Will AI Replace Antenna Engineers? A Deep Dive into AI's Role in Engineering

In this video, we explore the question: Will artificial intelligence replace antenna engineers? As an antenna engineer, I share my perspective on the current state and future of AI in our field. With the recent advancements from companies like Dell Technologies and Nvidia, there's a lot of buzz around AI and its potential impact on engineering.
Join me as I discuss:
The current capabilities and limitations of AI in antenna design
How tools like ChatGPT are being used in engineering
The importance of maintaining creativity and problem-solving skills as engineers
A fun experiment where I use ChatGPT to design a 900 MHz Yagi antenna
Don't miss out on this insightful discussion! Subscribe for more videos on antenna engineering, including upcoming topics on propagation properties in tunnels, helical and circular polarization, reflections on water, and the Magic Q RMS antenna radiation pattern system.
Timestamps:
0:00 - Introduction
1:30 - The AI Hype and Reality
4:45 - Recent AI Developments
6:15 - AI in Antenna Engineering
9:00 - ChatGPT Antenna Design Experiment
15:30 - Final Thoughts on AI in Engineering
17:00 - Upcoming Video Topics
Keywords:
AI in engineering, artificial intelligence, antenna engineering, ChatGPT, antenna design, AI vs human engineers, engineering creativity, Dell Technologies, Nvidia AI, Yagi antenna, antenna design tools, future of engineering, AI limitations, machine learning in engineering
Tags:
#ArtificialIntelligence #AntennaEngineering #AI #Engineering #ChatGPT #YagiAntenna #MachineLearning #DellTechnologies #Nvidia #EngineeringTools #FutureOfEngineering #AIvsHumans

[@Sam-cp6so]

remember: if a title or headline comes in the form of a question, especially yes/no form, the answer is next to always "no." and you can stop there.

[@flynx42]

My industry has been using antennas designed by genetic algorithms for awhile. It depends on what you consider to be "AI" but these antennas are mostly designed by the machine. However, it still requires a lot of human work to design the algorithm, the scoring routine, and the initial seed. Genetic algorithms only work if the solution space is continuous (any solution has to work, better or worse), antennas usually meet that requirement. en.m.wikipedia.org/wiki/Evolved_antenna

[@P-G-77]

WOW, wonderful objective... like the way made... interesting idea. Remain only one thing... HOW TO ASK CORRECTLY... TO DESCRIBE correctly... for sure we have to made a way, leaving IA to enter in the field of this objective... a discussion for example, searching for existing "papers", then ask to hypothesize, deduce, improvements... look at the results... then verify the results, then writing the results continue... in fact, in this way and only in this way I solved, and I built things and circuits that otherwise I would not have dreamed. I often find myself with solutions that maybe I had under my nose, simple, effective, but I didn’t see them... If then, you want to find a way to make an IA respond in "absurd" way well, it gets simple...

[@cagedgandalf3472]

I feel like it is a specific expertise and was probably not included in the dataset as much. It would be like asking a first or second year student to make one. Maybe if a company that had a lot of antenna engineers and lots of data of what the antenna engineers did. Then they train the AI on that data, maybe you would get a better response. But I still don't think they would be able to replace engineers as of now. Maybe in a few decades or so.

[@JNET_Reloaded]

ok you should of built it and then designed a (got ai) to make a better improved version to compare it to along with a couple standard ones and do tests that would be interesting!, but prompt engineering is key, you have to tell it what you expect of it, for example tell it you want Xn eliments or it wont know, its not a mind reader ...yet)!

[@SloeJuice]

If anyone rolled out AI for engineering, or specifically antenna engineering, it wouldn't be a general purpose chat bot. These 'AIs' can get quite specialised depending on how they're trained. Whether anyone will create such a tool or not will depend on economics of whether it's worth pooling resources into the creation of AI assistant tool for this niche target.
Assuming they do decide to do it, it would be great for those well established within the field, not so great for others.
The boring work has to raise the demand for antenna engineers. When you bring in extra automation, productivity increases as, for the sake of example, every 1 engineer can do the job of 2.
I don't know the economics of how much $$$ goes into engineer salaries when doing an antenna project, but if it's significant, then short-term you'd expect there to be more projects as every engineer would get the job done faster - meaning less money spent on salaries.
Long-term, if we're assuming the antenna demand stops growing (as enough are built such that any new ones are experiencing diminishing returns), then the salaries in this field will likely reduce for anyone but the most experienced engineers due to there being more applicants per position open than before. It would also get more difficult for the newcomers to squeeze into the field.
I wonder where would the replaced antenna engineers go? Would they go from the roles of doing the design to the roles of hands-on maintenance of existing antennas? Or are there other engineering fields to which the knowledge and skills they have would easily transfer?

[@FailTrainS]

You should do a bit of research into Physics Informed Neural Networks. Large language models are not meant to do scientific research. You can use the programming skills of a LLM to help you design some software to create antennas though. Eventually you'll probably have an agent swarm with varying strengths and weaknesses creating computer programs to solve difficult problems in design.

[@JNET_Reloaded]

talk wayh roo muh get 9n with it

[@BrianSheppard]

No. It will be an expensive plugin.

[@jooch_exe]

Engineers and programmers without talent will be replaced.The smart ones are always going to be ahead of AI, because AI is a tool and not a person with interests.

[@nandesu]

LLMs, no. They can’t do math. ML algorithms are already popular in this space. The LLM hype train needs to crash.

[@cheponis]

You're asking the image generator to make engineering drawings. This won't work. You are misusing AI.