« first day (3781 days earlier)      last day (1063 days later) » 

8:00 PM
I thought the Hello World feed would also help us find new answers by new users that may not be done well enough (although there's a review queue for that, I guess)
 
@hyper-neutrino I care enough about not caring that I downvoted the question
 
At least be sincere about your apathy
 
@pxeger here's a little lesson in trickery
Question: I have an idea for a question - print all the named oxygen compounds, as found on Wikipedia. Yea or nay?
 
(sandbox it :p) do you have a list
 
8:06 PM
Sounds like a boring KC to me.
 
^ What does it offer over existing kolmogorov complexity challenges? I don't see how they're easily compressible or something
 
@hyper-neutrino no
 
in general the answer to "should I make this KC challenge" is probably "no" and occasionally if there's some interesting compressability to it then maybe
 
I couldn't even write to files until yesterday at 23:45
@hyper-neutrino the internet is allowed for this one
 
8:07 PM
What changed then?
 
it means you don't have to hardcode the string
(s)
 
most KC challenges either fall into a) english and thus can be compressed by languages with dictionary compression, b) a lot of repeated tokens and thus can be compressed with a simple substitution approach which each language does pretty much the same way, or c) have no compression and therefore bubblegum just wins
 
@pxeger Personally, I'd be more than happy to just start it and stop if we don't want it. But, given that there has been pushback to it, I've been trying to get a better gauge of whether people would want it or not, as I don't want to impose my own opinions on the chatroom
 
@StackMeter well, you'll have to provide a list, in which case you could just scrape that list, in which case that sounds like a dupe
 
so basically no KC
ghhgtgg
sorry, just trying to get a piece of fluff out of my computer
 
8:09 PM
well no, the point was just that a lot happen to not be particularly interesting
 
Plus, there's disagreement about ticker vs message feed, and I especially don't want to have to mess around with the feed if we did have it
 
@StackMeter then don't do that over the chat window
also keep in mind that adding message feeds results in a ton of spam because SE feeds suck
 
@hyper-neutrino sorry, it was the thing I had open
 
@StackMeter well no, the point was that if you're going to make one, it needs to have something interesting to it
 
8:11 PM
if we all posted KC of "output this script" or "output those lyrics" we could come up with a challenge every 30 seconds and they'd all be equally boring
for example, recently a "output the bee movie script" challenge was posted and was hammered into "output the lyrics to rickroll" because they both fit into category a) that i mentioned and won't really offer anything new
 
@hyper-neutrino fair enough
rickroll was better anyway
 
@hyper-neutrino But also, the methods used in the rickroll are completely generalisable for any text
That's the important bit that so many people miss in the argument about closing most KCs as rickroll dupes - unless there is obvious compressible patterns, the rickroll methods are 100% generalisable and like 95% optimal
 
yeah
my output titin challenge is an example of something that isn't a rickroll dupe tho, despite people trying to close it as such, lol. it really wasn't interesting and I'm sad that it's my most upvoted challenge but the compression there is totally different lol
 
@hyper-neutrino 62 upvotes how lol
 
that's not even particularly exceptional
 
8:16 PM
@hyper-neutrino keep in mind my best question has 9, and my best answer has 6
don't concatenate those
 
TBF our favourite challenges are rarely our highest voted
 
we all got that joke before you even wrote it
 
right, but you also haven't posted that much, and our community is not nearly as active anymore. also our most highly voted challenges are actually kinda old and underspecified TBH
 
@hyper-neutrino fair enough
I've posted 30-ish things
 
Tbf 30 isn't a lot in the grand scheme of things
 
8:17 PM
the community still exists though, as evidenced by the 10 or so answers on a new question
need to get those numbers up
 
@StackMeter yeah, but it used to be insanely active. We're recovering from the loss of users from 2018-2020, but slowly
 
well it's still indisputably much smaller than we used to be
 
@mods btw should I continue flagging posts on Meta? Am I posting too many low-quality flags at once?
 
@user your flags look pretty good. they give a suggested action and an explanation / where I can look for why you made that call, which is quite helpful. you can continue with that :)
also @mods doesn't ping (i'm sure you knew that). you may as well just ping me TBH :P
 
That's good, just making sure I'm not doing anything wrong
@hyper-neutrino @hyper-neutrino @hyper-neutrino ping ping ping ping ping >:)
 
8:31 PM
*hovers over the 11 button*
 
9:00 PM
@hyper-neutrino My million-byte deadfish answer almost certainly lost :P
CMQ: How do you pronounce 'pxeger'?
 
