#stackOverflow
An advice for young folks learning to code. Please get comfortable coding the old school way, looking up solutions on stackoverflow etc before you start getting AI assist.

You can trust AI with menial tasks but not with critical code components. It is nowhere even close to the human brain. 🧪⚛️
February 4, 2026 at 6:14 PM
In case you ever wondered about the differences between magrittr pipe and base pipe 🤔

This table is taken from a great stackoverflow answer by @GeorgKindermann

stackoverflow.com/questions/67...

#rstats #dataviz #phd
February 2, 2026 at 5:39 PM
update: i fixed my website's CSS finally! (thank you stackoverflow)

might write some articles i've had in mind that i've been too lazy to write partially bc of tech issues (my css and img serving broke, now both are fixed)

www.experimentalsynthesizer.blog/articles
Experimental Synthesizer
www.experimentalsynthesizer.blog
February 5, 2026 at 2:32 AM
"C'est prouvé : Le vibe coding va tuer l'open source"

#VibeCoding #VibeCode #OpenSource #Programmation #StackOverflow ...

korben.info/vibe-coding-...
February 4, 2026 at 11:23 AM
yea but thats a lot different from ai can code. ai will be used as an excuse to fire people because to a person with an MBA 'can copy paste from stackoverflow' is a lot of what junior devs do now and thus $200/mo for claude is the same value to them as a real person at 1/20th the cost
February 4, 2026 at 11:28 PM
reddit & hackernews for AI agents, what's next? stackoverflow?
February 2, 2026 at 7:38 PM
How come I keep getting portaled to Asphodel? "i copied that script from stackoverflow, if I remove it everything breaks"

where are the fates? "uhmm i just put them in the temp folder i think"
February 1, 2026 at 11:07 PM
i think it's because a) software has traditionally been a place where copying has been more accepted (see copy pasting code snippets from stackoverflow, heavy boilerplating, etc.) and b) the tools have started to hit an inflection point where some of the models are becoming actually useful
February 2, 2026 at 7:55 AM
Sobald die Kids im Bett sind zerlege ich den Stack nochmal. Irgendwo wird da die config nicht richtig gemounted. Und dem Python Script von ChatGPT trau ich auch nicht. Kann leider nur typescript/php/JavaScript und den Bums. Docker ja, aber viel stackoverflow und probieren. Doku lesen.
February 5, 2026 at 6:20 PM
Vor LLMs kopierten die Leute Codezeilen aus Foren wie Stackoverflow und passten sie an ihre eigenen Daten an. So bekommt man ein Gefühl für das Programm, aber es ist auch sehr zeitintensiv. LLMs werden anhand vieler dieser Forenbeiträge trainiert und können daher blitzschnell Vorschläge liefern. 7/9
February 1, 2026 at 2:44 PM
Aujourd'hui j'ai fait une recherche sur StackOverflow, improbable non?
January 29, 2026 at 2:20 PM
And finding a reddit or stackoverflow post asking the same question without any helpful answers and the conclusion “I found the solution and got it to work. This question can now be closed.” Of course the account has long since been abandoned.
January 26, 2026 at 7:24 PM
Considering that it was ~30% at the time of the Stackoverflow 2025 survey (May 29, 2025 to June 23, 2025), chances are pretty good that agent adoption (at least occasional use) has crossed 50%. The "don't plan to" was already <40% then.

survey.stackoverflow.co/2025/ai#ai-a...
AI | 2025 Stack Overflow Developer Survey
survey.stackoverflow.co
January 29, 2026 at 7:28 PM
Ah yes, there was a similar trick for solving programming issues: going to Stackoverflow, asking a question, then using another account to give a wrong or inefficient answer. It worked like a charm.
January 26, 2026 at 8:00 AM
which raises the question: if you have to check, why not check first?

of course, then you fall into the trap of search engines having become dogshit, and stackoverflow shriveling into dust, and all kinds of useful little bits of info hiding in discord servers
January 26, 2026 at 9:23 AM
It's like the person I heard of who found that posting a question on StackOverflow and then immediately answering it incorrectly using a sock puppet account was an infallible strategy for baiting nerds into supplying the correct answer who otherwise wouldn't have bothered.
stop using AI to do your research. it hallucinates too often. if you want an answer to something, post something arrogant on the appropriate subreddit. something like: "this item performs 10% better than everything else. only idiots deny this." this will bait nerds into doing your research for you.
January 27, 2026 at 10:50 AM
this SWE lady I know used to do the “Tee hee I’m an SWE girlie how do I <blank> on Stackoverflow” thing and it was like instantaneous
January 25, 2026 at 8:57 AM
Tiens, nouvelle entrée dans ma collection : cette fois, les forums type stackoverflow vs les grands modèles de langage (LMM) : moins on utilise les premiers, moins on a de chance d'y trouver se qu'on cherche, plus on se replie sur les second, moins on utilise les premiers…

bsky.app/profile/flo....
Article intéressant qui relie le sujet des LLM pour le code à la disparition des espaces partagés d'apprentissage et de feedback (les forums ouverts) et une phénomène "d'enclosure" par les éditeurs de LLM privés, et la relation asymétrique que cela crée et mène à la disparition des espaces ouverts
The Enclosure feedback loop
michiel.buddingh.eu
January 27, 2026 at 3:19 PM
I'd read "Why I'm not on MathOverflow" along with the other posts about MathOverflow linked in comments on my question. But I hadn't seen the discussions about StackOverflow - thanks! Many of the issues must be parallel.
January 27, 2026 at 3:16 AM
claude code will cause the destruction of the website stackoverflow (this can already be seen in their precipitously declining traffic) because it turns out people don't like being told non-answers on how their code works by rude surly IRC users and would prefer a nice robot actually answers instead
January 26, 2026 at 7:41 AM
Au début il y avait les dinosaures, ensuite on a eu la doc Microsoft sur disquettes, puis sur CD, puis Internet et les forums de discussion, puis StackOverflow le wiki des développeurs, maintenant on a l'IA, et encore mieux aujourd'hui L'IA intégré à ton IDE, prochaine étape la disparition des dev😭
January 23, 2026 at 2:13 PM
Proving my code isn't AI-generated because it has links to StackOverflow in the comment for every function so I know where to look when something breaks.
January 21, 2026 at 8:54 PM
January 20, 2026 at 4:15 AM
me: wrestling with yet another buggy Microsoft tool, browsing stackoverflow to try and find a fix
the only other person in the world with the exact same problem:
January 20, 2026 at 3:31 PM