« first day (3679 days earlier)      last day (123 days later) » 

12:10 AM
REFRESH! There are 7306 unanswered questions (90.5126 answered)
2 hours later…
2:14 AM
there's actually another site for reviewing working code: codereview.stackexchange.comjhnc 59 secs ago
Hi @pacmaninbw! Thanks for contributing. Indeed, there are cases where iterative implementation works better than recursion. I remember that factorials are a good example even though there are not a data structure, wish I had an example with data structures now. BTW I tried to read your code review question but I didn't understand it, so I'll keep going working on my skills to be able to read code like that — Juan Jáuregui 12 secs ago
3:02 AM
Q: For loop to download files once and verify them on each invocation

Chris StoneThe complete script works as expected. I am able to download and verify the files. If the file is all ready present verification is retested and downloads if necessary. I'm in the process of teaching myself bash/sh scripting. Is there a cleaner way I can implement the following section of sh scri...

I’m voting to close this question because there is another site for reviewing working code: codereview.stackexchange.com — Nic3500 30 secs ago
This may be more on topic for code review SE. Be sure to take their tour if you're new to the site. — ggorlen 21 secs ago
3:52 AM
possible answer invalidation by saxbophone on question by The Compile-Time Comedian: codereview.stackexchange.com/posts/285218/revisions
possible answer invalidation by saxbophone on question by The Compile-Time Comedian: codereview.stackexchange.com/posts/285218/revisions
possible answer invalidation by saxbophone on question by The Compile-Time Comedian: codereview.stackexchange.com/posts/285218/revisions
4:59 AM
I was the first question that I asked. Can you explain what I am asking here? — Mansour Dalir 3 hours ago
Language barrier.
If anyone can fix that up, I'll be back later today.
5:35 AM
Q: Extracting area of the clusters of specific coordinates

th145I am working in GEE and I am trying to extract areas of clusters for some specific coordinates. I have a clustered map of the area and a list of coordinates. I am trying to extract values of those clusters whose coordinates are available. I have tried this approach but it isn't giving me the area...

2 hours later…
7:21 AM
This question might be more suitable for code review.se — mousetail 16 secs ago
7:36 AM
Is this more a question for Stack Code Review rather then overflow? Here we tackle problems, where in review they can help you with something like this a little better without people getting a little (lot) narky — Simon Price 15 secs ago
2 hours later…
9:11 AM
I’m voting to close this question because it belongs to codereview.stackexchange.comMichael Butscher 34 secs ago
9:48 AM
Q: Computation of average speed in Python

Fabiosomeone can help me with this code? I'm trying to compute the average speed every 100 m, starting from a list of spans that contains [distance, time, avg speed]. This is what I've tried to do, and seems to work pretty well (can you confirm?). The problem is that the code is a little bit bulky, an...

3 hours later…
12:49 PM
@Duga Does not invalidate either of the answers.
1:27 PM
@Mast I have read CR's What topics can I ask about here? (again) and I don't see why it wouldn't fit there. — Michael Butscher 42 secs ago
2:04 PM
@SimonPrice This question would be closed on Code Review as being too hypothetical. It is not real code from a real working project based on the type names. — pacmaninbw 23 secs ago
2:39 PM
...if what you want is just general-purpose code review, instead of a solution to a narrow, specific technical problem, we have a separate site for that: Code Review. — Charles Duffy just now
1 hour later…
3:59 PM
@MichaelButscher You might as well review it then, it's posted on CR nowMast 57 secs ago
4:15 PM
4:34 PM
Q: Julia Set Animation

Laramie DillingerI am new here as well as new to python and I was looking for a little help with this animation. Currently, my code outputs 10 graphs instead of an html5 video with 10 frames. Ultimately, I would like to be able to specify the number of frames and watch this julia map evolve over that time period....

