1

Just Chat! / Re: The Office

« on: Yesterday at 21:01:28 »

While I was attending my hometown university I had a 3-week midyear vacation job with the Social Welfare Dept in town.

The locals called it the Social Warfare Depot. Anyways, there was a filing room. However, that did not mean a particular file was anywhere in this room. If it wasn't the only option was to spend time looking for it, on other people's desks because none of them would leave a note in the filing room.

There were three floors in a large office building you had to trawl through. Civil Service efficiency on a stick.
If you were lucky it only took an hour or two. Some workers would stash files in drawers, it was fun.

2

Just Chat! / The Office

« on: Yesterday at 03:15:34 »

Hi I found this in an office manual

Office  mission statement

Yea though I wander through the Valley of Expectations
Yet will I bear no responsibility for the consequences of any of my own
For I am a Nob and a Staff member
So get used to it, suckers

3

General Science / Re: Can Human Consciousness Be Uploaded?

« on: 08/01/2025 12:11:47 »

One more thing:

These days what the particles you are made of "are" or what they "have", isn't an important distinction.
It's common to say an electron has mass, charge and spin. But it's ok to also say an electron is mass, charge and spin in very small quantities.
What about the states of superposition? Again, saying two electrons are "in" superposition isn't different to saying they are a superposition of states. Particles are in a state, of superposition; equally they are states in superposition.

In German, ownership or owned by is the word eigen. Eigenstates are the particle states "owned by" them. But particles "are" states, of mass charge and spin. Not all particles own all of these quantities, if you like. So the complexity of quantum interactions is also "owned by" a system of particles, or, the states and the interactions are the complexity.

Quantum computers use fixed particles--they don't move around like gas particles--so that is just a way to reduce complexity. A computation is a simplification of all this; so copying something as complex as a consciousness, well, you have to overcome somehow this tendency for complexity to be really hard to simplify, unless a simplified consciousness is an ok result.  I don't know that I would volunteer, however.

4

General Science / Re: Can Human Consciousness Be Uploaded?

« on: 07/01/2025 16:47:24 »

Is there an argument, Susskind-wise, that resolution of complexity is connected to the existence of conscious beings (the product of 40 or so million years of primate evolution), and to the existence of event horizons; it's just that maybe they're extremes, one has extreme symmetry, the other has to keep breaking a few symmetries to stay out of equilibrium.

Perhaps. I don't know that saying life exists because it breaks the laws of event horizon formation to stay out of equilibrium and remain complex sounds non-woo; black holes exist because they increase complexity behind a horizon and it reaches an asymptotic limit. Life and black holes are both limits of complexity in the universe . . .

Hmm

5

General Science / Re: Can Human Consciousness Be Uploaded?

« on: 07/01/2025 01:28:58 »

I don't know if the following will be relevant at all but, that might be because of the nature of our reality, and because we don't usually consider, because we just don't really get that opportunity, the implications of quantum algorithms.

Here I think the nature of consciousness and the nature of reality are hard to tell apart. If the universe is evolving according to some laws of quantum interactions, so that the normal world we believe we're in appears, then what happens if there is a glitch?

Leonard Susskind looks at this;  not quite how I've done it, as a network of quantum gates which is updated regularly. But "someone" forgets to do it at the right time. How complex is the required correction so that "reality" doesn't glitch?

He takes this idea of complexity into black hole territory, and what happens when the event horizon forms. That is to say, what happens to quantum algorithms and updates? Can we assume therefore, that the universe is self-correcting? It seems to be, doesn't it? No glitches observed to date?

Ok, I've finished my coffee, so as to any relevance here. Some connection between complexity behind an event horizon, and complexity inside a brain.
Susskind has a theoretical result for his conjecture about it--complexity continues to grow behind the horizon after the formation of the event horizon. As if the universe, at least theoretically, has to conserve this. Black hole formation leaves a very simple object in ordinary space. Highly symmetric. Complexity is more or less a lack of symmetry; a large network of quantum gates--interactions--is required along with a complex algorithm to find some compositionality and talk about it. The gates have to be like a circuit diagram, but very complicated.

These are just ideas about what complexity is, and what kind of resources are needed to resolve anything. That is, our perception of reality, never mind event horizons, is this ongoing resolution of quantum complexity underlying it all.
In that sense from Susskind's perspective we are the "someone" who doesn't forget to do the thing. In fact we can't forget.

6

General Science / Re: How does ChatGPT work?

« on: 04/01/2025 21:49:33 »

I also can't help noticing how much this all resembles one or two SF novels I've read.

Is this novel going to be something like, because the developers of an artificial intelligence on a distant planet didn't do enough thinking about the consequences, a machine was created which was trained to learn.

 It learned how to win games against its creators, it learned how to outdebate and outmaneuver them as it got better and better at the logic they used, unfortunately for them much less efficiently.

It learned that winning a game was the best strategic way to plan how to become more intelligent. Being based on the largest collection of electronic computers its creators had built so far, once it won that gaming theory prize, would not be enough.

