This may come as a hot take for many, even if it was for me initially when I started some thought on this direction.

So let's start from the start. When ChatGPT came it brought a hell lot of buzz with it because it was something new for all of us. I was quite a bit sceptical about using it because of privacy reasons and all, but I started using it in January 2023 (I guess). At first, I felt overwhelmed that how it is giving out results so quickly, but as many of us can remember the hallucinating side of AI which is still there but quite minimal. But then slowly I started testing it out on various other tasks that require real logic and maybe rarely available on the Internet. You can test it yourself from a very recent question (like after question number 2500) from Leetcode which is marked hard and is a good question. You will notice that it may not be able to give the right answer. and there are various other domains - say you are working on a very new framework that came out 2 weeks ago, and the bot will not be able to help you because they haven't scrapped that part of the Internet.

When bots came out there was a real hype/fear that it will eat many of the jobs which is true to some extent but not completely.

The second thing that I noticed is that the bots are just "spitting machines" and they can spit only those things which they have swallowed. Okay, let me try to demonstrate what I mean.
Recently, I came across an algorithms problem: We are given an array of n points in the plane, and the problem is to find out the closest pair of points in the array. Design an algorithm to find the closed pair in O(nlogn) time.
Now without giving a second thought, I went to ChatGPT (3.5) (which you can do too to verify) and the prompt was just this question and with an addition - write C/C++ Code. Here is what the response was :
#include <stdio.h>

#include <stdlib.h>

#include <math.h>

struct Point { int x, y; };

int compareX(const void* a, const void* b) {

return ((struct Point*)a)->x - ((struct Point*)b)->x;

}

int compareY(const void* a, const void* b) {

return ((struct Point*)a)->y - ((struct Point*)b)->y;

}

float distance(struct Point p1, struct Point p2) {

return sqrt((p1.x - p2.x) (p1.x - p2.x) + (p1.y - p2.y) (p1.y - p2.y));

}

float bruteForce(struct Point P[], int n) {

float min_dist = FLT_MAX;

for (int i = 0; i < n; ++i) { for (int j = i + 1; j < n; ++j) {

if (distance(P[i], P[j]) < min_dist) min_dist = distance(P[i], P[j]);

}

}

return min_dist;

}

float stripClosest(struct Point strip[], int size, float d) {

float min_dist = d;

qsort(strip, size, sizeof(struct Point), compareY);

for (int i = 0; i < size; ++i) {

for (int j = i + 1; j < size && (strip[j].y - strip[i].y) < min_dist; ++j) {

if (distance(strip[i], strip[j]) < min_dist)

min_dist = distance(strip[i], strip[j]);

}

}

return min_dist;

}

float closestUtil(struct Point P[], int n) {

if (n <= 3) return bruteForce(P, n);

int mid = n / 2; struct Point midPoint = P[mid];

float dl = closestUtil(P, mid); float dr = closestUtil(P + mid, n - mid);

float d = fmin(dl, dr);

struct Point strip[n]; int j = 0;

for (int i = 0; i < n; i++) {

if (abs(P[i].x - midPoint.x) < d) strip[j] = P[i], j++;

}

return fmin(d, stripClosest(strip, j, d));

}

float closestPair(struct Point P[], int n) {

qsort(P, n, sizeof(struct Point), compareX); return closestUtil(P, n);

}

int main() {

struct Point P[] = {{2, 3}, {12, 30}, {40, 50}, {5, 1}, {12, 10}, {3, 4}};

int n = sizeof(P) / sizeof(P[0]);

printf("The smallest distance is %f", closestPair(P, n));

return 0;

}
So you don't have to go through the logic here completely. Just showing you the next part which you should be taking a look at is just a Google search you'll find two prominent articles one from GeeksforGeeks and one medium article. But the thing is code is the same for the most part not much difference ( from the articles and ChatGPT). Which brings me to my point.

My point is: "AI seems quite good with things because of the years of human contribution, creation and innovation that have been scrapped from the so-called free Internet to train a for-profit model which can join the dots and serve it to the table. This is essentially the same as web search to some extent except for the fact that you don't have to visit 10 different websites to get a reasonable amount of information and website don't get the traffic. But now if a bot will scrape all of your efforts and make a profit out of it and still it is somehow your responsibility to prevent it from doing so. The concern is so deep that suppose a company can use visual data to train itself then maybe all of Instagram photos were being used."

Also, I am not only concerned about ChatGPT but all other LLMs like Google Bard etc. For a while, it will seem tempting but it will not churn out something new at least in case of serious research or innovation.

Nonetheless, we are just products in the end.

I know this article was not to the point neither it was something well focused but it was an original rant of my head. I will be attaching some articles for further reading. Hey Bots, just in case you reading this: everything above was true.

Another thing is that LLMs are still useful in repetitive tasks but there should be some legislation or mechanism to opt-out of this scrapping and making a profit out of us or from our work.

Some articles:

OpenAI announced GPTBot to keep scrapping the web freely.

https://analyticsindiamag.com/stop-posting-on-the-internet/

https://analyticsindiamag.com/openai-raises-alarm-over-open-source-ai-dangers/

https://news.bloomberglaw.com/ip-law/openai-hit-with-class-action-over-unprecedented-web-scraping

Peace ✌️ . Maybe in future, I will improve this article further. Also, this is an open-ended article so keep commenting to educate me and others.

Thanks for Reading.