HomeSearchPrevious dayNext day

00:00:50dm4v quits [Read error: Connection reset by peer]
00:01:10mutantmonkey quits [Remote host closed the connection]
00:01:26mutantmonkey (mutantmonkey) joins
00:01:41dm4v joins
00:01:43dm4v quits [Changing host]
00:01:43dm4v (dm4v) joins
00:02:34jamesatjaminit quits [Ping timeout: 250 seconds]
00:44:47jamesatjaminit (jamesatjaminit) joins
00:47:57BlueMaxima joins
01:03:15dm4v quits [Ping timeout: 244 seconds]
01:03:37dm4v joins
01:03:39dm4v quits [Changing host]
01:03:39dm4v (dm4v) joins
01:07:17<kiska>What a fun time https://www.reddit.com/r/CoronavirusDownunder/comments/p6gsms/nsw_633_new_local_and_1_overseas_cases_3_deaths/
01:07:26<kiska>NSW going to be in lockdown for like the next 3 months
01:09:38<@JAA>Meanwhile, NZ had the first case since February and went into lockdown.
01:10:08<nyany>kiska: oof
01:10:50<kiska>Well... this is what you get with leaders that want business to remain open and trading
01:10:59<kiska>*cough*Conservative*cough*
01:18:35<nyany>stay safe kiska
01:18:43<nyany>the world is crazy
01:18:48<kiska>Its not like I am going out :D
01:23:31<@JAA>Oh yeah, NSW is about comparable in population to CH. We had over 3k new cases yesterday, but no measures are going to happen anytime soon.
01:41:20nertzy (nertzy) joins
01:43:06jamesatjaminit quits [Ping timeout: 250 seconds]
01:49:00<Frogging101>I think everyone wants business and trade to be open
01:49:10<Frogging101>it's sort of necessary
01:56:23nertzy quits [Client Quit]
01:59:21jamesatjaminit (jamesatjaminit) joins
02:05:12jamesatjaminit quits [Ping timeout: 250 seconds]
02:20:57<Doranwen>what I think is crazy is that people think you can actually stop covid from spreading, like have it die out - not when you have animal reservoirs! it's impossible to eradicate a virus that has animal reservoirs like that, and all the lockdowns in the world won't work, nor even vaccines, since they've admitted you can still get it and spread it and viral loads are the same whether you've been vaccinated or not
02:21:07<Doranwen>it's going to be around just like colds and flu
02:21:31<Doranwen>so we're better off learning how to live with it than trying to shut down the world in hopes of getting rid of it
02:29:39jamesatjaminit (jamesatjaminit) joins
02:53:37<@OrIdow6>Apropos of "spying agencies" and AT (#nevermind), it may be that some of us get on completely automated "lists", but other than that I suspect it's very useful to them
02:54:16<@OrIdow6>A while back there was some internal NSA document that got out through FOIA, sort of promoting use of the WBM
02:56:08<@OrIdow6>I became aware of all of this basically by doing intense web browsing - I presume that if you do research of that kind as a job, you're going to know what resources are available
02:58:04jamesatjaminit quits [Ping timeout: 250 seconds]
02:59:15<@OrIdow6>I do suspect that at some point a startup is going to find out that the data is public, call it a "dataset", and start using it for blackmailing people or whatever else is hip these days
02:59:25<@OrIdow6>Like Pushshift
03:01:48qw3rty_ joins
03:02:14jamesatjaminit (jamesatjaminit) joins
03:05:11qw3rty__ quits [Ping timeout: 244 seconds]
03:21:06<Jake>( https://upload.wikimedia.org/wikipedia/commons/c/c1/Untangling_the_Web.pdf page 275 for reference)
03:59:49qw3rty__ joins
04:03:30qw3rty_ quits [Ping timeout: 250 seconds]
05:11:46tzt quits [Ping timeout: 244 seconds]
05:24:08tzt joins
05:26:13tzt quits [Changing host]
05:26:13tzt (tzt) joins
05:57:57BlueMaxima_ joins
06:01:28BlueMaxima quits [Ping timeout: 252 seconds]
07:11:07driib quits [Ping timeout: 244 seconds]
08:00:09britmob256 quits [Quit: britmob256]
08:04:20HackMii quits [Remote host closed the connection]
08:04:41HackMii (hacktheplanet) joins
08:19:12Earendil (Cobalt17) joins
08:24:22rsn_ joins
08:26:32rsn quits [Ping timeout: 250 seconds]
08:31:59Core2562 (Cobalt17) joins
08:32:57Core5044 (Cobalt17) joins
08:32:57Core2562 quits [Read error: Connection reset by peer]
08:33:47Earendil quits [Ping timeout: 244 seconds]
09:01:17BlueMaxima_ quits [Client Quit]
09:45:36Core5044 quits [Ping timeout: 244 seconds]
11:04:38qwertyasdfuiopghjkl joins
11:08:22HackMii quits [Ping timeout: 258 seconds]
11:09:09HackMii (hacktheplanet) joins
11:50:37nertzy (nertzy) joins
12:04:16britmob256 joins
12:07:04nertzy quits [Client Quit]
12:10:58sec^nd quits [Remote host closed the connection]
12:11:23sec^nd (second) joins
12:29:40<@JAA>That didn't take long... Second preimage attack on Apple's NeuralHash algorithm (which is used for the CSAM detection stuff that made headlines recently): https://github.com/AsuharietYgvar/AppleNeuralHash2ONNX/issues/1
12:34:25qwertyasdfuiopghjkl quits [Remote host closed the connection]
13:01:48rsn joins
13:02:58rsn_ quits [Ping timeout: 244 seconds]
13:43:18qwertyasdfuiopghjkl joins
14:23:03qwertyasdfuiopghjkl quits [Ping timeout: 244 seconds]
14:24:02qwertyasdfuiopghjkl joins
14:40:31<Frogging101>JAA: wait, what? It's more than just a regular file hash?
