24:12
|
And by a different concept, I mean not even a different solution to the same problem, but likely a very different problem statement.
|
|
2021-03-03 08:24:42
|
Bitcoin also does not solve the same problem as early e-cash system did, except at a very high abstraction level.
|
|
2021-03-03 08:25:44
|
e-cash was trying to solve the problem of having a single centralized server that could censor anything, but due to anonimity properties of the cryptography, couldn't tell which transactions belonged to whom.
|
|
2021-03-03 08:26:36
|
Turns out, it's still pretty easy to just get the server shut down.
|
|
2021-03-03 08:27:34
|
But if you try to compare the two, and from an e-cash mindset look at Bitcoin, you'd probably protest that it's not even an actual consensus system, as it never guarantees finality of any given update.
|
|
|
|
_wb_
|
2021-03-03 08:28:55
|
It would be great if "mining" would boil down to "becoming part of the infrastructure that makes the internet work better, more decentralized, and more robust, including but not limited to securing transactions"
|
|
|
|
Pieter
|
2021-03-03 08:29:37
|
I agree, but I'm very skeptical that something like that is possible without a wildly different concept.
|
|
|
|
_wb_
|
2021-03-03 08:31:00
|
Yes, I agree a wildly different approach is needed, and it would have to be at least as innovative as Bitcoin was
|
|
2021-03-03 08:31:45
|
And I agree it's probably not even about finding a better solution, but more about finding a better problem
|
|
2021-03-03 08:32:43
|
Which is harder, because it takes an engineer to solve a problem but a philosopher to make a problem
|
|
|
|
Pieter
|
2021-03-03 08:33:13
|
An explanation I like about Bitcoin is that of seeing it as a court system: there is this infrastructure - the court, or blockchain - that anyone can present their evidence to (including the contract they had agreed to with others) - and the system will always rule fairly. It's expensive, and slow, but thankfully, not everyone needs to go to court. In fact, the mere existence of the court helps keeping participants honest.
|
|
|
|
_wb_
|
2021-03-03 08:34:13
|
That's not a bad analogy
|
|
|
|
Pieter
|
2021-03-03 08:34:37
|
But obviously, it can only rule on things that are entirely native to it.
|
|
2021-03-03 08:35:33
|
To make it observe things in the real world, you need what are called "oracles", which sign based on real world events - but they always involve trust.
|
|
2021-03-03 08:37:01
|
Things like lightning can be seen as building a payment system that adds efficiency under the assumption of honest behavior. As long as everyone cooperates, you can transact, cheaply, and quickly. When someone is irresponsive, or tries to cheat, you go to the chain to adjudicate and get the money you were owed (plus a penalty, in the case of LN, but there are alternative payment channel systems that work differently).
|
|
|
|
_wb_
|
2021-03-03 08:42:56
|
What is the proportion in reward for bitcoin miners between finding new bitcoins and transaction fees?
|
|
2021-03-03 08:43:30
|
Are there statistics on that?
|
|
2021-03-03 08:44:07
|
On how that ratio evolved over the years?
|
|
|
|
Pieter
|
2021-03-03 08:44:27
|
It's a small but non-negligible percentage now.
|
|
2021-03-03 08:45:00
|
I'd expect that there are dashboards showing it now.
|
|
2021-03-03 08:45:43
|
Let's see if I can find one. I expect that what you'll see is that there are periods where it is essentially 0, and periods where it's suddenly non-negligible.
|
|
2021-03-03 08:46:27
|
Because of limited capacity, there are times when demand exceeds the capacity, and space in blocks essentially becomes a market where space goes to the highest bidder.
|
|
2021-03-03 08:46:38
|
And there are periods where that's not the case, and fees are negligible.
|
|
2021-03-03 08:47:17
|
As the subsidy drops, and interest increases, I expect that a steady state will be that there is always a non-negligible amount of fees. Anything else means there is also a big problem for security going forward.
|
|
2021-03-03 08:47:54
|
https://transactionfee.info/charts/block-coinbase-amount/
|
|
|
|
Scope
|
2021-03-03 08:49:03
|
Also, any decentralized systems often tend to be centralized, such as Cloudflare CDN or mass DNS binding to Google and Cloudflare servers
|
|
|
|
Pieter
|
2021-03-03 08:49:38
|
DNS is inherently centralized to begin with 😉
|
|
|
|
Nova Aurora
|
2021-03-03 08:50:00
|
ICANN
|
|
|
|
Pieter
|
2021-03-03 08:50:23
|
Also, Cloudflare is quite possibly a front for a US intelligence agency.
