« first day (4361 days earlier)      last day (778 days later) » 

00:24
@lyxal Doesn't that mean the potentially infinite list is really finite?
well no
because it's a filter over an infnite list
i think he means infinity list contains the finite list
like [1,2,3] and [1,2,3,4…]
more like [1,2,3], [1,2,3
Oh
where the second is only ever going to contain 1,2,3
but because it's filtered from an infinite list, it'll never actually terminate
00:26
ooooh
So it doesn't terminate after giving the right output
correct
imo it shouldn't be valid
i'd assume not but that's an interesting question
00:27
because you can't really verify that it will not emit some extraneous value
but the filter conditions exclude everything that can't be balid
Well, you can prove it, it's just that it doesn't terminate
i think yes if you can prove it will be the same
Like an answer that outputs through stdout that prints the right answer and has an infinite while loop at the end
in the context: 'L¹= will reject every number after the valid part because it'll be longer
00:32
the rule for default infinity list challenge output is:
1. finite list on input index
2. the item of input index
3. a generator
i think we can change those three rule or some of it to fit finite list challenges too(so that more output method is allowed)
@lyxal Would it be shorter if you used Take/Slice?
no, because then you've got a filter at the end that would need to be closed
1. whole list
2. every item seperately ( loop the list with print for example)
3. a generator that equals to the answer in the end
@user Þp'⁰øp;'L¹= is the infinite list in question
Oh well, you have 11 bytes already
00:36
I know :p
but still, it's an interesting question
although I suppose it's true that it can't be 100% proven it's the same
because there could be a number big enough that python's len fails
and gives a false positive
That's a different thing
That's a limitation of your hardware, and we can pretend our computers have unlimited time and unlimited memory
Well, get outgolfed lyxal
@emanresuA every byte shorter means less chance of being beaten by other languages :p
00:52
Well, competition is within languages, not between, so that wouldn't matter
@user it does for PR :p
 
1 hour later…
02:18
> Sorry for the wrong language No Habla C#
that just made my day
02:48
-70
Q: ChatGPT should be incorporated into the site

DialecticusChatGPT is a tool. We should adopt all useful tools. This site is in the best position to do this adoption. I propose that before the question is published the OP gets the opportunity to read the answer from the AI. The OP might already accept this answer. Or they might realize that the question ...

that might genuinely be one of the most profoundly wrong ideas i've ever seen
-329
Q: Shop system. Reputation points buyable for real currency

John LockMake a shop system, in which we can buy reputation packs for real money. There should be daily limit on how much of reputation we can buy a day. There could be also other features to buy. Here are my proposals: Golden frame around questions and answers. Bigger font of our comments or ability to...

oh yeah that
that's worse
Gotta love SO
There's probably some worse ones on MM
@emanresuA bruh whaaaat?!?! legit p2w lmfao
03:01
They should make it so, once you hit 25k, you can buy rep
cause at that point, it has no actual benefit, it's literally just vanity
Or sell it
*35k
(aside from I guess, a few extra delete votes)
Also don't you get extra flags from rep or something?
Don't think so
I think it's based on helpful flags
Ngl, I used to know rate limits, but since elected, I've entirely forgotten, as I'm not limited (⌐■_■)
 
1 hour later…
04:14
> One-time-use moderator actions for buy - except ban.
So you're saying for $1.99 I can change people's pfps? Sign me up :p
 
2 hours later…
05:58
06:21
classy
Copiloy is sometimes concerning
Like chatgpt, it writes stuff that looks like it works
06:37
Hello all
I am very tempted to pose a fiddly coding task I just had as a challenge
0
Q: Remove duplicates from my academic transcript

SimdConsider a list of subject, grade pairs. E.g. [("Latin", "A"), ("French", "A*"), ("Math", "B"), ("Latin", "A*")] The task is to return a list of pairs that has each subject at most once. Where a subject occurred more than once originally, the returned list should have the highest grade. Using ...