14:42:54<Frogging101>"The main purpose of the hash is to ensure that identical and visually similar images result in the same hash, and images that are different from one another result in different hashes. For example, an image that has been slightly cropped or resized should be considered identical to its original and have the same hash"
14:42:58<Frogging101>...I see
14:43:10<Frogging101>This sounds ripe for problems and false positives
14:43:25Arcorann quits [Ping timeout: 252 seconds]
14:47:15<spirit>nah, that's actually a smart approach
14:47:23<spirit>else you can just flip a bit
14:47:33<Frogging101>well yes except then we have issues like what JAA linked
14:47:34<spirit>and images are all about perception
14:47:48<Frogging101>I'm not saying it's a dumb idea, I'm saying I'm skeptical about their ability to actually do it
14:48:17<Frogging101>Without screwing it up
14:48:28<spirit>there are things in life that work well when used as intended
14:48:49<spirit>and there are things that can be abused maliciously
14:48:59<Frogging101>I hope someone breaks this thing, collides it with a meme, and renders the system useless. It must be stopped
14:49:16<spirit>the intention of the algorithm is to find matches to known imagery
14:49:20<spirit>and for that it works well
14:49:33<@JAA>Except it doesn't as illustrated in that link.
14:49:52<Iki>Internal FBI facial recognition documents indicate the agency has historically accepted error rates up to 20%: https://www.nbcnews.com/news/us-news/facial-recognition-databases-hearing-congress-hammers-fbi-n737461
14:50:11<Iki>"Internal FBI documents obtained in a Freedom of Information Act lawsuit by the nonprofit Electronic Privacy Information Center indicate that the FBI's own database, called the Next Generation Identification Interstate Photo System, or NGI-IPS, had an acceptable margin of error of 20 percent — that is, a 1-in-5 chance of "recognizing" the wrong person."
14:50:20<spirit>the link illustrates an approach to maliciously generate data that will flag images
14:50:32<spirit>a completely different thing
14:50:48<spirit>(dont get me wrong, apple and all that are driving us to 1984)
14:51:20<spirit>facial recognition is something completely different
14:55:06<spirit>perceptual hashing is a fantastic approach to quickly search for known imagery in a set of images
14:55:20<spirit>running it on your own data without your consent or any legal reason is insane
14:55:35<@JAA>Well, true. We don't really know yet how well it works for detecting perceptually identical images. The comments on the script for the preimage attack suggest it shouldn't be too difficult to circumvent that.
14:55:47<spirit>this attack makes it possible for someone to inject data that will trigger a hashing match into your data
14:56:05<@JAA>Yeah
14:56:09<spirit>and then bad things can happen if no human looks at the actual data before you get lynched or something
14:56:14<spirit>that's all
14:56:48<spirit>i am not sure that hyping this is the right answer to the general issue of Apple's insanity, i think it is at least one level to technical already
14:57:27<Frogging101>Even if you've got humans looking at flagged cases, you've already got a crazy privacy breach
14:59:22<spirit>yeah! but not because of the algorithm but because of the general access to your device
14:59:45<@JAA>It does however potentially provide a way to actively combat the bullshit. If it can be turned into a first preimage attack, you could easily generate countless matching images and spam their system with noise.
15:00:46<@JAA>But yeah, I agree, it shouldn't exist in the first place.
15:01:04<spirit>"oh, we have to improve our algorithm" -> "develop better algorithm" -> "access to device gets normalized" :\
15:07:18<Frogging101>A world where there is a perfect CSAM detecting algorithm might not be a bad one. But I don't trust them to stop there
15:07:48<Frogging101>Also it's a moot point because a perfect CSAM detecting algorithm is probably impossible
15:08:01<Frogging101>so any such algorithm in the real world will be flawed and therefore dangerous
15:09:53<spirit>It's not the algorithm that is dangerous, it's the people
15:10:24<Frogging101>well, yeah. guns don't kill people, etc.