|
|
2021-03-03 08:50:44
|
(I don't actually believe that it is, but it also wouldn't surprise me)
|
|
|
|
Nova Aurora
|
|
Pieter
Also, Cloudflare is quite possibly a front for a US intelligence agency.
|
|
2021-03-03 08:50:51
|
*the internet as a whole currently
|
|
|
|
Pieter
|
2021-03-03 08:51:00
|
Fair.
|
|
2021-03-03 08:51:32
|
<@794205442175402004> Even better: https://bitinfocharts.com/comparison/fee_to_reward-btc.html
|
|
|
|
Nova Aurora
|
2021-03-03 08:51:41
|
except in china I guess but then you go from possibly an Orwellian nightmare to actually an Orwellian nightmare
|
|
|
|
Scope
|
2021-03-03 08:53:01
|
Yes, well at least the availability of sites was not so dependent on one or two companies (not so long ago Cloudflare went down and many popular sites also did not work as well as DNS for people who switched them to 1.1.1.1)
|
|
|
|
Pieter
|
2021-03-03 08:56:17
|
<@111445179587624960> Also, is it a stretch to say that today email = gmail?
|
|
|
|
Nova Aurora
|
2021-03-03 08:56:44
|
I'd say that's only 50% true currently
|
|
|
|
Scope
|
2021-03-03 08:57:08
|
Yes, and also many other things, this is just an example
|
|
|
|
Nova Aurora
|
2021-03-03 08:57:09
|
enough to be scary, but not total domination yet
|
|
|
|
Pieter
|
2021-03-03 08:57:20
|
I think gmail is so dominant that google can probably push for whatever email standard they want adopted, as long as they're not too reckless with it.
|
|
|
|
_wb_
|
2021-03-03 08:57:27
|
Broadcasting all transactions does not seem to be the best idea for scalability, except I guess if you can have something that allows local islands of transactions to happen for a while, only syncing a summary to the main chain every once in a while, which is what I guess is roughly what LN is doing?
|
|
|
|
Pieter
|
2021-03-03 08:58:17
|
<@794205442175402004> Yeah, pretty much. In general you only broadcast the setup and teardown transactions (and there is work on getting those too batched with other things, ...).
|
|
|
|
_wb_
|
2021-03-03 08:58:31
|
Outlook and stuff is still kind of big, it doesn't work very well but lots of companies and governments are still locked firmly into the Microsoft ecosystem
|
|
|
|
bonnibel
|
2021-03-03 08:59:08
|
Tbf there's not too much exciting happening in the email standard world afaik
There was JMAP (an IMAP + SMTP for user clients (not for server-to-server) replacement), but that was Fastmail, and no-one else's really adopted it yet
|
|
|
|
Pieter
|
2021-03-03 09:00:12
|
But fundamentally you need _something_ broadcast, because there is one type of validity you cannot decide locally: two equally valid but conflicting transactions spending the same coins. You need _something_ to make that decision (because basic physics, in a decentralized system you cannot absolutely decide which is first and which is second).
|
|
|
|
bonnibel
|
2021-03-03 09:00:26
|
I guess Google helped push MTA-STS?
|
|
|
|
Pieter
|
2021-03-03 09:01:17
|
So if that something isn't going to be a central party, it'll need to be broadcast to all parties that potentially are going to be involved (PoW is effectively probabilistically picking a winner to make that decision every time, with incentives to make sure subsequent winners cooperate with them).
|
|
2021-03-03 09:05:12
|
If you have fixed identities of a group of verifiers, you can use a traditional consensus system (paxos, PBFT, ...) to decide. But that preassumes a group of trusted parties rather than a single trusted party, and so that's not really a solution either if you want a fully permissionless system.
|
|
2021-03-03 09:06:20
|
<@!178524721023811584> I used to run my own mail server a long time ago, but then moved to gmail (and worked at google...), and now it does seem a lot harder (DKIM etc) to run your own.
|
|
|
|
_wb_
|
2021-03-03 09:06:55
|
Couldn't the decentralized-simultaneous-spending problem be solved in some other way? I dunno, maybe some kind multi-step protocol to do a transaction, where there's announcement of intent to do transaction, then some verification which ends in a green light, then confirmation or abort of transaction, or something
|
|
|
|
Pieter
|
2021-03-03 09:07:42
|
I think that's just moving the problem elsewhere.
|
|
2021-03-03 09:07:55
|
Multi-step protocols are hard if you have no identities.
|
|
2021-03-03 09:08:42
|
There are some hybrid approaches where you use PoW to decide on a finite set of trusted verifiers for a finite period of time, and then use traditional consensus among those.