4:50 PM
@CaptainObvious RBA
5:16 PM
Why didn't anyone tell me there was a brainfuck zombie waiting to be slayed?
@SimonForsberg there are 3 others
do you care about ? there are 69 of those
@SimonForsberg Slayed another zombie.
5:32 PM
@SᴀᴍOnᴇᴌᴀ The others are brainfuck interpreters, not brainfuck code.
@SᴀᴍOnᴇᴌᴀ I'm monitoring all the incoming Kotlin questions, might go through some of the zombies over time.
@Mast What changed?
@SimonForsberg Hah, first thing I asked them is whether they were going to make a changelog.
@Mast It would for sure make the reading investment so much shorter.
git diff is underrated.
There are a lot of sub-page now.
The version of 4 days ago can be found on the Wayback Machine
@SimonForsberg Because code takes two levels of crazy!
TL;DR: Things got longer. More examples have been provided and if you found the previous one restrictive, you're going to love (/s) this one.
Most of the links you'll see in the new CoC were not subpages before.
5:41 PM
@Mast eh, I'll keep doing the same as before and adapt if anyone gives me a warning or informs me about what I'm doing wrong.
Especially pages like sensitive content, self harm and inauthentic usage are either new or used to be unlinked.
@SimonForsberg You'll be fine. There's a lot of rule lawyering afoot, but that shouldn't concern the average upstanding citizen.
For the bigger sites, there were some concerns about suspensions being handed out for no obvious reason (there were still good reasons, they simply weren't obvious for some people) and situations that technically weren't against the CoC yet that should've been.
Considering we've always used the unofficial rule that at some point, you're not breaking a rule but the community, this wasn't a problem for smaller sites yet.
SE is adamant that some things should be more strictly codified.
Which is a bit of a hassle for moderators, but on average shouldn't impact the users much.
However, there's also a strike incoming.
Not necessarily about this, but more about the recent upheaval about how the CEO and Team talk about AI participation.
And the disrespect to the community.
This could be '19 all over again.
So, fun.
For some values of fun.
5:57 PM
@Mast How do they balance the AI stuff with the inauthentic usage?
They do not.
Kindling dumpster fires.
Oh, this is beyond kindling already.
How can you have a rule about inauthentic usage and allow AI input?
6:15 PM
You think logic has anything to do with this.
I can understand the confusion.
But if you're ever foolish enough to get a diamond to your name, you'll find out SE and logic don't always go hand in hand.
We removed the automatic postings of the blog because it was a load of horse pucky lately, but consider the following posted today:
Prashanth Chandrasekar on May 31, 2023
With all the significant changes in the industry, one thing has remained the same: companies are committed to driving productivity and efficiency throughout their organizations, and we continue to help our customers and community deliver both.
Once you filter out the buzzwords, there's nothing of value left.
@RMunroe Earthlike stars and exoplanets in our solar system? Yea, that would be weird...
The dictionary would like a word with you.
@Mast yeah, I read all the comments on the "AI moderation policy update". I definitely got the feeling that "Oh, it's drama happening on SE again"
@SimonForsberg And that's the public drama.
@Mast I'm pretty sure that's published for the shareholders and the press, not the community.
6:25 PM
It is.
And as long as it's just buzzword bullsh** bingo that has no bearing on what we do, nobody cares.
But this time, it's all related.
There are more changes incoming.
Winter is coming.
Academics already lost a mod over the upheaval.
Others are presumably writing their letter of resignation.
Someone needs to put a stop to calling generative models "AI"
Others want to be fired over this.
@AriaHarper It's a lot of A and not enough I.
I know.
We used to call every 'machine' that did something pseudo-intelligent AI.
Names change.
"AI" A computer or program that tries to do something you thought only you could do
6:30 PM
Used to be that every computer player in a game, even as simple as tic-tac-toe, was considered AI.
@Mast I still call them AI and I will continue to do so.
Genetic algorithms too are like... "That's not intelligent tho, it doesn't "learn", it just... brute forces... somewhat efficiently"
(even though I know they're not "real" AI)
Even the current LLM are not 'real' AI.
@AriaHarper And doesn't the neural networks do the same thing? All they do is optimize a function to minimize the loss function.
@Mast Especially not when they do this:
You would think that a language model would know spelling, but nope...
6:34 PM
That's the neat part, it doesn't know. It just guesses the next token, really, really, really well.