@Ausername Hi there, "A"! I saw your cousin The just the other day :P
 
Lol
 
@Ausername puck-say-gur
 
pick-seg-eur
 
pʰə̆ɡsɛˈgə? I give up.
 
phaigsege is how I'll pronounce pxeger in my head from now on :P
Perks of not knowing IPA - all spellings look hilarious when spelled using IPA :P
 
True
 
9:31 PM
@Ausername "pecks-er"
 
Uh...
 
I don't pronounce most usernames as they're spelt :P
 
How do you pronounce 'Lyxal'?
 
They'd hate you for that./
 
9:36 PM
Let them, that's the 100% correct way to pronounce it :P
 
Likesal not Licks-all.
 
I also pronounce it "licks-all" except more of an "ull" sound (like in "gull")
 
Yeah, the emphasis should be on the "lick" not on the "all"
 
@lyxal!!!
 
I pronounce pxeger as puhk-seger, but the uh is so small it sounds like pks is basically one group of consonants
(Sorry for no quotes or code formatting, I'm on mobile)
 
9:50 PM
I don't actually really have an internal pronunciation of pxeger's name, I just sort of skip over it when formulating thoughts
but if I had to, I'd probably say something like "pseger" because rather than inserting a vowel between the P and K sounds I more or less just skip the K sounds lol
it doesn't make sense but that's how i pronounce it the few times i try to imagine how i'd pronounce it
 
I see the "p", "e", "x" and "er" and mix that into "pexer" :P
 
How do you pronounce 'Adám'?
Note the á.
 
the way I'd pronounce "adam", because I don't actually know what that accent does
 
Like Adam, sometimes like "ah-DAM" (using caps for emphasis)
 
but since Adám is his real name I should probably learn how to properly pronounce it
 
9:57 PM
How do you pronounce 'dzaima'?
 
"dz-eye-muh"
The "dz" as one "ligature"
 
I can't think of anyone else :P
 
as pinyin, zai ma (without tones) (since this is one of the few occasions where I can actually use pinyin to describe something i'm pronouncing :p)
 
How about I4m2?
 
10:00 PM
imo there is only one sensible way :p but i may be convinced otherwise
 
Wait, it's an I? I thought it was an l.
 
@hyper-neutrino "eye-four-em-two"? I say "I'm-four"
 
it is an I, and I copy-pasted it into TIO to check
@cairdcoinheringaahing yeah no offense but i'm going to consider that not a sensible way :D lol
 
@hyper-neutrino No offence taken, "hipper-neu-try-no" :P
 
good to see we're on good terms, "kye-red ko-eye-n-heering-ay-hing" :P
speaking of which what is your "last name" supposed to be pronounced as
i say "cared coin-herring-ah-hing"
 
10:06 PM
I used to do that
 
@hyper-neutrino Same
 
@hyper-neutrino It isn't supposed to be pronounced, just observed :P
And cut off by things that don't like long strings
 
well, I am not even supposed to be observed :P
@cairdcoinheringaahing yeah i have to zoom my chat window down to 90% to see your full name and i like to keep it at 110%
it works fine at 110% zoom when you monologue tho (is that a term for it? when you send enough messages in a row like i am now and it makes your profile picture expand and your name display under instead of beside it)
 
A little off-topic maybe, but I just uploaded a new track to my SoundCloud page
 
10:20 PM
@hyper-neutrino It's at least the term chat's HTML uses
Any group of messages is .monologue
 
@LuisMendo It's more on-topic than a lot of other stuff posted here :P
 
:-) And at least it's not rick-rolling
 

War on the heathens!

Jun 3 at 20:00, 33 minutes total – 235 messages, 12 users, 20 stars

Bookmarked 2 days ago by user

 
10:53 PM
@cairdcoinheringaahing okay carrid coinengineering
@user mmm I would if you actually had profile text
@user you just don't know your memes very well: youtu.be/KgK0g0M4yLM
 
@lyxal That was the joke :P
 
@Ausername there once was another user here named A
 
@lyxal Looked it up right after looking at your profile :P
knowyourmeme is a very useful resource :P
 
@user very good
Making me proud
@Ausername peh gecks er
Which, come to think of it, isn't even close to accurate
 
11:09 PM
@RedwolfPrograms No one will dare go against the mod for fear of being 11'd :P
I hate it when you ask a question and no one answers or comments on it immediately. Answers are so much easier
 
@RedwolfPrograms I have a proposal: we use AI to settle it
 
How could that ever go wrong?
 