 
3 hours later…
09:22
> Ability to remove some reputation from user. The more reputation for remove, the bigger cost.
awesome
It's SE Nitro all over agaain
I'll say this for your idea: turning SO into a PvP MMPORG might be better than turning it into a social network. — jscs Jun 20, 2016 at 17:50
En garde, nerds
I cast fireball!
09:39
Is ~0 the same as -1?
10:07
only in 2's complement
10:37
Basically for signed numbers yes
11:59
That seems more like TTS than chatGPT
obviously
the whole point is that it's speech synthesis
like dalle but voice
12:41
Not really much like dallle either
it is
it's text to some non-text medium
voice clips instead of pictures
It's literally the same text though
Unlike dalle that has to understand the prompt
it's still a text to non-text medium service
that's the comparison I was making
Ok I think the comparison is weak but yo are entitled to your opinion
DALL-E = text -> image (not text). VALL-E = text -> speech (not text). That was the whole point of the joke.
12:46
Ok I didn't get it was supposed to be a joke
13:04
CMQ: Which of these two generalised fizzbuzz questions should be closed as duplicate? (if any at all)
because they're basically the same except for input formats
I don't like the old one since it requires a full program and has a very strict input format
 
2 hours later…
15:29
Hello all
15:53
hey
16:43
hi
CMQ can codegolf.stackexchange.com/questions/256441/… be done with regular expressions only?
No clue, probably not
@WheatWizard I hear an infinite number of wizards writing code from an infinite number of ivory towers will eventually produce something that compiles correctly. (The existence of the PHP standard library would seem to suggest that this has already happened.)
17:00
@Ginger I need the view of a regex enthusiast
hey d:
@Simd Maybe, by hardcoding all 36 possible combinations
All A not followed or preceded by a A*, All B not followed or preceded by a A or A* etc.
Would be very long with this apprach though
@lyxal I'd say the old one should be closed and the new one should be a dupe target for any fizzbuzz questions
@mousetail that sounds interesting. That's not no many
@mousetail what about if you only had A*, A, B?
@lyxal (also, my memory's a bit foggy, why did I comment "Welcome to Code Golf Stack Exchange!" when Radvylf was already an established user here?)
17:10
@Simd Oh nevermind, you can't use groups in lookbehinds so it's impossible to tell if a subject has appeared before
@Simd Probably a regex-based language, like Retina
This will match unique subjects: \[(\w+) (\w)](?!.*\[\1 \2\]) but I don't know how I could make it match only the best one in the case the B comes after the A
Unless you can guarentee grades will be in descending order
nvm I can't read
Unless... I may have an idea
17:28
Nope it doesn't work
Would using Perl's flavor of regex make it possible?
What to do when an answer have trivial improvement and author is away(grey name)?
Post your own answer, I guess
Or just leave it alone if you don't want to bother
@lyxal so how does W fit in? GIGO?
17:50
I feel this looks like a regex challenge now :)
@lyxal That's identical to URL shortening isn't it? Just a custom URL shortener
18:27
^
18:46
Was this answer written by AI you think? codegolf.stackexchange.com/a/256466/91213
I think so
That answer made some amount of sense, this one doesn't codegolf.stackexchange.com/a/256467/91213
We should definitely not presume anything is AI without hard evidence tho
I mean it doesn't even matter, it should be deleted regardless
Yeah. But just so that nobody goes and comments "using AI to write answers isn't allowed here" or whatever
18:52
Hugging Face thinks it's AI-written with a certainty of 59%
That's not a very high certanty
@mousetail Can we assume that every answer which starts with "To solve this problem" and ends with "I hope this helps!" was written by gpt3?
IMHO yes
But Radvylf doesn't think so
@mousetail that one Hugging Face thinks is AI with 99.97% certainty
(Completely unrelated but hugging face sounds like the enemy from a horror film)
19:02
I think we should have a community consensus on this, "It looks like it was made by GPT" won't do the job.
0
A: List of bounties with no deadline

Wheat Wizard50 reputation for first answers in FP languages I find golfing in functional programming languages, and I'd like to get more people involved to make it more competitive. Therefor I am offering 50 reputation for each user's first answer in: Haskell Curry Elm You may claim all 3 bounties. In orde...

