The h Bar

PM 2Ring

07:10

This guy physics.stackexchange.com/a/511376/123208 has a classical theory that replaces relativity & quantum theory: "Energy Field Theory". But he's not your usual crank: he apparently has a science degree, and his articles don't have obvious flaws or the common tell-tale features of typical crank work, apart from claiming that relativity is WRONG. (Yes, I flagged the answer as promoting a non-mainstream theory).

yuvraj singh

07:36

@PM2Ring yesterday you guide me this I have doubt which I already mentioned in the question. astronomy.stackexchange.com

PM 2Ring

07:47

@yuvrajsingh I think you're talking about the edit you made to this question. That's actually a very important question in relativity theory.

The name of this topic is "the relativity of simultaneity". There are many questions on this topic here on the Physics site. To understand this, first you need to understand the 3 different types of separation that are possible in spacetime: timelike, lightlike, and spacelike.

Relativity of simultaneity

In physics, the relativity of simultaneity is the concept that distant simultaneity – whether two spatially separated events occur at the same time – is not absolute, but depends on the observer's reference frame. == Description == According to Einstein's special theory of relativity, it is impossible to say in an absolute sense that two distinct events occur at the same time if those events are separated in space. If one reference frame assigns precisely the same time to two events that are at different points in space, a reference frame that is moving relative to the first will generally assign...

yuvraj singh

OK. @PM2Ring

PM 2Ring

Have a look at that Wikipedia article. I like the animated diagram near the top of the article. I've been looking for a good page here on Physics, but I can't find one that is aimed at beginners.

It may help to read this first: en.wikipedia.org/wiki/Spacetime#Spacetime_interval

yuvraj singh

Can you give some, so that I can read this, and then I can come. Up with my doubt. @PM2Ring

PM 2Ring

08:29

@yuvrajsingh Sorry, I still can't find a suitable page on Physics.SE. But take a look at this page by John Norton: pitt.edu/~jdnorton/teaching/HPS_0410/chapters/…

Novalium Company

11:46

I'm planning to code the first general AI on my PC during the weekend.

Something cute to keep me busy.

PM 2Ring

12:27

@NovaliumCompany Good luck. Please make sure it's friendly. I don't want to get turned into paperclips; at least, not for a few more years.

Abhas Kumar Sinha

14:16

@PM2Ring hehehehe Paperclips. I got it.

@NovaliumCompany Cool, what you'll code? deep nets or ML related something like classifier/hopfield or what?

@PM2Ring AI talking over humans is BS, humans are always above all.

@PM2Ring if someone thinks that AI is about to overtake entire humanity, just unplug the computer cable. THAT'S GENIUS!

JMac

@AbhasKumarSinha ... That's the kind of arrogant thinking that gets us enslaved by computers. The actions that a super intelligent entity might take to prevent itself from being unplugged could be beyond our recognition. By the time we notice something is wrong it could be too late.

Abhas Kumar Sinha

@JMac Not true, computers are just doing for what they were programmed for, they can never enslave any one or a small mouse without one's permission, that's funny that people think like that. Probably, they don't have much exposure to AI. AI is something that is as profound as water or electricity in the modern-day world. I'm waiting for the day when the AI revolution will come and people will realize that was all for good.

We should stop making movies like the terminator/robot, they ain't doing good.

JMac

@AbhasKumarSinha It's funny that you assert that AI could never do that without one's permission. The paperclip example is the perfect thought experiment to explain it. It's not uncommon for humans to make something more capable than they are at some task; only to have that thing cause unexpected damage to some other area.

Abhas Kumar Sinha

14:36

@JMac The universe is meant to change, resistance to change will hit harder than change itself if you aren't ready for that. Paperclip maximizer is pure nonsense. Please stay away from myths and superstitions. AI would never take over humanity because it won't get pleasure by doing so. AI is nothing but a branch of mathematics that takes data and processes it using a function to get output in the desired set/forms. No big deal.

@JMac For example what it is called Bayesian probability in mathematics, is bayesian classifier in AI, they take the experience from past data (event, outcome) and produce most probable outcome, more they do it, more set of input data, more perfect output. (specific example, how things are in AI)

PM 2Ring

@AbhasKumarSinha Maybe, but we can't be sure of that. People have been working on AI for decades, and it's still very primitive. OTOH, AI can now do some pretty impressive stuff that computers couldn't do 20 years ago. Who knows what future AI will be capable of?

Abhas Kumar Sinha

@PM2Ring Paperclip maximizer can be stated in another way. "Suppose you have an AI that plays chess against humans. It gets a reward each time it wins against human players. AI realizes that the lesser the players are, there are more chances that it'll win the competition. So, it kills all the humans in a single go."