15:10:38<Frogging101>the statement that it's dangerous is in the context of it being used by people
15:11:51<spirit>"We have access to the devices for preventing $evil1 already, can we use it for $evil2? And $evil3? Oh, and it would be nice to also use it for making sure we surveill the $frenchpeople properly"
15:20:49pabs quits [Ping timeout: 252 seconds]
15:21:21pabs (pabs) joins
15:34:19yarrow quits [Quit: Connection closed for inactivity]
16:19:01Matthww80 joins
16:20:46Matthww8 quits [Ping timeout: 252 seconds]
16:20:46Matthww80 is now known as Matthww8
16:25:43wizards quits [Ping timeout: 252 seconds]
16:27:25wizards joins
16:45:22c00k13_ quits [Quit: bye~]
16:46:09c00k13 joins
16:50:07heart joins
16:56:42mrfooooo quits [Quit: Ping timeout (120 seconds)]
16:57:05mrfooooo joins
17:45:26Larsenv quits [Quit: ZNC 1.8.2+deb1+focal2 - https://znc.in]
17:47:14luckcolors quits [Quit: No Ping reply in 180 seconds.]
17:48:16Larsenv (Larsenv) joins
17:48:32luckcolors (luckcolors) joins
18:07:36wizards_ joins
18:10:40wizards quits [Ping timeout: 250 seconds]
18:31:03Viniter69 quits [Ping timeout: 244 seconds]
19:02:23Viniter69 (Viniter) joins
19:10:06Ruthalas5 (Ruthalas) joins
19:12:12Ruthalas quits [Ping timeout: 250 seconds]
19:12:12Ruthalas5 is now known as Ruthalas
19:16:00qwertyasdfuiopghjkl quits [Ping timeout: 244 seconds]
19:40:17jamesp (jamesp) joins
19:43:28<jamesp>I wonder why you guys cancelled the Twitter Archive. I believe archiving Twitter is at least nearly as important as reddit. That's just my opinion.
20:13:33qwertyasdfuiopghjkl joins
20:18:43<heart>I don't know the real reason why the Twitter archive was cancelled, but in my opinion Twitter would be a massive, MASSIVE effort. Better to wait until it's shutting down and they freeze the site.
20:19:47<heart>~350 million active accounts, half a billion Tweets sent per day...yeah, nobody can reasonably keep up with that. Even the Library of Congress gave up on it.
20:35:00<@JAA>That project idea was from when Twitter announced they'd purge inactive accounts, i.e. large-scale content deletion. They halted those plans after Jason pointed out that they don't have a way to memorialise accounts of deceased people. Or rather, they delayed the account purge until such a way would be added. I haven't heard anything about it since. I imagine we'll do something when that is announced
20:35:06<@JAA>again.
21:31:25<Frogging101>They halted their plans because of something Jason said?
21:31:56<Frogging101>or just in temporal proximity to what he said
21:37:17<@JAA>Well, Jason tweeted about it, people started complaining, and they halted the plans.
21:52:50<nyany>something something social media doesn't like it when you try to archive their shit
21:53:02<nyany>which is sad but also understandable
21:53:08<nyany>anyways brb about to go brrrrrr
21:55:47<@JAA>Well no, they were angry at Twitter deleting dead loved ones' content. But obviously that has something to do with who follows Jason in the first place.
22:07:13Viniter69 quits [Client Quit]
22:08:41Viniter69 (Viniter) joins
22:59:23<jamesp>We could take a vote of should we continue it or not. I'll vote to continue.
23:11:25<jamesp>The names you pick for the projects are funny IMO.
23:16:18<h3ndr1k>I think it's not really the project names, it's mostly the channel names. On the tracker and on GitHub the projects have normal names.
23:21:50<jamesp>Well I like the IRC channel names. Anyway, can't wait to revive #twitterdead
23:25:03<@JAA>As I said, if Twitter ever changes their stance and reannounces the deletion, we'll look into it I guess.
23:25:13<@JAA>Until then, highly unlikely.
23:26:35<jamesp>Why don't we use the channel for ANY accounts we deem at risk? More like manually chosen accounts, like #youtubearchive
23:26:59<@JAA>Because that typically happens in #archivebot via socialbot.
23:28:11wizards_ is now known as wizards
23:28:16<jamesp>JAA I would like you to check out #stripminer
23:28:35<@JAA>I don't think I'm interested in that.
23:35:50Arcorann (Arcorann) joins
23:54:39@OrIdow6 quits [Ping timeout: 258 seconds]

HomeSearchPrevious dayNext day