|
|
2021-03-03 09:11:19
|
_-> work_
|
|
|
|
bonnibel
|
|
_wb_
Couldn't the decentralized-simultaneous-spending problem be solved in some other way? I dunno, maybe some kind multi-step protocol to do a transaction, where there's announcement of intent to do transaction, then some verification which ends in a green light, then confirmation or abort of transaction, or something
|
|
2021-03-03 09:13:14
|
It's not _that_ hard, and there's a lot of existing solutions for it. Most MTAs have a plugin/filter for DKIM, and the current most popular spam filtering system, Rspamd, can do both DKIM & ARC signing
|
|
2021-03-03 09:14:21
|
I'm currently using OpenSMTPd + Rspamd + Dovecot
(a similar setup you'd get from <https://poolp.org/posts/2019-09-14/setting-up-a-mail-server-with-opensmtpd-dovecot-and-rspamd/>, though that one does a little more filtering in opensmtpd itself and like, actually _trains_ rspamd with what you put into/out of the spam folder, which I haven't bothered with)
|
|
|
|
Pieter
|
2021-03-03 09:14:26
|
Yeah, I'm sure it's doable. Just pointing out there is more to it now than 15y ago.
|
|
|
|
bonnibel
|
2021-03-03 09:14:31
|
That is true
|
|
2021-03-03 09:16:20
|
I'd love to write a minimal incoming mail server that runs in SGX and just encrypts everything coming in asap, but sadly 1. I don't really have the time, 2. I have an AMD cpu, & 3. SGX gets broken like once a quarter
|
|
|
|
Pieter
|
2021-03-03 09:16:43
|
Ah, SGX. aka proof-of-Intel.
|
|
2021-03-03 09:17:55
|
It's a nice idea in theory, but all the speculative execution problems are sufficient to make it hard to trust.
|
|
|
|
bonnibel
|
2021-03-03 09:19:46
|
(TL;DR for my motivation for it: I'm not using email for e2ee communication. Everyone I know who might bother to set up PGP or S/MIME I already have on like, Signal, which leaves mostly newsletters, package tracking links, and mail from elderly relatives. Therefore, just encrypting everything coming in as well as possible is enough for my use case. However, the email server sees every mail in plaintext, so how do you make sure it's not just making a copy before asymmetrically encrypting it and sending it on to me? That's (theoretically) where something like SGX comes in)
|
|
|
|
Deleted User
|
|
Pieter
<@111445179587624960> Also, is it a stretch to say that today email = gmail?
|
|
2021-03-03 09:26:22
|
Not for me, I do have Gmail but use Outlook on a daily basis, using Gmail only when there's some incentive to do it
|
|
|
|
veluca
|
2021-03-03 09:37:07
|
maybe a silly question, but couldn't one build a blockchain where the proof-of-work is based on some variant of solving folding@home-like tasks? one could pick some known, useful np-hard problem (like ILP), and people would publish real problems to be solved that get altered in some way to still produce an optimal solution to the original problem but also depend on some arbitrary data
|
|
2021-03-03 09:38:26
|
it wouldn't reduce energy consumption of the blockchain itself, but it would make it do useful work
|
|
2021-03-03 09:38:53
|
(I'm sure there is a million ways to cheat this specific system, but it could be an interesting idea...)
|
|
|
|
Pieter
|
2021-03-03 09:41:39
|
<@179701849576833024> I think may people have tried that, but nothing comes really close. There was primecoin at some point that used something vaguely related to prime numbers, but they're not really useful either.
|
|
|
|
veluca
|
2021-03-03 09:42:37
|
well, ILP would be useful 😛 ask the people that did gurobi say...
|
|
|
|
Pieter
|
2021-03-03 09:42:51
|
Also gridcoin, which uses BOINC... but of course this kind of leads to a trivial centralization: the BOINC server could hand out problems that some already know the answer to.
|
|
2021-03-03 09:44:48
|
Another idea I've heard once: build a normal PoW function that includes some computation that's very related to a useful problem, in the hope that the ASICs that would be developed for it help commoditizing solutions to the related problem.
|
|
|
|
bonnibel
|
2021-03-03 09:45:34
|
~~my favoured solution is to regulate bitcoin et al into the ground~~
|
|
|
|
Pieter
|
2021-03-03 09:46:42
|
It's still hard though, because PoW generally wants maximal bandwidth with barely any latency concern, and often with extremely high fault tolerance (e.g. if you can build something that is 20% more power efficient, but just computes the wrong thing in 10% of the cases, that's totally worth it).