^that doesn't happen, even with very complex deepnets like Google DeepMind which is far complicated and better deep net than any other in the world now. It played the game like a normal human being without killing humans. LOL XD

JMac

14:54

@AbhasKumarSinha You're projecting way too much on my comments. The point of paperclip maximizer isn't that "AI will kill all humans, we shouldn't use AI". That's not what I'm saying, and I'm not against progress in AI. But we have to be very careful about implementation as we give AI processes more sway over physical processes. If AI doesn't understand human values and we don't fully understand the AI, it could lead to unexpected and bad consequences.

PM 2Ring

@AbhasKumarSinha It might not be easy to simply "pull the plug". What do you think would happen if (for example) Google announced tomorrow that a powerful AI had escaped, and was roaming the internet, and that the world was in danger, so the entire internet had to be turned off? Would people actually turn off every internet server? I don't think so, unless it was really obvious that the rogue AI was doing bad stuff.

Abhas Kumar Sinha

Powerful AI escaped and roaming internet...?

PM 2Ring

On the topic of containing a powerful AI, please see wiki.lesswrong.com/wiki/AI_boxing

Abhas Kumar Sinha

You are right...

I don't know what AI is capable of doing tomorrow, so unless it happens, it'll take over whole humanity and will not too! (Schrodinger's Cat's Logic)

PM 2Ring

You may notice that I'm posting lots of links to LessWrong. Eliezer Yudkowsky has his faults & biases, but he has spent a lot of time thinking about AI, and discussing it with many experts & AI enthusiasts. It's worthwhile browsing the LessWrong archives on Artificial Intelligence just to get up to speed on what those people have thought & said.

Note that AI systems do not need to be sentient in order to have a powerful influence over humanity. To an extent, they already do, because AI tech is now a significant factor in trading on the stock market.

I assume that big media organizations are starting to use Deep Learning to analyse news and trends on social media. Which means those AI systems are starting to influence what gets reported. In a few years, it may be very hard to tell real news from fake news created by AI systems...

@AbhasKumarSinha I agree that stuff like Terminator is a bit silly (and probably totally impossible, due to the time travel stuff). I don't believe that we'll create an AI monster that wants to wipe out humans. But we might create an AI that has its own goals, and which considers humans as mostly irrelevant, and somewhat annoying.

JMac

15:27

@PM2Ring It's things like subtle manipulation that I would be worried about from the AI, not so much mass genocide.

yuvraj singh

@PM2Ring I ask you a question. Can I say about that A completely empty universe is static i.e. unchanging in time. So it makes no difference where we place the origin. Mathematically we can arbitrarily choose a point as the origin and that's a perfectly fine thing to do. The universe is a 4D manifold and we can choose any random 4D poitn and say "that's our origin".

Is there any false in it.

PM 2Ring

@yuvrajsingh That sounds perfectly correct to me.

@JMac Exactly. And it doesn't need to be all that subtle. Just consider the profound affect that media organizations like Disney have had on the world.

yuvraj singh

@PM2Ring. But then again a similar question comes in my mind.

Aaron Stevens

16:30

@PM2Ring Ugh I am so tired of seeing huge Frozen 2 displays at Target.

Related to the above discussion, just went to a talk yesterday by Shriram Ramanathan from Purdue University. His group is doing some cool work in using various materials that exhibit properties similar to neural plasticity at various temporal scales. This is interesting because one of the key requirements of learning, memory, etc. in biology is plasticity at multiple scales.

His work is at the "proof of concept" stage in terms of how these materials exhibit these behaviors, but it was an interesting talk.

Abstract of the talk

The survival instinct is ubiquitous across organisms. Intelligence and cognitive capability however are correlated with the complexity of the nervous system, e.g. neuronal diversity and is enhanced by transfer of evolutionary knowledge and lifelong learning. We will consider examples where such organismic behavior can be realized in programmable quantum materials.

Using the metastable perovskite nickelates as a model system, we will discuss insulator-metal transitions that are controlled by hydrogen doping highlighting recent discoveries on electronic structure modulation approaching the predictions of Verwey and Mott.

From understanding the binding of the charge carriers to the lattice, we will describe two experiments in nickelates that mimic classic observations in biology, namely habituation in the sea slug Aplysia and ancestral intelligence in sharks. We will conclude with suggestions for future research to design AI machines that can exploit strong Coulomb interactions in ionic lattices.

JMac

@AaronStevens Like how water has memory! /s

Aaron Stevens

@JMac I see the "s/", but what are you referring to here?

