Is electronic engineering dead?

I don't agree - while you see  a great many published designs executed in this way this is because their is little protectable (or saleable) IP embedded in them.

Real products made commercially still (often) use original hardware design to gain cost and performance advantage. But having invested a lot of money in a design it's not usually a good plan to publish it.

The picture shows an example (a current project):

There are two processors and an FPGA but there are filters ADCs, bipolar DACs, switching supplies, differential amplifiers etc. None of these are copies of reference designs (although obviously based on the huge body of analogue design knowledge available in books, on the internet etc.).

And while I can happily show you a picture of the pcb I'm not ready to share the detailed design.

There will be work for good hardware designers for a long time - and I suggest that such work is much more resistant to attack by AI  than any kind of software work.

MK

No way Jose.

We live in reality, not Tron. Nano tech can make a ton of building blocks, but never totally the infinity of creativity. 

Oh my goodness I hope not! I worked in medical research as an electronics development engineer and nearly every project required novel solutions.
I relished the challenge of working on demanding problems that often required unconventional approaches to achieve the final goal.
Developing equipment for use in a medical environment of course has the added constraints of stricter safety standards, and whilst (as our questioner points out) standard circuit elements are employed, but it's the design engineers job to use her/his skill pull these together to provide a satisfactory outcome.
That's what I got out of bed every morning for, the excitement of pitting my skills against what seemed an insurmountable electronic puzzle.
It's the creativity that made me love the job.

Daveb

Out of curiosity, how do you chose the ICs, chips... to use in such medical devices?
In most datasheet that I have read over the years, there is always a text saying "this device is not for human health, we won't be liable for any injuries/death..." even for the littlest component.

Well you have to use the chips that are available of course but its the careful development of the rest of the design that mitigates these safety issues. That's the challenge and skill of being a development engineer I guess. Yes I would lie in bed pondering these issues but still couldn't wait to get to work the next day. Pretty sad aren't I?

Daveb

I suspect that AI will be used as a productivity enhancement tool for EEs for a fairly long period before completely replacing EEs, long enough to have a good career if that's what's needed. However, the demand for EEs will also consequently decrease. One has to be flexible enough to learn not just the specifics of a particular technology as practiced today, but the overriding meta-rules that can be applied outside of one's primary educational focus.

AI will have customers whenever something is very expensive. NRE for small board runs is less likely to get heavy AI treatment than RE, management, and tasks that require extensive interaction with large portions of the environment that cannot be easily modeled.

Remember "computer" used to be a job title.

If you can't have a better product if everyone is implementing the same thing from datasheets, then it's a little hard for firms to innovate. Apple would cease to exist, and we would be buying phones from Alba or Amstrad.

Each time more performance/features are required, then you need to insert engineers.

You can tell that some student or person completely removed from the world of engineering work has asked this question.

It reminds me of one of my first jobs and my desk was piled high with ring binders full of ITU and ANSI docs, and a non-tech friend asked what it was all about. I mentioned that they explained how the phone system worked, so that I could design bits of it. The response I got was, why is all that needed to explain that you just need to lift the phone and dial?

There are whole TV comedy series' dedicated to this type of thing : )

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

RE: Is electronic engineering dead?

Poster

Upload Preview

Ok

Cancel

Electronic Engineering has been iterated in many ways.

One reason is the advancements in chip designs and availability of SoC, SIP etc that comes with everything that is needed. Consider a timer IC 555. we used to create timer and comparator circuits using this IC in the past but nowadays it is not that much required to do so as integrated chips such as microcontollers or processors will handle the tasks more efficiently. And there could be many such examples for us.

This does not necessary mean that "EE" is dead. But the number of jobs in circuit design and analog electronics should surely have reduced. 

I don't think it's dead. I think that the popularity of Arduino and other platforms made electronics a lot more accessible, and this has created a huge market for "ready-made" modules and prototyping platforms that are either code-based or sometimes even "zero code". As you mention it's easy now to build stuff just connecting modules together, and if you use Arduino or Raspberry Pi you will probably also find example code and libraries on the internet, so it's just connecting stuff and modifying some lines of code.

BUT, that's not how real products (i.e: devices expected to be mass produced, certified, and put on the shelves) are designed. Those are still built from the ground up.
I use modules to prototype and test, but that's never how I intend to build things.

And even if we were to move into a future where *everything* is just made of modules, someone has to design those modules : )

Regarding "reading the datasheet and copying the reference design". That has always been the starting point of any application, in my experience, but never the whole final product.

We are not yet at a point where electronic design can be done by an AI. It is true that there are a vast number of designs that have been developed already, and these can pretty much be duplicated without electronic engineers, but the number of new and unique electronic designs that have not been designed yet is a much larger number.

It is also true that electronic engineers will use reference designs and previously developed IP in new designs, but this just avoids having to reinvent the wheel and provides more time to spend developing new IP.

AI is developing at an accelerating rate so it is very hard to predict when it will become sentient enough to translate any need or desire into a complete solution.

The question of whether to specialize in hardware versus firmware is more about personal preferences - both are still in high demand and both may give way to AI eventually. I suspect one hardware engineer could keep many more than one software engineer busy so maybe more software engineers are needed.

The overarching question of deciding what the needs are, is another tricky topic. I doubt people will want to give these decisions away to an AI because it represents giving away power.

It may seem like designs published on the internet are mostly derivative, but that is because commercially important designs that have lots of new IP, would not be published on the internet. Many of them are too secret to even risk applying for patents.