Lan Todak

22 hours ago, swansont said:

We’re not denying its presence, but we can opt not to use it in certain situations, like where it’s unreliable or unethical.

I don’t see how that would comply with the rules. We want the thoughts of people. If we have questions or critiques we want the person whose idea it is to engage with us

Ok, I got it. When I am free, I will start fresh topic with no AI. 👍👍👍

On 12/22/2025 at 3:20 AM, Phi for All said:

Like other sites, we're adapting to the sudden surge in AI generated content. We think it can be used as a LLM for those who can't always put their thoughts down as eloquently as they'd like. Many people, however, use it to draw conclusions, search for evidence, and site their sources, all of which we've proven the AI is incapable of doing honestly.

Even this small interaction is much more valuable to me than anything you might do with AI.

I have to admit that I'm heavily biased against AI in general. In the US, preemptive laws are in place that prohibit us from regulating AI companies. I've heard that AI in 2026 will use more water than all the bottled water companies in the US. And it just pisses me off in general since it's obvious the billionaire class hopes AI will eliminate the need to employ actual people. To me, AI is like a fascinating toy that kills people by making them useless.

This is an unavoidable situation. We can't deny the presence of AI companies. It's up to each country's regulations to protect citizens from them taking our jobs. I don't really know how to handle this situation, but I hope my country will do what's best for our people.

16 hours ago, Sensei said:

The problem with LLM is that there are many models, even under the same name, with different computing power and different capabilities.

Let me give you an example from last week.

I launched ChatGPT and asked it to convert a piece of C/C++ code (brute-force calculation) into a mathematical function. It was the following piece of code:

#include <stdio.h>
#include <stdlib.h>

int calc_total( int digits ) {
	int result = 10;
	while( digits-- > 1 ) {
		result *= 10;
	}
	return( result );
}

int extract_digit( int value, int digit ) {
	int result = value;
	//printf( "value %d digit %d ", value, digit );
	while( digit-- > 0 ) {
		result /= 10;
	}
	//printf( "result %d\n", result % 10 );
	return( result % 10 );
}

int calc_checksum( int value, int digits ) {
	int result = 0;
	for( int i = 0; i < digits; i++ ) {
		int digit = extract_digit( value, i );
		if( digit == 0 ) return( -1 );
		result += digit;
	}
	return( result % 10 );
}

int main( int argc, const char *argv[] ) {
	if( argc == 2 ) {
		int digits = atoi( argv[ 1 ] );
		if( digits >= 2 ) {
			int total = calc_total( digits );
			printf( "total possible: %d\n", total );
			int count = 0;
			for( int i = 0; i < total; i++ ) {
				if( calc_checksum( i, digits ) == 0 ) {
					count++;
				}
			}
			printf( "checksum possible: %d\n", count );
		}
	}
	return( 0 );
}

If you (reader, whoever you are) are an expert in mathematics/physics, stop reading right away the text below and try to solve this problem yourself as a mathematical puzzle.

If you find this difficult, imagine an LLM doing this.

The first ChatGPT told me f(n)=10^n/10

I answered, NO! you did not take into account that zero is ignored!

It told me: my mistake, first zero is ignored, and gave yet another function. With just first zero skipped (sigh!)

I answered, NO! all zeroes are ignored, not just first one!

It told me: my mistake, f(n)=9^n/10 is the right answer.

I answered, NO! For n=2, 9^2/10 = 81/10 = 8.1 which is a fraction! How can a fraction be the answer?! For n=2 the correct answer should be 9. For n=3, 72.

It agreed with me, and used... some rounding operator...

I lost patience, I asked: what version are you?

It answered: I am ChatGPT-2.

What? WHAT?!

I have never saw ChatGPT-2. How on earth it is here, when it normally runs ChatGPT v4, v4-mini, and v3.5 was the oldest one...

This time it completely locked up, and the only answer it could get was f(n)=9^n/10

It was impossible to skip it. Deadlock.

A day later, at night, when I was expecting ChatGPT servers will be less busy, I asked what version are you.

It told me: ChatGPT v4 with v5.2 engine. Let's test it.

And it gave the correct mathematical answer for my C/C++ algorithm. Which was:

The correct answer

f(n)=(9^n+9*(-1)^n)/10

Shock, it did it!

ps. And did you manage to do it yourself? I doubt it.