JMac

@AaronStevens homeopathic medicine. en.wikipedia.org/wiki/Water_memory

Aaron Stevens

In relation to that, slime molds that have no brain still have the ability to remember and predict events basically using biological oscillators

@JMac Interesting. Never heard of that

In the above talk he kept saying, "mathematical models can be used to explain these things," but he never went into the modeling, which is the part of biophysics I find the most interesting. oh well haha

JMac

@AaronStevens Yeah it's one of the new ones. They dilute chemicals with water... so many times that it's statistically likely there isn't even a single molecule of the chemical left... but it's supposed to work because the water "remembers" the chemical frequency or some BS. definitely not the same as that talk

Aaron Stevens

16:42

@JMac Yeah weird. That talk somewhat got me thinking about what one could define "memory" to be though

Like for the slime mold example, the memory/prediction aspect seemed to just arise from the establishment of some sort of oscillation process.

The experiment he referred to they would change the environment of the slime mold periodically and the slime mold would react

And then they would stop the period changing of the enrvironment but the slime mold still reacted at the time where the change would have occurred

So it really was just like training an oscillator I suppose, but I can see how one could argue that this is like a form of memory

He made the comparison of how the Egyptians survived by understanding the periodic flooding of the Nile

JMac

I think it's very exciting, and possibly existential crisis inducing, to know we're figuring out things about how thinking actually manifests

PM 2Ring

@AaronStevens It's certainly worth investigating. It might be a lot more efficient to build neural networks like that than to implement them with standard logic circuitry. Of course, according to Turing, Church et al, all general computers can compute the same stuff, but that certainly doesn't mean that all computer designs have the same efficiency.

Aaron Stevens

@PM2Ring Right. The middle of the talk was more materials science focused, which isn't my area of expertise. But I think he was saying how the materials he was working with that involved Hydrogen doping had better properties for this type of stuff then the silicon materials that are really well-understood and being used today

But yeah, it was a different way of thinking about it

Essentially mimicking the processes of the brain at a smaller level to build up AI that can learn, remember, etc.

Rather than, as you say, using standard logic circuitry

At least I think that is what he was getting at haha

PM 2Ring

There are single-celled creatures with very short lifespans that live in the ocean. They need to use different enzymes at different times of the day, due to differences in temperature & light. The enzyme production is controlled by a biological clock. Some of these organisms only live for a couple of hours, so they inherit the current clock setting from their parent, which is pretty cool.

Aaron Stevens

@PM2Ring Yeah it's pretty astounding. Oscillations are so important in biology. Which is why one of the recent Nobel prizes was given to the studies of Circadian rhythms in fruit flies I believe

@JMac Yeah... I start thinking about it and then I stop haha

PM 2Ring

17:07

A few months ago there were some conversations between enumaris & b degnan about neuromorphic chips. That's not as radical as what Ramanathan is talking about, but it is rather different to the usual bog-standard logic chips.

Novalium Company

18:10

I'm actually noob at programming AIs. I need to look into that. I've only done simple linear regressions and supervised learning based on label data but with libraries.

(That's ML, though. Neural Nets are an interesting start. I'll look into TensorFlow)

Emilio Pisanty

19:12

Los Panchos - Perfidia

►

@DanielSank currently playing at the restaurant I'm at

Novalium Company

20:26

I'm confused. If I was to stay on Earth my whole life, would I live more than if I was to spend my life on the ISS?

or I would live equally long for myself but not equally long relative to my peers on Earth or ISS?

If aliens are coming to Earth at 99% c, then we shouldn't expect them soon, since they'll be travelling super slow from our point of view. Maybe that's why they haven't come yet.

Is there an equation that tells me at what velocity others are perceiving me to travel if I was travelling at let's say 87% of c?

Loong

20:44

Heh. If you stay your whole life on ISS at a dose rate of about 20 µSv/h, you will probably not really live longer than people on Earth. ;-)

PM 2Ring

21:12

@EmilioPisanty Rita Payés from Barcelona on trombone & vocals, with her mum on guitar.

Rita Payés & Elisabeth Roma - Melodia Sentimental

►

PM 2Ring

21:54

@NovaliumCompany If you're heading towards Earth with a constant speed of .87c relative to Earth, then in your frame you are at rest, and Earth is moving towards you at .87c. At that speed the Lorentz factor is about 2, so if the Earth observer measures your ship to be at a distance of 1 light-hour, you will measure the distance to be 0.5 light-hours. And if your ship is 100 m long (in your rest frame), Earth will observe it be 50 m long

Also, you will each measure the other's clock to be running at half the speed. Yes, that sounds crazy, but it works. ;)

Sir Cumference

22:44

@NovaliumCompany No worries :)

Jake Rose

23:49

@NovaliumCompany Check out the lorentz transforms

Transcript for