Are you using AI?

Certainly not using AI.

I keep reading about how wonderful it is and how it will design all my circuits, write all my code etc.

And then I try it (again) ........

Last attempt was the AI help engine available with Maple (paid for software for doing Maths).

It (like always with AI) wrote  a nice looking chunk of code that didn't work.

It was moderately useful as a help engine for explaining commands but not very good.

I read about wonderful results but never see any of them published - I wonder why.

MK

So far AI is hopeless at designing circuits.

In what way. I'm curious.

I recall this glib remark which has stuck with me... The quality of the answer starts with the quality of the question.

AI certainly won't do the work for you (as in, produce a wonderful set of gerbers), but it does make design decision making quicker, especially if you view AI as a sounding board. E.g. I needed to include an microSD card holder on a PCB. It's not something I would do often. I couldn't quite remember how the card detect pin works for the part I wanted to use as it usually differs depending on the manufacturer. Choose an unknown oem part and risk escalates. Asking AI saved me 10's of minutes searching for the answer. As such, it's an accelerator.

AI can do lots of things but they don't shorten design time or improve designs - yet.

AI can select components, but it takes a lot of work to get it to select an optimal component and then how can it prove it did a good job without showing you hundreds of datasheets. Sure it can organize the data sheets, with enough coaching, but a parametric search engine is going to compete favorably with AI at this task.

AI can generate a netlist, but how can it prove it is a good design? Do you have to translate the net list to a schematic and PCB layout and build the card and test it to know if it was hallucinating? Verifying the design from a netlist introduces far more problems than it solves.

I haven't seen  anyone show that AI is good at designing circuits. Presumably it will happen at some point. 

Ah, but is that not automation - you tell it do this and it does it. In my mind, I don't think that's what AI is about. What it can or what it should probably do is give you the components, or a short list of components within spec, to achieve the result you're after. It can certainly tell you which pin to connect to what component etc.

AI cannot draw a "proper" schematic of even a simple circuit, and certainly it cannot design a PCB layout. Even if you supply datasheets and reference designs, it cannot copy the symbols properly.

Try asking your AI to draw a schematic of an inverting operational amplifier with a gain of 10.

It can sort of describe some simple circuits with text in a net list type of way, but that doesn't help in my workflow.

AI is improving dramatically every week, but the last time I checked, it had not been trained on schematic capture.

AI cannot draw a "proper" schematic of even a simple circuit, and certainly it cannot design a PCB layout. Even if you supply datasheets and reference designs, it cannot copy the symbols properly.

Try asking your AI to draw a schematic of an inverting operational amplifier with a gain of 10.

It can sort of describe some simple circuits with text in a net list type of way, but that doesn't help in my workflow.

AI is improving dramatically every week, but the last time I checked, it had not been trained on schematic capture.

AI can do lots of things but they don't shorten design time or improve designs - yet.

AI can select components, but it takes a lot of work to get it to select an optimal component and then how can it prove it did a good job without showing you hundreds of datasheets. Sure it can organize the data sheets, with enough coaching, but a parametric search engine is going to compete favorably with AI at this task.

AI can generate a netlist, but how can it prove it is a good design? Do you have to translate the net list to a schematic and PCB layout and build the card and test it to know if it was hallucinating? Verifying the design from a netlist introduces far more problems than it solves.

I haven't seen  anyone show that AI is good at designing circuits. Presumably it will happen at some point. 

Ah, but is that not automation - you tell it do this and it does it. In my mind, I don't think that's what AI is about. What it can or what it should probably do is give you the components, or a short list of components within spec, to achieve the result you're after. It can certainly tell you which pin to connect to what component etc.