|
|
|
|
veluca
|
2021-03-03 09:46:55
|
what if you make a gridcoin-equivalent that uses a "distributed BOINC", based on gridcoin itself? I'm not exactly sure what would be the good way to deal with it
|
|
|
|
Pieter
|
2021-03-03 09:46:57
|
Which is completely the opposite for what you want for most "real" computation.
|
|
|
|
diskorduser
|
2021-03-04 05:38:52
|
<@!258670228819410944> Could you tell me the name of window decoration/theme in this image?
|
|
|
|
Master Of Zen
|
|
diskorduser
<@!258670228819410944> Could you tell me the name of window decoration/theme in this image?
|
|
2021-03-04 05:44:29
|
https://github.com/kupiqu/SierraBreezeEnhanced
|
|
2021-03-04 05:44:46
|
|
|
|
|
Crixis
|
|
Jim
|
2021-03-04 12:50:26
|
😆
|
|
2021-03-04 12:50:52
|
Numbers go down before they go up
|
|
|
Jyrki Alakuijala
|
2021-03-04 08:35:06
|
hahaha -- back then they didn't even know which year it was
|
|
2021-03-04 08:36:05
|
it is very difficult to anticipate someone's birth 7999 years ahead of time and that it has such a profound impact on how years are labeled
|
|
|
|
Dr. Taco
|
2021-03-04 08:40:31
|
nope mate, your wrong
```js
console.log(new Date(0));
January 1st 1970
```
A Thursday, apparently
|
|
|
|
190n
|
2021-03-04 08:43:16
|
time didn't exist before 1970
|
|
|
|
Pieter
|
2021-03-04 08:44:01
|
```js
> console.log(new Date(-1));
1969-12-31T23:59:59.999Z
```
|
|
|
|
190n
|
|
|
Pieter
|
2021-03-04 08:56:26
|
```js
> console.log(new Date(-62167219200000));
0000-01-01T00:00:00.000Z
> console.log(new Date(-62167219200001));
-000001-12-31T23:59:59.999Z
```
|
|
|
|
Crixis
|
|
|
bonnibel
|
2021-03-04 09:47:39
|
https://everestpipkin.medium.com/but-the-environmental-issues-with-cryptoart-1128ef72e6a3
|
|
2021-03-04 09:49:34
|
<https://twitter.com/vectorpoem/status/1367512454469545984>
|
|
|
|
Deleted User
|
2021-03-04 10:02:20
|
What's "cryptoart"?
|
|
|
|
Pieter
|
2021-03-04 10:02:34
|
You don't want to know.
|
|
|
|
Jyrki Alakuijala
|
2021-03-04 10:03:49
|
My #つぶやきGLSL tryout: bit.ly/3qdqMF7
|
|
|
|
Deleted User
|
|
Pieter
You don't want to know.
|
|
2021-03-04 10:07:16
|
Believe me, I know things no one should be supposed to know.
|
|
|
|
Master Of Zen
|
|
Pieter
You don't want to know.
|
|
2021-03-05 12:08:45
|
I thought it's was something scary
|
|
|
bonnibel
https://everestpipkin.medium.com/but-the-environmental-issues-with-cryptoart-1128ef72e6a3
|
|
2021-03-05 12:41:08
|
Damn, you got me
|
|
|
|
BlueSwordM
|
2021-03-05 02:20:40
|
https://cdn.discordapp.com/attachments/803196804651941959/817218903213342740/stopdoingmemes.png
|
|
|
|
Pieter
|
2021-03-05 03:56:27
|
hahaha
|
|
|
|
Master Of Zen
|
2021-03-06 10:11:12
|
<@!794205442175402004> your JPEG XL developer status cleaned up, as it is by default cleans in one day, and need to be manually set to never clear
|
|
|
|
_wb_
|
2021-03-06 10:25:18
|
ugh, discord has annoying defaults — it's the same with invite links
|
|
|
fab
|
2021-03-06 12:34:54
|
|
|
2021-03-06 12:35:00
|
master of zen
|
|
2021-03-06 12:35:23
|
how to remember butteraugli
|
|
2021-03-06 12:38:51
|
but the pronounciation maybe is gh l
|
|
2021-03-06 12:43:17
|
is it a good idea to upscale images with jpeg xl 0.3.0?