We're already handing too much control over to this AI...next it'll be "should we adopt a strike-first nuclear policy? let's ask GPT!"
 
There's literally a recent MM discussion about giving bots mod powers, seems like letting computers run the SE network is a hot idea tonight :P
 
@cairdcoinheringaahing admittedly, smokey being able to superping users in regards to invalid flags would be useful
 
11:21 PM
Yeah, my objection isn't to the superping :P
 
Bots can fail
 
But you can't have sp without mod powers
 
Its with all the other stuff. Think about it: if smokey has mod powers, it can raise and validate its own flags
 
^
 
Only if we tell it to
 
11:22 PM
^
 
We need humans to confirm what the bots think. There's always false positives
 
The whole point of smokey is that it's human operated
 
@lyxal Not normally. Smokey is unique enough to SE that the staff could always give it the ability to superping
 
@cairdcoinheringaahing That's like saying "if I give my friend a knife, that gives them permission to kill me with it"
 
@user we already do that though
 
11:23 PM
@lyxal With a significant of mods, who can superping
 
@cairdcoinheringaahing but what if the mods aren't available
 
@lyxal I know, and we should continue to do that
@lyxal Elect more mods ¯\_(ツ)_/¯
 
I don't see why an existing mod can't just run a script that superpings users for them
That doesn't require a new mod account specifically for the bot
 
Mods have incredibly powerful tools. I don't know about you, but I am not at all comfortable with giving those powers to anything other than a user I trust
 
I trust smokey
 
11:25 PM
^
A tool that can automatically raise large numbers of flags already has incredible power (and trust)
 
But it can't act on those flags
 
Yes, and it still has a significant number of things that it can't do, and that's the point
 
This isn't some rogue AI that's going to use its powers for anything it's theoretically capable of; if it's programmed to only superping, it will only superping.
 
^^
 
if smokey isn't a mod it won't be able to autoflag lol
 
SD, for example, cannot autoflag on CGCC using my account. it will use my account for any other site except this
 
@RedwolfPrograms If all the bot can do is superping, then it'd be okay, I guess. As long as it doesn't use other mod privileges
 
Which is not what was proposed. Superpinging was all that was needed.
 
The point is that there isn't even consensus within the team question linked in the meta proposal
 
I like having bots around, they can be very helpful resources. However, this network isn't for bots, its for people. Bots should be used when needed and only when needed, and humans should handle the rest - and bots should not be given the most powerful tools we have on the network
 
11:28 PM
I do not at all trust smokey with a diamond. even if the devs all promise to only use it for superpinging. it's just that you now have an all-powerful program (which can make mistakes much more drastic than a human could) and it's hard to hold someone accountable
to put it this way - the bot is the servant of the human, not the other way around
 
And the humans developing it are already trusted members of the community, right?
 
granted, it's hard to imagine how a dev could even accidentally abuse mod tools - like, if only superpinging is involved, that's literally a matter of adding another @ and using user.id instead of user.name or something like that
 
@RedwolfPrograms I trust you a lot, as virtual trust goes. I voted for you in the RO election, I'd probably vote for you in a mod election if you stood. I don't trust NP/SP with RO powers, let alone a diamond
 
@RedwolfPrograms But those humans' programming skills aren't necessarily trusted
 
It's hard to accidentally go flag a bunch of stuff with a script written to superping a list of people
 
11:30 PM
but yeah, ^ + ^^. a mod making a mistake might delete a user by accident. a mod bot coded incorrectly might accidentally start deleting every user
 
That would be quite the mistake
 
Hell, I trust myself more than I trust anyone else, I still wouldn't want to give OSP RO powers and I coded the damn thing
 
@RedwolfPrograms Better safe than sorry
 
I trust a bot 20x more than a typical user
 
@RedwolfPrograms same
 
11:31 PM
@cairdcoinheringaahing See? Even the Dictator of the United States doesn't trust themselves that much :P
 
if it's only superpinging, sure, that's hard to mess up. but that's a lot of trust to be placing in the bot and all devs. i'm not saying they're not trustworthy to me, but the point stands that you need the trust of the whole network for something like this
 
@RedwolfPrograms Trust a bot to do what?
 
