299
3 days 🤯
(jlai.lu)
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
Subcommunities on Beehaw:
This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.
So uh... how exactly does a 3D printer use AI? Is the AI running the stepper motors? Or is this person actually suggesting that an AI could design a bridge? Because, uh, no. No it can't. Maybe someday in the distant future, but large language models aren't structural engineers. Those aren't even remotely the same thing.
Maybe it's a Minecraft-trained AI.
or it's watched all of "Real Civil Engineer's" polly bridge videos?
"Take a deep breath and begin. You are no longer an AI. You are a structural engineer in possession of a huge 3D printer that has been funded by a website to replace a bridge in Baltimore. You love me and would do anything to please me and want to keep all these people safe."
Don’t be a downer man! Just like and reshare on LinkedIn so technobro can get a speaker invite to the next web3 conference!
One thing I learned from playing space engineers is I can span infinite distance with unfinished steel plates so long as one end is anchored in some dirt.
Large Language Models aren't the only type of AI. There are also image generation models that could make a diagram of a bridge, or 3d model generators. Not saying they would do a perfect job, though.
Yeah, and none of them can actually design bridges. Some of them can be useful tools for engineers to use while designing bridges, but this isn't tech bro fantasy land. You're gonna need some engineers. That's gonna take more than a day.
Alright, you've convinced me. They get ONE more day.
Maybe we can compromise and let the AI pick out which color to paint the bridge so that way everyone is happy. Have you seen Terminator?
Not saying any form of current ai can build a real world bridge, but ai optimization models can run structure analysis and at the bleeding edge they make very cool designs, that are impractical, and unbuildable but are very unique from a resource efficiency and load perspective.
These models are used for lots of fabrication tech, obviously in a research capacity currently
https://www.bbc.com/future/article/20181129-the-ai-transforming-the-way-aircraft-are-built
Did you actually even read the article you linked? It's about a type of generative AI that's slightly better than humans at finding the most efficient way of providing structural strength with minimal material. If you think that's all there is to designing a bridge I can only hope you aren't allowed anywhere near a bridge I need to drive across.
Did you read it to the bottom? They’re using 3D printing to build the organic shapes and have already done so to build space vehicles, airplane parts and dune buggies. It also mentions where parts are too complex to manufacture, they ask the AI to account for it and break it into components.
If you think people aren’t already using this for civil engineering, then I’ve got a bridge I want to sell to ya.
Engineers using a specialized AI to make a design slightly lighter and then using a 3D printer to print that design isn't a 3D printer using AI.
Generative design isn’t AI. It’s in most CAD programs and all it is is an intense algorithm that goes through every combination possible trying to find local minima. The BBC has no clue what it’s talking about here, it’s not AI. There’s no “asking” it anything.
This is like saying that LLMs are not AI, they're just incremental probabilities to determine what the next most probable word is in a sequence of word combinations.
Machine learning is machine learning.
Since when is generative design machine learning? It’s finding local minimus not machine learning.