If humans cant solve the halting problem does that mean we can be simulated by a computer

[Linus No Beard]

you can program a Turing machine to do anything calculatable and with enough training and time a human can do anything calculatable just food for thought 

[Vishera]

19 minutes ago, Linus No Beard said:

you can program a Turing machine to do anything calculatable and with enough training and time a human can do anything calculatable just food for thought 

As matter of fact humans are programmed,the rules and terms are written in our DNA.

But if we crack the code of our DNA and reverse engineer it - We will be able to program ourselves and change the limitations we have and the way we function.

[tikker]

25 minutes ago, Linus No Beard said:

you can program it to know it's limitations

But knowing a limitation doesn't equal understanding that limitation. We knew things fall if you let them go since the dawn of mankind, but we didn't understand why until the last few centuries. I think this is where we diverge. As far as I understand it a Turing machine gives a description on how to compute, not how to understand. Going with @Vishera's example of space simulations, if you put the known phyiscal laws into a Turing machine and let it go it'll simulate a universe with those rules and those rules only. It will never invent a new force and change the simulation, because it doesn't "know how", it just computes what it has been told to compute.

[Linus No Beard]

3 minutes ago, tikker said:

But knowing a limitation doesn't equal understanding that limitation. We knew things fall if you let them go since the dawn of mankind, but we didn't understand why until the last few centuries. I think this is where we diverge. As far as I understand it a Turing machine gives a description on how to compute, not how to understand. Going with @Vishera's example of space simulations, if you put the known phyiscal laws into a Turing machine and let it go it'll simulate a universe with those rules and those rules only. It will never invent a new force and change the simulation, because it doesn't "know how", it just computes what it has been told to compute.

but aren't we just following the programing of our dna

[Vishera]

Just now, Linus No Beard said:

but aren't we just following the programing of our dna

The difference is that we were programmed to learn,the space simulation was programmed to just follow specific rules of physics and nothing more.

The limitation is still the rules and terms of the software,just like we can program computers that can learn like machine learning technology.

[Linus No Beard]

1 minute ago, Vishera said:

The difference is that we were programmed to learn,the space simulation was programmed to just follow specific rules of physics and nothing more.

The limitation is still the rules and terms of the software,just like we can program computers that can learn like machine learning technology.

but we can program a Turing machine to learn what do you think machine learning is

[Vishera]

1 minute ago, Linus No Beard said:

but we can program a Turing machine to learn what do you think machine learning is

That's the point - If you want the machine to learn you have to program it to learn,It will follow what it was programmed to do.

[Linus No Beard]

Just now, Vishera said:

That's the point - If you want the machine to learn you have to program it to learn,It will follow what it was programmed to do.

and that is the point i am trying to get at with enough time and the right programing a Turing machine could fully simulate a human

[Vishera]

Just now, Linus No Beard said:

and that is the point i am trying to get at with enough time and the right programing a Turing machine could fully simulate a human

As long as humans don't understand the whole code of their DNA and how the body and brain work - humans won't be able to fully simulate a human.

With partial understanding of it we can only partially simulate a human like in video games.

[Linus No Beard]

6 minutes ago, Vishera said:

As long as humans don't understand the whole code of their DNA and how the body and brain work - humans won't be able to fully simulate a human.

With partial understanding of it we can only partially simulate a human like in video games.

yes which why i have been stressing WITH THE RIGHT PROGRRAMING 

[AbydosOne]

8 minutes ago, Linus No Beard said:

yes which why i have been stressing WITH THE RIGHT PROGRRAMING

Except you'd never know if it were the right programming because we don't know if the "is computer a human" program halts.

(semi-/s)

[WihGlah]

2 hours ago, Linus No Beard said:

we are talking about a theoretical machine with an unlimited tape and does not break down and does not succumb to entropy

Hence Logical Hubris.

 

 

[Sauron]

2 hours ago, tikker said:

Would a Turing machine know it can never resolve the halting problem?

No, because that would mean knowing whether the program will halt.

 

We also can't do that because it creates a paradox.