@RedwolfPrograms Which is why hyper is a good mod.
 
But we aren't talking about typical users. We're talking about the 0.00001%
 
@user *director
 
11:31 PM
@user What it's told
 
@RedwolfPrograms A bot may do what it's told perfectly, but it's not guaranteed to be told the right things. A typo here, a logic error there, and you could end up in a lot of trouble
 
TBH you know what's a much better solution?
 
All it takes to superping is sending @@... ... to /messages/.../new, with an fkey and some auth cookies. Making it do anything else would take a ton of work. Not something you can easily do by accident, or even on purpose.
 
just require people with autoflag to be in the teams and have some sort of notification system within it or something
 
@user However, I do agree that superpings would be hard to mess up, and as long as those bots are tested extensively before being deployed, I'd probably be fine with that
 
11:33 PM
@hyper-neutrino dang
That's clever
 
like
there's still a chance it'll fail
but there's also a chance people just won't respond to superpings
i superpinged doorknob a couple hundred days ago and to my knowledge he still hasn't visited chat since
 
Nah making autoflaggers join teams is a whole lot better
 
I guess the problem is that a malicious programmer can always sneak in (or an existing developer turn the side) and insert something malicious
 
There are two ways a superpinging bot could go wrong: accidents, and malicious behavior. The first is almost impossible.
 
again, the invalidated autoflag message already just does @{name} so just editing the code to do @@{id} instead isn't particularly prone to error
 
11:34 PM
If HN makes a mistake as a mod, I know that I can ping him here and let him know. Or raise a flag and let at least another mod know. Or even, in some cases, I can go back and fix that. At the very least, I know that I have at least one way of reverting any mistakes he might make. I can't know that with a bot, no matter what people say they made the bot do.
 
somewhat like a bitcoin mining script in a widely used npm package
 
but keep in mind that you also have to think about what SD already does that might change if it is given mod powers, and what damage could arise from that
 
Make it a totally different script that handles the pings.
 
at this point this just sounds like giving a nuke and then thirteen different workarounds and limitations to someone who just wanted a hammer
 
@Bubbler eventstream, yes.
 
11:36 PM
To me it sounds like giving someone a hammer and a list of instructions on what to do
 
The TL has a tool that allows you to summon mods from any site and it seems like it has the ability to superping / send an inbox notification (it doesn't even actually leave an @@<id> message in chat) - this may be possible to implement elsewhere, and would certainly be a much better idea than handing over a diamond just to notify people. — hyper-neutrino 10 mins ago
 
Then getting worried they'll suddenly decide to kill you with it
 
@hyper-neutrino "Here's a nuke. This is the button to launch it, and here's the glass cover over that button. Here's the hammer you'll be using to knock in these nails"
But also, doesn't Smokey require you to sign up to flag in the first place? So, like HN said, just require them to be part of the Team
 
Has anyone else here written a low-level implementation of a chat bot? It's not easy to do even simple things. Accidentally doing something like deleting a user is almost certain to require extensive debugging, not something a mistyped number could do.
 
Here's my final thought on this: if they won't let you use GPT-3 with a bit, why should we let a chat bot have mod powers?
 
11:38 PM
4 mins ago, by Bubbler
I guess the problem is that a malicious programmer can always sneak in (or an existing developer turn the side) and insert something malicious
I still think this is important to consider
 
That'd be my only concern
 
Even if the entire community verifies it, there's always the chance of something slipping in. I don't think we should risk that
 
But the script shouldn't need to change, and I think it should be run on the mod's computer rather than whatever hosts any bots
 
@hyper-neutrino Who? If Redwolf could've done that, they would've done already.
 
@RedwolfPrograms If its being run on a mods computer rather than like, by SE, then extra no
 
11:39 PM
@Ausername ?
@cairdcoinheringaahing Why?
 
That's like so much easier and open to exploits
 
@cairdcoinheringaahing We trust hyper, right?
 
