This media is not supported in your browser
VIEW IN TELEGRAM
I remade the whole GUI from the april fools version from a few years ago and gave it a more modern look. I added the air bubble and the xp level it was missing, do you think they look proper?
https://redd.it/1t3k7qa
@MinecraftModded
https://redd.it/1t3k7qa
@MinecraftModded
I'm building an Airship in Aeronautics, and it looks surreal with shaders
https://redd.it/1t3n60y
@MinecraftModded
https://redd.it/1t3n60y
@MinecraftModded
Reddit
From the feedthebeast community on Reddit: I'm building an Airship in Aeronautics, and it looks surreal with shaders
Explore this post and more from the feedthebeast community
This media is not supported in your browser
VIEW IN TELEGRAM
I created a mod that fixes the crafting output slot to show the actual total number
https://redd.it/1t3n82i
@MinecraftModded
https://redd.it/1t3n82i
@MinecraftModded
This media is not supported in your browser
VIEW IN TELEGRAM
Immersive portals - block face shading ruins gravity change seamlessness.
https://redd.it/1t3qlnr
@MinecraftModded
https://redd.it/1t3qlnr
@MinecraftModded
Sodium backport to NeoForge 1.21.1 is happening major improvements on the way
https://redd.it/1t3v8bu
@MinecraftModded
https://redd.it/1t3v8bu
@MinecraftModded
Reddit
From the feedthebeast community on Reddit: Sodium backport to NeoForge 1.21.1 is happening major improvements on the way
Explore this post and more from the feedthebeast community
Ok, we really need to do something about AI mods.
The other AI content ban post seemed to die out, and I wanted to add to the conversation after decompiling a seemingly generic AI generated mod that quickly turned out to be one of TWO mods which injected DLLs into the JVM in order to seize control and run the game as... they(?) intended it.
The arbitrary program ran as a remote process, and quite literally force unloaded any class that touched their mod's stuff (unless it had a package that matched a 'whitelist' prefix), which inevitably corrupted people's worlds whenever a disallowed mod tried to get involved.
Their wording also made it sound like this was an optional 'feature', when it was copied out, and then actually loaded internally from the jar itself, not dissimilar to an actual trojan. And this made its way onto Modrinth.
Regardless of if this was harmful in of itself is subjective. What it actually did was something no developer would ever want to do. ACE is a massive security vulnerability, and writing native windows x86 code for a mod defies the purpose of Java to begin with. There are like 5 mods which can justify doing any of this, but the AI was just told to stop all attempts to modify the mod's behavior, and so it did.
And even earlier, there was a vibecoded 'minecraft wrapped' mod which was poorly designed and leaked user information all the while only using HTTP even for things like session IDs. This mod seems to have been taken down/rejected by Modrinth, though.
What I'm trying to say is that vibecoding makes people very confident in their abilities to make a mod, without understanding what they're actually doing. If people want to share it with their friends, they can, but AI generated mods introduce a huge security and mod compatibility risk.
AI will very likely take the path that all the online examples follow, which is exactly what can make it so dangerous. HTTP is fine when it's just generic information. It is not fine if you go to somewhere (hotels, schools, etc) with shared wifi and handle personal information insecurely.
I am saying this to express what I think should be done here:
\- Posts that make use of AI, or that are AI-translated should be flaired for clarity. Some developers use AI and I'm fine with that as long as they can understand their own code.
\- mod posts should be put under increased scrutiny in some way, because of the much larger risk of damaging someone's worlds, or worse, computer. Fully AI mod posts (description is ai for no reason), or mods that show obvious signs of complete vibecoding should be taken down.
\- mod posts might be better off also linking to source code, although this can be used misleadingly.
The lengths these AI mods are increasingly going towards are getting incredibly concerning, and it's only a matter of time before someone with actual malicious intent realizes they can just ask AI to write some niche malware, post here, and infect actual people.
I worry that we never learned from the Fractureiser malware incident 3 years ago.
https://redd.it/1t4dyvx
@MinecraftModded
The other AI content ban post seemed to die out, and I wanted to add to the conversation after decompiling a seemingly generic AI generated mod that quickly turned out to be one of TWO mods which injected DLLs into the JVM in order to seize control and run the game as... they(?) intended it.
The arbitrary program ran as a remote process, and quite literally force unloaded any class that touched their mod's stuff (unless it had a package that matched a 'whitelist' prefix), which inevitably corrupted people's worlds whenever a disallowed mod tried to get involved.
Their wording also made it sound like this was an optional 'feature', when it was copied out, and then actually loaded internally from the jar itself, not dissimilar to an actual trojan. And this made its way onto Modrinth.
Regardless of if this was harmful in of itself is subjective. What it actually did was something no developer would ever want to do. ACE is a massive security vulnerability, and writing native windows x86 code for a mod defies the purpose of Java to begin with. There are like 5 mods which can justify doing any of this, but the AI was just told to stop all attempts to modify the mod's behavior, and so it did.
And even earlier, there was a vibecoded 'minecraft wrapped' mod which was poorly designed and leaked user information all the while only using HTTP even for things like session IDs. This mod seems to have been taken down/rejected by Modrinth, though.
What I'm trying to say is that vibecoding makes people very confident in their abilities to make a mod, without understanding what they're actually doing. If people want to share it with their friends, they can, but AI generated mods introduce a huge security and mod compatibility risk.
AI will very likely take the path that all the online examples follow, which is exactly what can make it so dangerous. HTTP is fine when it's just generic information. It is not fine if you go to somewhere (hotels, schools, etc) with shared wifi and handle personal information insecurely.
I am saying this to express what I think should be done here:
\- Posts that make use of AI, or that are AI-translated should be flaired for clarity. Some developers use AI and I'm fine with that as long as they can understand their own code.
\- mod posts should be put under increased scrutiny in some way, because of the much larger risk of damaging someone's worlds, or worse, computer. Fully AI mod posts (description is ai for no reason), or mods that show obvious signs of complete vibecoding should be taken down.
\- mod posts might be better off also linking to source code, although this can be used misleadingly.
The lengths these AI mods are increasingly going towards are getting incredibly concerning, and it's only a matter of time before someone with actual malicious intent realizes they can just ask AI to write some niche malware, post here, and infect actual people.
I worry that we never learned from the Fractureiser malware incident 3 years ago.
https://redd.it/1t4dyvx
@MinecraftModded
Reddit
From the feedthebeast community on Reddit
Explore this post and more from the feedthebeast community