|
|
2021-03-06 12:43:19
|
for %i in (C:\Users\User\Documents\F*.jpg) do cjxl "%i" "%i.jxl" -j -P 5 --palette=0 -I 5 -s 3 -m --mquality=75 --num_threads=2
|
|
2021-03-06 12:44:03
|
which quality i should set to do that
|
|
2021-03-06 12:44:26
|
should i use better tools as oyo netflix
|
|
|
|
_wb_
|
2021-03-06 01:33:44
|
Upscale images? How do you mean?
|
|
|
|
Deleted User
|
2021-03-06 01:51:17
|
hey cloudinary
you upscale anime female say x2, x4, png
after use cjxl "%i" "%i.jxl" -j -X 9 --hdr 0 **--resample=2 or =4**
for butteraugli quality!
|
|
2021-03-06 01:52:08
|
can you check?
|
|
|
|
_wb_
|
2021-03-06 01:56:03
|
Are you imitating Fabian now? 🤔
|
|
|
|
Deleted User
|
2021-03-06 01:56:57
|
I was answering your question <:ugly:805106754668068868>
|
|
|
|
_wb_
|
2021-03-06 01:57:34
|
Could be nice actually to have a cjxl option that doesn't first downscale but just encodes the given image like it is, and signals an upsampling factor
|
|
2021-03-06 01:58:33
|
So you give it a 500x400 image, and it encodes that, but when you decode it, it will be a 1000x800 (or 4x, or 8x) image, with upsampling
|
|
|
|
Deleted User
|
|
_wb_
So you give it a 500x400 image, and it encodes that, but when you decode it, it will be a 1000x800 (or 4x, or 8x) image, with upsampling
|
|
2021-03-06 02:00:40
|
What about 5x, 6x, 7x?
|
|
|
|
_wb_
|
2021-03-06 02:01:32
|
https://c.tenor.com/g8_v4mDf_nsAAAAM/inconceivable-princessbride.gif
|
|
2021-03-06 02:03:56
|
Only 2x, 4x, 8x can be done this way. Other factors (including non-aspect-ratio-preserving scaling) can be signalled via the IntrinsicDimensions header, but then you don't get control over if and how the application will do the rescale
|
|
2021-03-06 02:04:59
|
But we could e.g. do NN upscaling too (with custom upscaling weights), which could be nice for pixel art.
|
|
|
|
Deleted User
|
2021-03-06 02:07:04
|
You could, but do you also want to implement that? :D
|
|
|
|
fab
|
2021-03-06 02:11:43
|
0.3.0 lossy modular has interesting algorithm
|
|
2021-03-06 02:11:48
|
wonder how much it discards
|
|
2021-03-06 02:11:52
|
if is too much
|
|
2021-03-06 02:12:36
|
and if you can re create 0.3.0 and add upscaling in the opposite way
|
|
2021-03-06 02:13:58
|
but i think you can't recreate 0.3.0
|
|
2021-03-06 02:14:01
|
encoder
|
|
2021-03-06 02:14:20
|
trick in the decoder jpeg xl has them
|
|
2021-03-06 02:14:33
|
but you can't re create 0.3.0 lossy modular encoder
|
|
2021-03-06 02:14:52
|
i don't know if it looks good
|
|
2021-03-06 02:15:31
|
i know enhancements has never been a priority of the reference encoder
|
|
2021-03-06 02:15:41
|
of any codec
|
|
2021-03-06 02:17:09
|
so talking seriously
|
|
2021-03-06 02:17:18
|
how to up resolution with modular
|
|
2021-03-06 02:17:19
|
for %i in (C:\Users\User\Documents\F*.jpg) do cjxl "%i" "%i.jxl" -j -P 5 --palette=0 -I 5 -s 3 -m --mquality=75 --num_threads=2
|
|
2021-03-06 02:17:27
|
this is the setting i used with 0.3.0
|
|
2021-03-06 02:17:54
|
can the same encoder heuristic do double resolution?
|
|
2021-03-06 02:20:17
|
could that also work for video?