Hugginface sounds like the PG version of facehugger
@mathcat Want to make a meta post?
I guess I could do it, yeah
which reminds me: at some point I'm going to make a bot that analyzes incoming SE posts with that ChatGPT-detection AI
19:06
Seems like AI detection is something that'd be useful in something like Smokey
^ @Ginger consider submitting a PR to smokey instead
nah
Smokey's job doesn't extend to that
and
there are already non-Smokey bots that help fight spam
making one bot do everything isn't a good idea
@Neil are you a regex expert?
I try to be but the real expert is @Deadcode
19:22
@NewBountiesWithNoDeadlines Do polyglots collect multiples of 50?
for that latest question of yours you'd like both variable length lookbehind and inline code expressions
0
A: Sandbox for Proposed Challenges

Joao-3ABACABA pattern without recursion ABACABA pattern on Wikipedia Taking a number n as input, generate the ABACABA pattern with n symbols as stated in the above page, without using any form of recursion in your code. You can use any output method; characters, a list of numbers, an image, whatever is...

it could probably done with just variable length lookbehind and conditional expressions but it would be a bit of a nightmare to write
@SandboxPosts well that's going to get closed as unobservable
19:42
@Neil more of a nightmare than writing in Jelly? :)
@Neil is there a library that supports those?
19:58
@mousetail yep
20:32
@Simd I don't know of one
@Simd isn't there already a Jelly answer to your question?
@Ginger Why not? Smokey has a bunch of people listening to it already. If you made your own bot, would you make your own chatroom and alert system for it?
From the Charcoal HQ room description:
> This room is for Charcoal, a volunteer organization focused on detecting and eliminating spam and rude/abusive posts on all SE sites. (emphasis mine)
Ignore my last two messages, I misinterpreted your message about non-Smokey bots fighting spam
20:55
> Did you know that you can define a string with a negative length in PL/1? Excellent for eating the last characters in a string: just concatenate your string with the negative-length string
21:09
what
source?
@Neil yes there is!
@Neil I am never sure what pcre2 covers
@Neil look behind seems supported rexegg.com/regex-lookarounds.html
@Simd not truly variable length though
21:32
@Neil ah. Is that a problem for my short strings?
@Neil The issue is that variable length lookbehinds don't seem to work properly
@mousetail is that in a particular library?
I tried a few of them on regex 101
I think the issue is it creates a type of self-contradiction that most parsers are not able to handle
@Simd not sure, might be able to bodge something up with recursive regex
21:58
@mousetail the algorithm in the answer is awfully similar to what it just gave me
@Neil I like the sound of that
@mousetail ah
@Neil I would pose it as a challenge but I don't know if you are allowed to ask for regex only answers
@lyxal whoa. What do you think?
Can you ask it for a regex solution? :)
No, because it'll almost surely butcher it
And the point of the gist was to show that the answer in question is most likely ai written
@lyxal this is 90% proof
22:19
-1
A: Division between two strings

Mohamed FetouhCheck if the multiplied string is empty. If it is, the multiplied integer is 0. Check if the multiplied string is the same as the result of the multiplication. If it is, the multiplied integer is 1. Check if the multiplied string is the reverse of the result of the multiplication. If it is, the m...

The way it just lists the stuff at the top is similar to what you get when you copy paste something the ai has formatted as a list
Because it looks like a numeric list on the OpenAI site, but doesn't copy across as so
Is there any way to get confidence intervals in python? For example for a correlation coefficient you might compute. R is stuffed full of libraries to do that
@lyxal hmmmm
that is 100% ai too
23:19
Continue — l4m2 4 hours ago
What does this mean?
the link is probably for padding to get to the 15 char min limit
but maybe it's an attempt to trick the account into making a new post if it's a bot doing it :p

« first day (4361 days earlier)      last day (778 days later) »