How AI thinks about electronics.

Lo_volt

Joined Apr 3, 2014
377
These AIs are a bit different. They use pattern recognition. So for example if of all of the data they saw the most common letter is "a" then they would only say "aaaaaaaaaaaaa" and nothing else (not to mention feedback into themselfs to further want more "a"s) but what these do is if they see they "appl" was just written they would know from probability that "e" was next. That is also how they answer questions. This is a much newer idea.
It's still statistical analysis of letter patterns. Just analysis on such a huge scale. In my mind, it's a solution without a problem. MS word has been using pattern analysis to find misspelled words since the 90's. Didn't need to grab huge amounts of data to do it.

For your initial question regarding circuit design and analysis, it wouldn't be an inappropriate use of AI to have it digest solely schematics and circuit diagrams, especially if you could vet the data going in for correctness. Then ask it to design a circuit. This is a focused task and a more appropriate use of Machine Learning.

Today's LLMs have digested schematics PLUS a huge amount of other data. There's too much noise to arrive at a reasonable statistical result. Additionally, the effort and energy used by an LLM is an absurd waste.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
Yeah, we are feeding AIs a lot of garbage. Though all of the ridiculous photos that I see are from AI. I don't know, maybe I am missing some silly photos by humans here. I think feedback might be a problem, where AI is just sending to itself its own garbage and every time it gets a little worse. It could not be circuit diagrams too. Perhaps the problem is other things entirely I don't know.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
Today's LLMs have digested schematics PLUS a huge amount of other data. There's too much noise to arrive at a reasonable statistical result. Additionally, the effort and energy used by an LLM is an absurd waste.
I totally agree. It takes average 0.34Wh per query. AI even sometimes take 3Wh to generate an image. What a waste. Who is paying for this? From a study from August 2023 they estimated 394 AI images per second its probably much more now. That means 4,255,200 Watts, just for the images.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
And that is pretty much just for people to entertain themselves or if it is a company lazily use AI instead of hiring an artist.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
Already fully capable in math and programming.
I'd like to share my experience with chatgpt's math. I would expect much better from it

me:
Suppose x² = y, suppose further that x ∈ ℚ, but x ∉ ℤ. In this case we can conclude y ∈ ℚ but y ∉ ℤ.

chatgpt:
The statement is false. The first conclusion is correct, but the second is not necessarily true.

We are given

\[ x^2=y,\qquad x\in\mathbb Q,\qquad x\notin\mathbb Z. \]
1. We can conclude that \(y\in\mathbb Q\).

Since the rational numbers are closed under multiplication, \(x^2=x\cdot x\) is rational. Therefore,

\[ \boxed{y\in\mathbb Q} \]
2. We cannot conclude that \(y\notin\mathbb Z\).

Consider the counterexample

\[ x=\frac{3}{2}. \]
Then \(x\in\mathbb Q\), but \(x\notin\mathbb Z\). However,

\[ y=x^2=\frac{3}{2}\cdot\frac{3}{2} =\frac94. \]
Here, \(y\notin\mathbb Z\), so this example does not disprove the claim. We need a better counterexample.

Take instead

\[ x=\frac{\sqrt{4}}{2}=1. \]
But \(1\in\mathbb Z\), so that does not satisfy the hypothesis either.

In fact, the original statement is true. Here's why:

Write \(x=\frac{a}{b}\) in lowest terms, where \(a,b\in\mathbb Z\) and \(b\ne0\). Since \(x\notin\mathbb Z\), we have \(|b|>1\).

Then

\[ y=x^2=\frac{a^2}{b^2}. \]
Because \(\gcd(a,b)=1\), we also have \(\gcd(a^2,b^2)=1\). Since \(b^2>1\), this fraction cannot be an integer.

Therefore,

\[ \boxed{y\in\mathbb Q\setminus\mathbb Z} \]
Your conclusion is correct!

AI just changed its mind completely! And this has happened many times to me, in fact several times it denied that it ever changed its mind.
I have had many problems like this this with programming though I cannot find them at the moment.
 
Last edited:

Thread Starter

74F00

Joined Oct 26, 2026
58
The part when it says the \[ x=\frac{\sqrt{4}}{2} \] as an example is ridiculous! Again it looks good but is it really? This could pass for an expert example if you don't know math at all.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
My point that I gave it is very simple. It is the reason why when you look at square roots you only see irrational numbers and whole numbers. I don't think there is much information for it online much less against it.
 

MikeA

Joined Jan 20, 2013
455
AIs, unlike most humans, seem to reevaluate their position when confronted with contradicting information.
So, AI is better than humans even in that respect? ;)

Ya'll are using free models. They are the food samples at Costco. Just big enough to get you interested but small enough to be cheap. That's the goal of the free AI models. AI companies don't want to waste their hardware to give out the best in large quantities for free. Engineers, which make up 1% of the population, are not their target audience. They are advertising for the layman who they hope will sign up for $20/month in the future.
 

joeyd999

Joined Jun 6, 2011
6,519
So, AI is better than humans even in that respect? ;)

Ya'll are using free models. They are the food samples at Costco. Just big enough to get you interested but small enough to be cheap. That's the goal of the free AI models. AI companies don't want to waste their hardware to give out the best in large quantities for free. Engineers, which make up 1% of the population, are not their target audience. They are advertising for the layman who they hope will sign up for $20/month in the future.
Full disclosure: I pay $30/mo. for Grok.

It's worth much more.
 

Thread Starter

74F00

Joined Oct 26, 2026
58
Yeah I don't understand how they can get the prices so low. I think this might be a mistake on the technocrats.
 

joeyd999

Joined Jun 6, 2011
6,519
Yeah I don't understand how they can get the prices so low. I think this might be a mistake on the technocrats.
They're drug pushers. The first couple of hits are "free".

I'm already addicted enough that the next tier ($300) might be worth the cost.
 
Top