The moral of this story is that you have to be an expert in a given field to detect LLM's mistakes, because its answers are very credible, yet often wrong, and it cannot admit its mistakes unless they are pointed out to it directly and it cannot digest them itself. It is impossible to use it to come up with something completely new, such as new theories of physics.

Whether it answers will be highly prone to error depends on what you ask it, whether it's something trivial or something complicated. Asking it for help with the basics of computers carries a low risk of error (provided it's v4/v5).

You have to be very careful which version is started. Different versions of LLM have different window sizes (ask it about its window size and it will tell you). v2-v3.x have 4k tokens, v4 has 16k, v5 has 16-32k tokens. Once the window size is exceeded, it does not remember what was previously written to it during the same session. The more you talk, e.g., for hours, the less it knows what you wrote at the beginning and loses context. And the chance of making critical mistakes increases significantly.

Receiving code and writing code consumes tokens very quickly, so it will soon start writing nonsense. A few hundred lines and you're already outside the window size.

That's not how I usually work with AI. I know they make frequent mistakes, but that doesn't mean we can't use their data. Ask them to define something and then ask them to analyze other things related to that definition. Do they fit? For example: you ask them to define a pulsar, and then you give them several celestial objects for verification. Can they recognize which one fits the definition? If they successfully execute the task, you're done. Unless they fail, you can start feeding them logical contradictions between their data and yours. It works. If it doesn't, that's your problem, not theirs. I often use logical fallacies and contradictions to counter AI's reasoning by forcing them to generate a premise for each definition they've created. After that, I give them trick questions. This is where logical inconsistency happens. AI can sometimes make mistakes, but not always.

On 12/22/2025 at 5:01 AM, swansont said:

You should also see lots of moderator notes telling them it’s against the rules, if it was used to make content, and lots of such posts in the trash.

Read the rules. 2.13, in particular

Can I just post the summary of my discussion? I don't think this is prohibited too.

14 hours ago, swansont said:

It will be moved to the trash, as such content is against the rules.

I saw many people post a.i contents here. How is such content against the rules?

14 hours ago, Phi for All said:

Can you continue discussing this with AI by yourself? I'm more interested in what people who study science have to say about these things. We know the AI is more interested in pleasing you than it is using actual science. Even if it was short I wouldn't be interested. "Quite long" is probably a "framework" for obfuscation.

They will please you if you don't filter the contents. You can ask them to be unbiased towards your contents and preferences. They will give you direct answer, clean. Btw, I don't think I will continue 😁.

After discussing with a.is in several weeks, they finally agreed with me that gravity is independent of mass and eventually proves that core isn't dense like what people think. I will post some parts of it when I am not busy later. 👌. Be prepared because It's quite long.

On 9/2/2025 at 10:53 PM, swansont said:

Why do you think they would melt? How can there be one at the center? The universe has no center. Why do they have to be balls? Is the ability to move around important to the problem?

Can we increase gravity intensity by stacking up objects? If you are given a task to create a core of objects, how many stacked masses of objects required before gravity start crushing the masses to create core. Yes or no? Give me explanation.

On 9/2/2025 at 11:33 PM, Phi for All said:

This is science, so there is no "answer". An explanation is the best you'll get, but if you want one based on mainstream knowledge, you need to clarify what you mean.

Why can't science give answer. I don't understand. It's because science can't solve problems or science is more to philosophy?

3 hours ago, swansont said:

In what way is this a “brain teaser” or puzzle?

I want an answer. It's yes or no. So what is your answer and what's your explanation?

3 hours ago, exchemist said:

Perhaps you have to work out that noble price means Nobel Prize? 🤪

Ops, my mistake. It means noodle peace ✌️ 😁😁.

Schwarzschild metric has a flaw. If you arrange steel balls from one end to another end across universe, would the one at the center melt due to gravity?is it no or yes? If You get it right, you could probably win noble price.💪

1 hour ago, exchemist said:

Ballocks.

Bellocks

A tiny little creature, hiding in the cloud of heavy elements, surrounded by 7 layers of strong gravitational lensing effects(including earth GLE), within cold and dark area where light and temperature only cover up to Neptune

this is sun, a black hole.