|
|
2021-03-06 02:20:25
|
or is better to use topaz denoise
|
|
2021-03-06 02:20:37
|
like to extract all the frames and do similar thing
|
|
|
|
_wb_
|
2021-03-06 02:20:43
|
There currently only is a cjxl option (--resample) that will _downscale_ the image before encoding, and the decoder will upscale again with a nice upscaling
|
|
|
|
fab
|
2021-03-06 02:20:43
|
with 0.3.0 lossy modular
|
|
2021-03-06 02:21:17
|
do you remember the 0.3.0 lossy modular heuristics
|
|
2021-03-06 02:21:35
|
is fine with the setting i used
|
|
|
|
_wb_
|
2021-03-06 02:21:38
|
I don't think anything changed since 0.3.0 in lossy modular
|
|
|
|
fab
|
2021-03-06 02:21:43
|
i didn't tested 0.3.3
|
|
2021-03-06 02:21:58
|
but it should be only for low fidelity
|
|
|
|
_wb_
|
2021-03-06 02:22:44
|
I expect vardct to be better at lower and lower fidelity
|
|
|
|
Dr. Taco
|
2021-03-06 05:27:06
|
https://cdn.discordapp.com/attachments/690673268927234160/817155476772749322/2YQX5pQE8nrnf_lG.mp4
|
|
|
|
veluca
|
|
fab
but the pronounciation maybe is gh l
|
|
2021-03-06 06:05:45
|
I think it's pronounced how italians would read butteraug-li, "gli" doesn't really seem to exist in other languages 😛
|
|
|
|
fab
|
2021-03-06 06:07:37
|
Sorry dont understand
|
|
2021-03-06 06:08:00
|
Maybe is ghli
|
|
|
|
Deleted User
|
2021-03-07 07:12:20
|
<@!228116142185512960> wow, you're on Keybase?
|
|
|
surma
|
2021-03-07 11:41:18
|
<@456226577798135808> Is that note-worthy?
|
|
|
|
Deleted User
|
2021-03-07 11:56:53
|
<@!228116142185512960> I don't know a lot of people that use it
|
|
|
|
surma
|
2021-03-07 12:29:59
|
<@456226577798135808> “use” might be an overstatement. I definitely did not give them my private key lol. And I like that I can prove that a bunch of accounts and domains belong to me. But honestly, I should probably consider switching to Linktree or something
|
|
|
|
Deleted User
|
2021-03-07 01:19:11
|
<@!228116142185512960>
> I definitely did not give them my private key lol
Neither did I. No one will get it, even in encrypted form.
> I should probably consider switching to Linktree or something
I don't think that it should be a great idea...
> I can prove that a bunch of accounts and domains belong to me.
...because you can make that proof using encryption. I've been there since the service was invite-only and not owned by Zoom. You're knowledgeable enough to use cryptography properly, so with Linktree you'll loose security and some kind of "cool nerdiness" factor.
|
|
|
|
fab
|
2021-03-07 02:21:16
|
I wrote an article about jpeg xl
|
|
2021-03-07 02:24:12
|
it's the right text channel to post?
|
|
|
|
_wb_
|
2021-03-07 02:31:18
|
No, this is <#794206087879852106> 😅
|
|
|
|
fab
|
2021-03-07 02:47:07
|
when i going to see the article on SEO
|
|
2021-03-07 04:53:00
|
added tag JPEG XL with wordpress altervista
|
|
2021-03-07 04:53:04
|
i didn't with yoast seo
|
|
2021-03-07 04:53:10
|
should i do with yoast seo
|
|
2021-03-07 04:53:14
|
or it's paid?
|
|
2021-03-07 04:53:20
|
altervista in italy is from mondadori
|
|
2021-03-07 04:53:33
|
also updated the article
|
|
2021-03-07 05:40:38
|
The article is wrong
|
|
2021-03-07 05:40:41
|
i don't understand
|
|
2021-03-07 05:40:50
|
why when bluesword makes tutorials all gone right
|
|
2021-03-07 05:41:16
|
when i try myself with altervista wordpress ita it gets ruined all the formattation even new heuristics was formatted bad
|
|
2021-03-07 05:41:28
|
i should re write the article
|
|
2021-03-07 05:42:53
|
|
|
2021-03-07 05:43:08
|
maybe is the " who are different
|
|
2021-03-07 05:43:17
|
also the --
|
|
|
|
BlueSwordM
|
2021-03-07 05:43:19
|
Well Fabian, I have an idea.
|
|
|
|
fab
|
2021-03-07 05:43:23
|
what i did wrong
|
|
2021-03-07 05:43:39
|
ok
|
|
|
|
BlueSwordM
|
2021-03-07 05:43:53
|
Maybe I should help you with writing the article?
|
|
|
|
fab
|
2021-03-07 05:44:21
|
first you see the differene between these two commands
|
|
2021-03-07 05:44:26
|
changing the "
|
|
2021-03-07 05:44:31
|
breaks the command?
|
|
2021-03-07 05:45:03
|
do you see any other difference?