@Ausername This is just hypothetical
 
Yes
 
@Ausername Yes, I trust hyper
 
11:41 PM
If I make a bot that uses my account to do something malicious, I'm still accountable. A mod doing that is exactly the same.
 
How does superpinging work anyway?
 
A bot with mod privs can do incredible damage, but if the mod wants to do that they could write the bot themselves anyway.
Deleting a user on main is likely much easier than anything tangentially related to chat to automate.
 
@Ausername @@id pings the user with that chat id, whether they can be "pinged" or not. Also bypasses the 15 minute wait to the main inbox
 
I trust the moderators.
 
So only for very important situations.
@RedwolfPrograms Ye
 
11:43 PM
But as soon as you give it the tools, whether they're supposed to even be used or not, you've opened that can of worms
 
If it's a script written and run by a moderator, without anyone else being able to modify the code, that removes any potential for malicious behavior IMO.
@cairdcoinheringaahing A bot is not an AI. It does what it's told.
It's not going to randomly decide to delete a user, flag a post, or leak moderator-only information.
 
SE got hacked 2 years ago, where a new user managed to get developer privileges. You think that no-one is going to be able to hack a mod-Charcoal to make bad things happen?
 
to this date, moderators have always been "the people trusted by the community and SE to represent both to each other and maintain a safe environment and uphold community policies". even if we know the bot can only possibly superping, giving a diamond to a bot to grant it one of the weakest privileges we have access to, is honestly dangerously close to,
if not directly yet another violation of trust by SE towards its community, saying that we now have a script controlling and account granted the highest level of trust and privilege that's only supposed to be given to people that the community trusts
 
I'm not suggesting a mod account specifically meant to be a bot
 
11:45 PM
@cairdcoinheringaahing link plz?
 
That was never something I was arguing for
My suggestion was having an existing mod run a script using their account.
 
It's less about what the bot could do, being realistic, and more about what it says, allowing an automated system to have these exclusive powers that represent the ultimate trust
@Ausername Link
 
@Neil know any good local banks?
 
@cairdcoinheringaahing I don't think that's significant enough to matter, SD already has a ton of power in theory, superpinging hardly seems like much next to it
 
11:47 PM
 
@Ausername wow, imagine using imgflip
 
honestly, thing to consider - charcoal itself doesn't even want this
makyen has made it quite clear that this idea is not only not endorsed by charcoal, but that the team actively rejects this proposal and does not wish for it to be implemented
 
22 mins ago, by lyxal
The point is that there isn't even consensus within the team question linked in the meta proposal
 
@SonictheAnonymousHedgehog I understand that you thought it might be useful, but don't drag us into it. The way you've presented this, it sounds like Charcoal is seriously considering, or at least discussed internally and thought it might be useful to have SmokeDetector have moderator privileges, at least in chat. Charcoal DOES NOT want SmokeDetector to have moderator capabilities. The possibility was not brought up on the question you link. Any time the possibility has been mentioned in passing in other discussions it has always been immediately shot down by us. — Makyen 3 mins ago
(emphasis not mine)
 
Right. I'm just saying I don't see anything wrong with the idea, hypothetically, as long as it's implemented well.
Compared to bots, humans are dangerously unpredictable and shouldn't have nearly as much power :p
 
11:55 PM
That's why unpredictable humans aren't elected mods
 
@RedwolfPrograms It would've been perfect if you had logged into NMP and messaged "That's what you think" just then
 
We only elect predictable AIs pretending to be humans as mods, such as hyper-neutrino :P
 
@lyxal Try running this discussion throught GPT-3 pls?
 
TBH i personally more or less subscribe to the belief that a user shouldn't be given more power than necessary
 
@RedwolfPrograms Your proposal of a mod running a userscript sounds somewhat better than making a new bot with mod powers, since it both doesn't have the same significance and lets someone be responsible
 
11:57 PM
since the discussion is about giving smokey superpings, giving them a whole suite of mod tools to go with it is at the very least entirely unnecessary
 
@hyper-neutrino What do you have against me? :( (but seriously, I agree)
Could SE make an exception just for Smokey and let it superping?
 
it's possible, I suppose. and a much better idea than this would be
 
@Ausername convert to a nice format and I will
 
@lyxal Chat doesn't generate transcripts for you? :/
 

« first day (3781 days earlier)      last day (1063 days later) »