Doomsday is near 😲😲😲.

3 hours ago, billiebn151 said:

Scientific progress is often seen as a linear path toward truth. Each discovery builds on the last, bringing us closer to a complete understanding of the universe. But is that really the case?

Philosophers of science, from Thomas Kuhn to Karl Popper, have debated whether science moves forward objectively or if it's shaped by shifting paradigms. Kuhn argued that scientific revolutions don’t just add knowledge but replace entire worldviews. Popper, on the other hand, believed that falsifiability—the ability to prove theories wrong—drives true progress.

But what if science isn’t leading to ultimate truth at all? Some argue that our theories are only approximations, useful but never final. The history of physics illustrates this: Newton’s laws were replaced by relativity, and quantum mechanics challenges even deeper assumptions. Will today’s “truths” be tomorrow’s misconceptions?

How do you see scientific progress? Is it a straight path to understanding, or are we simply refining models that will one day be replaced? Let’s discuss.

I don't know about others but fusion progress will eventually fail. Once they exceed threshold, longer chain reaction will happen. That will drain up energy. They will s when that happens.

Hi Everyone, could anyone here help me comfirm that this calculation verifies prime numbers?
This is the calculation:

DSum(Mod((x^2−A),(2B−2x))==0,x,0.0,(B−2))

A is a reminder of a number to be tested while B is a factor of a top closest square that is used to subtract the test number. For example,

If you want to test 7, the top closest square is 3^2 or 9. B is 3 as it is a factor of 3^2. By using the square to subtract the test number, 
3^2- 7=2, u get 2. Plug into the formula, u get

DSum(Mod((x^2−2),(3⋅2−2x))==0,x,0.0,1.0)

If you want to test 19, the top closest square is 5^2 or 25. B is 5 as it is a factor of 5^2. By using the square to subtract the test number, 
5^2- 19=6, u get 6. Plug into the formula, u get

DSum(Mod((x^2−6),(5⋅2−2x))==0,x,0.0,3.0)

It works for any prime number(big or small). If u do it rite u should get all false.

Tq...

I think I should post this in puzzle section instead of here. The answer is 1. You can get this answer by rearranging Fermat's little theorem. 

(I think my question wasn't interesting enough because it's easy)

8 hours ago, Sensei said:

Rainbow table does it.

 

Then write a computer program to show this.

 

I think you didn't get it. Should I say "any kind of machine calculation?". As I mentioned before, by looking at it, you could probably get the answer right away, with no machine involved (for example calculator).

Test it first. if nobody gets it, I will show what I've got.

8 hours ago, KJW said:

Is this connected to Fermat's little theorem?

 

I could say yes, but the X value is constant. It doesn't vary with a or p. Can you guess one of the composite numbers for 23^37-x. If X is 23, the composite numbers is 37, right? But you can't use 23 to other numbers. If a = X then yes.

Hi everyone... I have an interesting test to do. If given to you a^p, where p and a are available from 3 to infinity( although this set of numbers actually started from 0 but I removed them for some reasons), what is the special number(x) when you subtract from a, some composite numbers of a's reminder are always known? 

Here's an example. 
Pick any number for a,X and p. For this test I choose 8 for a, 11 for p and 31 for X so the structure looks like this, (8^11)-31=8,589,934,561. Can you guess the composite numbers for 8,589,934,561 without factorization? Some of its composite numbers are 13 and 660,764,197.
By using X =31, test X for every available number for a^p. Guess its composite numbers directly without using factorization.
If you are able to that, that means X is the special number. If u can't then it isn't.
Here are tips for you.
1. X value is always near one of the composite numbers ( that's why you can always guess them right away) within no more than 2 digits
For example, if X is 321, one of the composite numbers is within 300 to 350 width. More than that, you might calculate it wrong or X isn't special.
2. It works for any number for a^p. Yes, any number.
3. Use small numbers for test like 3^3-7 or 5^4-16 if you have issues with huge numbers

4. Remember to keep x as constant.only vary a  and p.

Good luck.

Hi everyone... Why Don't all of you talk about algorithm ( a set of instructions that allows a machine to be responsive). I mean why do AI look Like zombies to us although they are cooperative and responsive. What are the missing parts that weve overlooked. I think by discussing about this, allowing us to understand how unique free will is and why humans look Alive. 