[Sauron]

3 hours ago, Linus No Beard said:

you can program it to know it's limitations

No, you can't. You can't program a Turing machine to know if, for an arbitrary program, the halting problem can be solved.

[tikker]

2 hours ago, Linus No Beard said:

but we can program a Turing machine to learn what do you think machine learning is

Machine learning isn't some magic that suddenly makes computers like humans. It's still following rules to find patterns or something and honestly advanced as the tech is, it's still not like an actual  human. We don't have a general AI yet either. An AI that plays Go will be best at playing Go. One that plays DOTA will be best at DOTA. Ask the former to play the latter and you'll get nowhere.

 

2 hours ago, Linus No Beard said:

yes which why i have been stressing WITH THE RIGHT PROGRRAMING 

It's not much of a discussion if you just go "we have the required compute power, correct understanding, right programming etc. so we know what to simulate". Of course the answer is yes in that scenario. The problem is that we don't know this yet. We don't know what the "right programming" is. We don't understand what a human is or what is required to simulate it. Some parts of humans will be easier to implement: if hungry then eat,  if thirsty then drink. Others not so much. When will the human decide to play videogames? When it gets bored? What decides when it gets bored, a random number generator? Our consciousness is still very much unknown.

 

3 minutes ago, Sauron said:

No, because that would mean knowing whether the program will halt.

 

We also can't do that because it creates a paradox.

Oh nice one. Didn't realise that. That's a nice proof that human != turing machine I guess then haha.

[Sauron]

Just now, tikker said:

Oh nice one. Didn't realise that. That's a nice proof that human != turing machine I guess then haha.

But humans can't do that either you can't know if the halting problem can be solved for any given program. Try going over Turing's original example and see if you can answer it as a human.

[tikker]

3 minutes ago, Sauron said:

But humans can't do that either you can't know if the halting problem can be solved for any given program. Try going over Turing's original example and see if you can answer it as a human.

Hmm, the intention of my original statement was more whether a Turing machine would even be able to somehow deduce(compute) that there is a halting problem. If it can't do that without creating a paradox, wouldn't us proving the problem to exist proof that we aren't a Turing machine? I'm probably missing some things. Time to dive deeper. Although we humans indeed also don't know if we'll find a proof and halt when we set out to proof something...

[Sauron]

Just now, tikker said:

Hmm, the intention of my original statement was more whether a Turing machine would even be able to somehow deduce(compute) that there is a halting problem. If it can't do that without creating a paradox, wouldn't us proving the problem to exist proof that we aren't a Turing machine? I'm probably missing some things. Time to dive deeper. Although we humans indeed also don't know if we'll find a proof and halt when we set out to proof something...

You might be able to write a program that detects some unsolvable instances of the halting problem, for example you could specifically look for Turing's example, but not for an arbitrary program. I don't know the math well enough to prove this to you myself but I found this. And more intuitively, you can see that if you want to determine if a halting problem can be solved for any program then on some programs that would involve waiting until they halt to check whether the oracle was correct, which of course is itself the halting problem.

 

This is a problem that is just mathematically impossible to solve; it's not a physical limitation of the Turing machine, it's just something we don't have a way of solving.

[sounds]

Part of why people get into heated arguments about this topic is that they think they'll become famous or something if they "crack the code" and prove that -

• We're all living in a simulation, or...
• The human brain is just a big turing machine, or...
• Turing machines are better than human brains, or...
• Something even weirder

If you take out the ego-centric bias, and start searching for actual discoveries related to consciousness, you start finding some really interesting stuff though.

 

For example:

 

https://people.duke.edu/~ng46/topics/evolved-radio.pdf

 

Summary:

 

They used real analog circuits. The result was that they got results impossible to get if they just "simulated the circuit."

 

The real analog circuit evolved a way to pick up nearby radio signals. The scientists had never programmed it to do this, it just "figured it out on its own."

 

The human brain figures stuff out on its own. It is likely something like what happened in this paper will lead to better insights into human consciousness.

[Invalid Human]

I personally dont think we are a game of sims.