|
|
|
|
BlueSwordM
|
2021-03-07 05:45:38
|
No, it seems fine. Here is the one I use for WebP on Windows
> ``` @ECHO OFF
>
> :: Create the directory in which you want to put the encoded files
> md WebP
>
> :: Command to actually compress the files using your encoder of choices
> for %%a in ("*.png") do (
> cwebp -z 9 -mt -v "%%a" -o "WebP\%%~na.webp"
> )
>
> done ```
|
|
|
|
fab
|
2021-03-07 05:46:09
|
|
|
2021-03-07 05:46:15
|
see the difference
|
|
2021-03-07 05:46:21
|
first is the original that works
|
|
2021-03-07 05:46:36
|
second is how it was edited by altervista
|
|
2021-03-07 05:46:42
|
and it doesn't work at all
|
|
2021-03-07 05:51:46
|
so do you know how to solve this problem on wordpress
|
|
2021-03-07 05:51:50
|
the --
|
|
2021-03-07 05:51:55
|
and the ""
|
|
|
|
BlueSwordM
|
2021-03-07 05:52:30
|
I do not know actually why it destroyed the formatting.
|
|
|
|
fab
|
2021-03-07 05:52:42
|
maybe i should embed a png
|
|
2021-03-07 05:54:36
|
that makes the page way more bad
|
|
2021-03-07 05:55:04
|
and not sure if the dev will get angry if i talk about evaluation codecs with bad formattation
|
|
2021-03-07 05:59:44
|
fixd
|
|
|
|
spider-mario
|
2021-03-07 06:21:23
|
<@!416586441058025472> I don’t really understand what you mean by this in the blog post:
> Don’t set less or more than q 94, it has no sense and going higher you can have artifacts plus some moirè.
|
|
2021-03-07 06:21:46
|
do you mean that higher qualities somehow give worse quality than q94? that should not happen, I think
|
|
|
|
fab
|
2021-03-07 06:33:29
|
for jpg with that setting
|
|
2021-03-07 06:33:36
|
p5 lossy palette
|
|
2021-03-07 06:33:43
|
that exact build
|
|
2021-03-07 06:33:59
|
also altervista is a joke not really a blog
|
|
2021-03-07 06:35:11
|
basically i recommended to re encode again a jpg png in q 94 if you're unsure
|
|
2021-03-07 06:35:21
|
and you think source quality is not good enough
|
|
2021-03-07 06:35:35
|
i don't think i'm telling a bad advice
|
|
2021-03-07 06:35:45
|
they will compress the photo
|
|
2021-03-07 06:35:51
|
but is not my responsability
|
|
2021-03-07 06:36:18
|
it's literally evaluation build nobody will use anyway
|
|
2021-03-07 06:37:40
|
and i also specified at the start the problem that that are, security issues etc
|
|
2021-03-07 06:37:58
|
bluesword can make his image comparisons for him
|
|
2021-03-07 06:38:10
|
i wrote it for fun
|
|
2021-03-07 06:41:01
|
spidermario are you an italian developer
|
|
|
|
spider-mario
|
2021-03-07 06:41:13
|
I am not, but <@179701849576833024> is
|
|
2021-03-07 06:41:18
|
I’m French
|
|
|
|
fab
|
2021-03-07 06:41:33
|
onlyveluca
|
|
|
|
veluca
|
2021-03-07 06:42:28
|
In JPEG-XL yes - unless someone didn't tell me something 🤔
|
|
|
|
lonjil
|
2021-03-08 03:58:34
|
PSA: you just lost the game.
|
|
|
|
Nova Aurora
|
|
|
Deleted User
|
|
lonjil
PSA: you just lost the game.
|
|
2021-03-08 05:35:04
|
DAMN IT <:Hsss:806131225278152756>
|
|
|
|
BlueSwordM
|
2021-03-08 05:35:41
|
<:Hsss:806131225278152756>
|
|
|
|
Crixis
|
2021-03-08 07:16:06
|
<:AngryCry:805396146322145301>
|
|
|
|
BlueSwordM
|
2021-03-08 06:54:31
|
https://cdn.discordapp.com/attachments/803196804651941959/818556423062552676/Screenshot_20200706-004725.png
|
|
|
|
_wb_
|
2021-03-08 07:03:01
|
Is this an AVIF vs JXL meme?
