LabVIEW
ChatGPT AI Assisted Programming with LabVIEW Discussion
@paul_a_cardinale wrote:
I think that "AI" is misnamed. I think is should be called 'Artificial Stupidity'
I think the issue is Labview, not the "Artificial Stupidity" for example:
Hello chat GPT, please make me a PWM controller in C:
Ok here you go:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#define PWM_PERIOD_US 20000 // PWM period in microseconds
void pwmInit() {
// Set up GPIO pins or any other hardware initialization
// For example, if you're using the Raspberry Pi GPIO:
// Initialize GPIO pin as output
// gpioSetMode(pinNumber, PI_OUTPUT);
}
void pwmWrite(int pin, float dutyCycle) {
int pulseWidth = (int)(PWM_PERIOD_US * dutyCycle);
// Set the pin high
// For example, if you're using the Raspberry Pi GPIO:
// gpioWrite(pin, PI_HIGH);
// Delay for the pulse width duration
usleep(pulseWidth);
// Set the pin low
// For example, if you're using the Raspberry Pi GPIO:
// gpioWrite(pin, PI_LOW);
// Delay for the remaining period
usleep(PWM_PERIOD_US - pulseWidth);
}
int main() {
pwmInit(); // Initialize PWM
// Example usage: Generate a PWM signal with a duty cycle of 50% on pin 18
int pwmPin = 18;
float dutyCycle = 0.5;
while (1) {
pwmWrite(pwmPin, dutyCycle);
}
return 0;
}
Hello Dall-E, please make me a PWM controller in Labview:
Response:
... lol : )
I think our jobs are safe for now.
@alexderjuengere wrote:
@Jay14159265 wrote:LAbviak, the spooderman version of LabVIEW
I just got the notification for this thread and didn't remember the context (and didn't scroll up), so I googled "LAbviak" to see what it was.
Maybe don't Google that one on your work computers... 😉
@Kathlynn wrote:
Yes, ChatGPT usually acts pretty sure of itself and might make up believable stuff instead of saying it doesn't know. So, I don't think programmers will be out of work anytime soon. But even if AI can write most of the code and messes up sometimes, you can still fix those mistakes and get things done faster.
I disagree. I think that it would be difficult to debug. No one wrote it, so no one is familiar with the code enough to know where to look if something goes sideways.
@Kathlynn wrote:
Yes, ChatGPT usually acts pretty sure of itself and might make up believable stuff instead of saying it doesn't know. So, I don't think programmers will be out of work anytime soon. But even if AI can write most of the code and messes up sometimes, you can still fix those mistakes and get things done faster.
Another problem with that assertion is that things like ChatGPT are often used by people who barely are able to write a "hello world" program in their programming language of choice. They have no possibility whatsoever to understand what the program really does and even less to spot inaccuracies or outright errors. Let's not talk about fixing them.
There are enough examples in other areas. AI fabricated legal cases that are used by professional lawyers in front of courts. Judges usually seem to actually check if those cases really exists, but once they use ChatGPT too for that research, things spiral totally downwards.
Before you know it, you have a self enforcing spiral of fabricated truth that is almost impossible to penetrate. Well we already have in a way without AI. It's proven that hearing a falsehood often enough will eventually make many people believe them despite actually knowing that it is not the truth. Thanks to social media, it's very easy to reach lots of people over and over again and in that way influence what is felt as actual truth by the masses.
