when’s the next mash im FIENDING for a mash right now
10 Likes
ive thought a bit about vibecoding a really condensed down mash app (like 1 hr game length total and like 40p games). This surely could not go wrong.
7 Likes
NOOOOOOOOOO
1 Like
id play it
2 Likes
As someone trying to code action resolution for Mafia Bot, have fun.
3 Likes
I mean thbthbthtbth if you’re willing to restrict yourself to simple-ish roles it could probably do it
3 Likes
i mean itd just be as simple as targeting the player with the most votes or whoever the mafia killed
Mountainous is the only good setup so you shouldnt bother making a tool for anything else
2 Likes
just put in little-to-no roles that mess with other roles and it should be fine
1 Like
it would be pretty light on roles. days would be like 10 minutes long or something so its not like I could really fit in too much mech anyways. I wrote out a little plan for it it’d be a fun proof of concept
2 Likes
SCOPE MANAGEMENT [Trivial: Failure] - But what if I don’t…? Break past your self-imposed chains! Bring the light to the masses!
1 Like
yea fair
vibecoding, on the other hand
3 Likes
LMAO
1 Like
its knighr errant but cool
1 Like
Theres like not even anything really wrong with vibe coding, it just usually creates things with subpar quality
And like, that can be fairly easily fixed if you just like, know what ur doing (or care enough to put in the time to make it not subpar quality)
4 Likes
kay fair
1 Like
actually if you vibecode you go straight to the seventh circle of hell and also lose your robux
12 Likes
I’ve started to listen to covers that are way less good than the actual song but I’m kinda tired of the actual song but it still has good lyrics
5 Likes
We’re seriously considering locking down the repositories for our international tooling because vibe-coded MRs are wasting too much time. Like they’re being submitted reviewed and ultimately rejected over and over and over again without any improvements. My team lead vibe-coded an app because we got told out of nowhere it needed to be done in a week, and what the LLM spit out was technically functional. Every part of it has since been replaced because it was bad. I’ve used a frontier model to produce code because my company is paying for it and the COO asked me to try it out to get my opinion on it actually having it used it, and it like… does produce technically functional code, but it takes me a couple hours to review everything and it’s usually weird and bad in ways that are hard to wrap your head around quickly unless it’s doing something that’s a sophisticated find/replace. Anything novel or requiring a broader understanding of the problem/project or dealing with architecture is mostly unacceptable garbage that would be actively bad to keep in the project. I am unconvinced it saves us any time in the long-run because more and more we find little hidden pieces of shit an LLM left for us in a dark corner of the code-base that’s usually technically functional, but bad for the long-term health of the project.
6 Likes