Except for when it doesn't guess really well.
@SimonForsberg Ask it for 5 examples of anything at all. Count them.
It usually defaults to 6.
Meanwhile I can use ChatGPT efficiently to reformat some stuff for me, it mostly saves some time
Last week I asked it to vectorize a couple of nested for-loops for me and it came up with dataframes that didn't exist.
The ironic part is that since it was trained on data from the internet, the types of content it's really good at generating is the type of content they don't want it to generate.
Meanwhile the industry is split between making models to make believable answers, and making other models to distinguish fake answers from real ones.
The general consensus is that if you want good data to be trained on SE, you shouldn't allow AI on SE.
Otherwise the AI will just train itself on AI.
@CaptainObvious @pacmaninbw Maybe you'll like this one.
6:43 PM
@Mast I know C++, but this.. screams internally
Yea, but pac knows a bit more C++ than you do :-)
I'm not that familiar with modern C++
I realized today that unit tests are pretty dang useful if you want to be able to review efficiently
Thinking that you know that the code does what it says and verifying that the code does what it says are two different things
@skiwi It can get really bad with low level languages too, because sometimes they let you do some really cursed stuff.
@skiwi I used to be a hardware test engineer. I know the feeling :-)
In the past it usually wasn't an issue because there was no formal review process
6:49 PM
When they do finally figure out AI, it's just going to be a simulation of a brain (a true neural network) on a computer...
From a coding perspective, I'd think that'd be pretty easy to do (We already know how neurons in animals work). But I know the scaling is horrible.
It wouldn't surprise me if the brain works similar to the AI we see these days but then on a vastly superior level
Maybe the brain also guesses the next token very very well ;)
The language processing part of your brain is only one piece of the puzzle :3
I mean, current AI models are layered. They have no state (minimal state, reset after every run).
The brain is just a mess of neurons connected to each other, with some general guidelines for connections across different sections to allow communication.
As long as there is this concept of a soul we don't understand, we have no idea about how a human brain actually functions.
Emergent behavior
We know the mechanisms of individual neurons, and how they interact with each other
Stick a couple billion of them together and it controls a flesh mech
Add some more, you might just get the ability to develop unique languages
The human species as a whole isn't really ready for a true AI yet, tbh. We have too many disagreements.
I hadn't thought about yet... think of the human brain, but a thousand times more powerful
7:04 PM
@AriaHarper I think it took another world war before we reached Star Trek levels of agreement.
@skiwi Think of a human brain, that survives multiple generations.
Currently politics is too much focussed on borders.
Hundreds or thousands of years.
Get rid of borders and you got a whole lot of unfocussed energy to put on something else.
Q: Identifying trees in text using python nltk

Stefan BartellThe goal was to identify trees in segmented text conforming to a context-free grammar (CFG) using the python package nltk. The CFG has rules like "DosageP -> substance num amount_unit", and "text1_tokenized" has sublists like ['substance', 'num', 'amount_unit'], which correspond to sublists in “c...

7:08 PM
@Mast The easiest way to have universal agreement is to have nobody left that can disagree with you
That's a zero-sum game.
2 hours later…
9:13 PM
Q: go: cli app to spin up a cluster

user219820This is my first go project. I've written a cli app to provision a database cluster on docker containers. Below are the flags currently used (to give an idea of its capabilities): version := flag.StringP("version", "v", "", "CBS version") noinit := flag.Bool("noinit", false, "Do not initialize...

1 hour later…
10:29 PM
Q: Coffee Machine practice project

BeginnerThis was simply practice to touch on somethings I have learned. I have also tried to take into account Pep 8 formatting practices to my best ability and applicable advice from my previous posts. It was difficult to keep the indents to 2 levels in. Constructive criticism much appreciated. main.py ...

10:51 PM
Your second question is off-topic here. Questions asking for peer review to improve working code belong on Code Review, which was created specifically for that purpose. — Ken White 27 secs ago

« first day (3679 days earlier)      last day (123 days later) »