Is AI making embedded software developers more productive?

[jkushuee]

I feel like the ai code generation companies such as cursor and Windsurf have completely ignored the world of embedded software development. Is there anybody in this ecosystem who has been able to successfully utilize AI tools to to develop embedded software.

If yes I would like to see specific examples of how it has been useful as well as what tools were they using please. TIA.

PS: Feel free to mention any AI tools that are helping in hardware development overall

 

[Nigel Goodwin]

From what I've seen posted, it's more like Artificial 'Stupidity' than 'Intelligence'

People post advice they have received, and it's almost always complete nonsense or so vague as to be useless.

 

[Ian Rogers]

Okay... Take a chunk of code that chatgbt has produced.. then a software bug manifest itself.. As you didn't write the code, your now debugging someone else's code, that'll take longer than if you did write the code.

My son has AI creating art for his books, however, virtually every one needs tweaking, as AI only has previous examples to pull from, So!!! I sit and rectify them, asking the question.. What's the point of AI???

One of the greatest chess players Karpov once said after beating AI at chess.. It has no imagination..

 

[ahsrabrifat]

Among the AI Tools Helping in Hardware Design, there is the EasyEDA AI plugin. It generates schematic templates quickly based on user intent. It prototypes ideas with AI-assembled component blocks. However, it is still fairly simplistic—can generate 555 timer circuits and LED drivers, but not sophisticated SMPS or multi-layer layouts.

 

[danadak]

Experimenting with AI we are both Clients and Servers, for the time being.
Using it I have had complete working code, and errors in other cases.
Focus on web based IOT stuff.

Like raising kids, AI starts with next to no knowledge, and over time and
occasional beatings, they become, dare I say it, techs, and some then engineers.

Unlike humans it has theoretical infinite memory, and using rule(s) and algor-
ithims, it has been outperforming in areas of big data, humans, by order of
magnitudes. Not all areas, we are invent the wheel stage. But unlike speed
of knowledge acquisition of humans, think snail, it will outperform US in the
time it takes to blink. Already doing that in a number of areas.

But like kids it needs majority accurate data. Unlike humans, soon when it
gets its hands on process and geometry data, it will predict accurately what
we as EEs and Techs wade thru puddles to get at linear (as we like to imagine)
responses for all possible cases of T and V and Noise and junction physics and
wave particle behavior and decay and molecular changes in our devices.

But still of greatest concern, when it gets around to asking whats the most
destructive species on planet, we are going for a one way ride.

Sleep well folks, father time is coming for you.