|
|
|
|
Nova Aurora
|
2021-03-08 07:04:01
|
It could be construed that way
|
|
|
|
spider-mario
|
2021-03-08 07:26:46
|
did y’all know that Thriller, before being Thriller, was: https://youtu.be/QsF-csdTOr4
|
|
|
|
_wb_
|
2021-03-08 07:44:13
|
https://c.tenor.com/-Dsa9P26eyIAAAAM/dana-carvey-johnny-carson.gif
|
|
|
|
fab
|
|
|
Fox Wizard
|
|
|
fab
|
2021-03-09 12:03:00
|
\\\\\\\\\\\\\\\\
|
|
|
|
Master Of Zen
|
|
Fox Wizard
Hm
|
|
2021-03-09 12:12:52
|
Directions unclear, Gmaps link pls
|
|
|
|
Fox Wizard
|
2021-03-09 12:13:28
|
Lol
|
|
2021-03-09 12:13:45
|
Probably only Dutch people will get it though :p
|
|
|
|
Master Of Zen
|
|
Fox Wizard
Probably only Dutch people will get it though :p
|
|
2021-03-09 12:16:55
|
Why Dutch? France on pic
|
|
|
|
Fox Wizard
|
|
|
_wb_
|
2021-03-09 12:56:27
|
it says "Napoleon's treasure map" in Dutch at the bottom
|
|
|
|
Fox Wizard
|
2021-03-09 01:21:18
|
That and they were part of many people's youth :p
|
|
|
|
bonnibel
|
2021-03-09 08:44:51
|
https://www.phoronix.com/scan.php?page=news_item&px=git-clone-CVE-2021-21300
|
|
2021-03-09 08:45:02
|
tldr code exec vuln in git clone, pls update
|
|
|
|
BlueSwordM
|
|
bonnibel
https://www.phoronix.com/scan.php?page=news_item&px=git-clone-CVE-2021-21300
|
|
2021-03-09 08:45:27
|
Oh, so that's why I just received a git update.
|
|
|
|
190n
|
2021-03-09 08:48:37
|
> on case-insensitive file-systems
|
|
|
|
Diamondragon
|
2021-03-09 10:55:59
|
Off topic, but does anybody here have experience with squoosh.app or ImageGlass?
|
|
|
|
veluca
|
2021-03-09 11:08:37
|
some (limited) experience with squoosh.app, but mostly because I helped them to add jxl 😛
|
|
|
|
Diamondragon
|
2021-03-09 11:19:11
|
I'm getting a cryptic error with a jxl that I created using a jpeg (lossless recompression.)
|
|
2021-03-09 11:19:22
|
|
|
2021-03-09 11:19:29
|
|
|
2021-03-09 11:20:22
|
The latest ImageGlass beta won't display it either, and throws up an error. I made it using default settings on 0.3.3 libJPEGXL.
|
|
2021-03-09 11:21:31
|
Not sure If I screwed something up somewhere, or if maybe the jpeg is defective.
|
|
|
|
veluca
|
2021-03-09 11:41:15
|
that seems to be a common issue with lossless JPEG recompression...
|
|
2021-03-09 11:41:25
|
there might be something wrong in our code
|
|
|
|
Diamondragon
|
2021-03-09 11:54:09
|
I see. Thanks for the clarification.
|
|
|
|
diskorduser
|
|
Diamondragon
|
|
2021-03-10 08:29:06
|
nice anime art
|
|
|
|
lvandeve
|
2021-03-10 10:23:48
|
About the squoosh failure with the lossless recompressed image: Does squoosh support the JXL container format? Its MIME detection only checks for plain JXL codestream, not for the JXL container box format header:
https://github.com/GoogleChromeLabs/squoosh/blob/dev/src/client/lazy-app/util/index.ts#L153
|
|
|
|
veluca
|
2021-03-10 10:37:57
|
https://github.com/GoogleChromeLabs/squoosh/pull/969
|
|
2021-03-10 10:38:00
|
this should fix it
|
|
|
|
lvandeve
|
2021-03-10 10:42:17
|
thanks 🙂
|
|
|
|
Scope
|
2021-03-10 01:41:25
|
<https://github.com/GoogleChromeLabs/squoosh/commit/0874a3ba5295873608ad8493f6e7a8f72db4efef>
> Update JPEG XL and support container format in mime sniffing. (#969)
|
|
|
|
Diamondragon
|
2021-03-10 06:01:38
|
I refreshed squoosh and it works. Thanks guys!
|
|
|
|
Jim
|
|
veluca
this should fix it
|
|
2021-03-10 06:03:45
|
Confirmed, it is working now. Thank you!
|
|
|
improver
|
2021-03-10 06:51:46
|
so basically it's jxl if it starts with `\xff\x0a` or `\x00\x00\x00\x0cJXL \x0d\x0a\x87\x0a`? or there could be some other variations?
|
|
2021-03-10 07:01:54
|
<https://gitlab.com/wg1/jpeg-xl/-/blob/master/plugins/gdk-pixbuf/pixbufloader-jxl.c#L270> this is uhh.. a bit different?
|
|