I think it has something to do with yes-no process. You know what I mean.

I have something to share to all of you. As the title says, i will talk about earth's gravity growing with a proof. Don't worry, the proof I will provide can be tested in a lab just like you usually do, a very simple lab test.

// I have asked admin to repost the topic. I hope this is ok.

I have something to share to all of you. As the title says, i will talk about earth's gravity growing with a proof but I need at least 1000 likes for precaution. Don't worry, the proof I will provide can be tested in a lab just like you usually do, a very simple lab test.

On 6/2/2022 at 11:19 PM, Phi for All said:

What kind of charity are you talking about? Do you want us to say, hey, good for you for having an idea, don't worry about the parts that aren't supported by science, just keep having those great thoughts! Is that the charity you want?

A publisher needs to profit somewhere, so you either have to write something that people will pay the publisher for, or you have to pay the publisher yourself. The best way to get support from a publisher of scientific publications is to write some good science. They love that. 

Iirc, you don't support your ideas very well. Any science publisher is going to want to see the evidence we've been asking you for. What would you tell them?

On 6/2/2022 at 11:50 PM, swansont said:

Most serious science that is published does not generate income for the author(s). On the contrary, it costs money to get published in scientific journals.

Popular science writing can generate some modest income, however. But for good writing that lets non-scientific people understand scientific things.

So if you just want to get your ideas out there, it's as Bufofrog says - publish it here. But keep in mind that you will have a critical audience. If you don't want that, start a blog somewhere and turn off comments.

My phone screen is broken. I will try to do that soon. Anyway i Aman activist.

😃If I can't get charity from this site, can I get support from publisher? Is nyone here working on publishing company? I would like to publish my idea. If anyone could help me I will start a new topic right away 😃

Hi everyone...

18 hours ago, QuantumStrangeness said:

why is general relativity not compatible with quantum mechanics?. can they ever be compatible?.

Most of the time, when we understand one system or nature, we can create a simple and smooth math from it. But for an unknown, you might get wrong anticipation. 

For example, a machine produces an output using collatz conjecture. You ask some experts determine the system and create a math from it. What will you expect the results? 

They will create different maths. Oddly every math will work just fine on certain level but not on others. 

This is how we deal with gravity.

10 hours ago, Aman Uensis said:

I'm making the inquiry what is existence? Because I do think there is a very very subtle difference between the purpose/intent for which we were made (ie. God made us) and why we exist at all (which includes why anything including God exists at all).

That's Interesting inquiry but I believe in aliens or programmers over god or gods. I don't really care if they're powerful or not. 

Regarding your inquiry, you shouldn't take it lightly against gods in which you rely heavily on intuitions solely. Otherwise, it's a faith based opinion. Anyway, I still find discussing about existence is intriguing.

3 hours ago, Aman Uensis said:

To all those who do believe in a faith system, I'd like to know, would it bother you if your God was not exactly all powerful?

I'm asking out of curiosity because on a reddit forum recently I proposed an article that God does indeed have a limitation, however, as far as concerns the creation of humanity we could still regard God (if one believes in it) as the most powerful being in our immediate orbit of existence. So, would it bother anyone if this was the case? Or does your faith hinge on the belief that your God is absolute in all respects?

The link to the article is below for anyone who wishes to read in full:

https://medium.com/@thelittleblackbook/what-is-the-limitation-of-the-all-knowing-all-seeing-insert-belief-here-god-6068823c1a7

 

For those that don't want to read, basically I'm saying that the one limitation of God is that God cannot exist as anything other than God in essence or spirit. Not fully. In other words, existence imposes a condition on each and every thing (including God).

Do you mean god can only be himself, don't you?

In my sensei 😎

4 hours ago, studiot said:

But I could have written those numbers in a different order; the set would still have been the same.

It has no centre. That is the point.

It should be noted that it is also a representation if used to illustrate properties of the universe since it only shares some properties with whatever manifold the universe actually is.

The edges are 1 and 2 with 1.5 as the centre. Your set can't represent our universe. 😑

😜

On 8/20/2021 at 8:19 PM, Bufofrog said:

Please present evidence to support this.

 

I dont have... 🤐