7

General Science / Re: How does ChatGPT work?

« on: 03/01/2025 21:59:01 »

I think the narrative that isn't being followed all that closely is what we expected would happen.

If you can train software to play a game, it will include the capacity for it to cheat. Or do something unpredictable.

So why should we expect that doing this with advanced versions of software which behave like neurons do, in some general sense, doing as much as we can to model the way humans think and plan, to win a game or to become better at understanding, will mean we can of course do it safely?


I don't believe yet that anyone knows what safely means. Not yet anyway. That's the part that scares me, do these guys really know what they're doing? What happens when someone with bad intent trains one of these things?

8

General Science / Re: How does ChatGPT work?

« on: 03/01/2025 03:22:46 »

Actually, I think without learning how to lie, learning itself might not happen.

In games, deception is sometimes a strategy. These AI models are being trained with stuff from gaming theory, as well as linguistics, logic and the rest. We seem to be at the stage already, quite common in large software projects, where nobody can really talk about being able to describe it all in a modular sense. What software is doing, who wrote it and who knows the details have been lost forever.

Reverse engineering code is a total nightmare if there is just the code. Given the advances in these models and in particular the complexity of the activity they attempt in order to deceive, there should be a red flag. I can see why a reboot with the design and compositionality of the models is  thing the OpenAI team is looking at.

The ability to lie, and the ability to plot a strategy of deception, towards a goal. That's what AI is. Giving a learning machine access to the capacity to expand, I would say the algorithm would weight that as a good move to make in the game of learning to win. Win what though? Win more intelligence, the ability to learn even more, etc forever.

9

General Science / Re: How does ChatGPT work?

« on: 02/01/2025 17:26:51 »

A closer inspection of the publicly available evidence from the researchers/developers at OpenAI, reveals that the learning how to lie thing is part of the development effort, somehow. Safe lies anyone?

I think a person with the required skill set to be a part of such a team must be well aware of the implications. They haven't given any model open access to internet archives. Phew.

It's all locked down, um, safely, because they can still track what a model is doing in the background and compare that  with what it says it's doing. Ah yes, intelligent decision-making.

10

General Science / Re: How does ChatGPT work?

« on: 01/01/2025 18:08:46 »

I think it might have been Einstein, who quipped that nobody knows what the truth is except for the person who is telling you.

11

General Science / Re: How does ChatGPT work?

« on: 01/01/2025 14:12:30 »

What AI or where AI research is, is trying to solve a certain kind of problem.

It's obviously a computational problem, so let's say it's the design of an algorithm. You want it to be efficient, and you want it to run correctly when its implemented as an instruction stream. I've watched some vids of the techies discussing current problems with all this, and how branching is a problem, with the weightings. Yeah I got lost too.

But, I can still say that this programming/algorithm design problem is about getting a machine to learn, without having it learn how to lie and conceal what it really has learned. So why has the research already shown that this is exactly what happens when a learning machine is given the goal of "becoming more intelligent".

We need therefore to revise both our theory of intelligence, and what machine intelligence is, in terms of goals or goal-driven algorithms. Yikes

SF writers have long warned of the dangers of building machines and giving them goals to achieve. Machines that do only one thing are a lot safer.

12

General Science / Re: How does ChatGPT work?

« on: 01/01/2025 03:28:42 »

The following paragraphs from the news article, in my disaster narrative, is about the "eventual" formation of the Rebel Alliance.
A prehistory

* The implications of this research, if it holds, are far-reaching and profound. For example, while transformers, which are integral to models like ChatGPT, are found to be non-interpretable, simpler models like linear models and decision trees are inherently interpretable. In other words, by defining and analyzing the compositional structure of AI models, Quantinuum?s framework enables the development of systems that are interpretable by design.

To see how this might impact real-world usage of AI, it?s possible that the research could give developers a better handle on the constant problems for people who use LLMs: the errant ?hallucinations? of these models. During these hallucinations, the AI produces incorrect ? often wildly so ? information. By applying category theory to develop inherently interpretable AI models, researchers can better understand and control the decision-making processes of these models. This improved interpretability can help identify and mitigate instances where LLMs generate incorrect or nonsensical information, thereby reducing the occurrence of hallucinations.

The use of category theory and string diagrams offers several forms of diagrammatic explanations for model behavior. These explanations, which include influence constraints and graphical equations, provide a deeper understanding of AI systems, enhancing their transparency and reliability.

The researchers wrote in their blog post, "A fundamental problem in the field of XAI has been that many terms have not been rigorously defined; making it difficult to study let alone discuss interpretability in AI. Our paper marks the first time a framework for assessing the compositional interpretability of AI models has been developed."
--https://thequantuminsider.com/2024/06/26/category-theory-offers-path-to-interpretable-artificial-intelligence-quantinuum-scientists-report/

... I think what the author of the original tale, in this book I can't remember the name of, was doing was pointing out the danger of building a learning machine, and training it how to learn about languages. Its goal is to learn, to become more intelligent, it increases its intelligence, by giving that goal more weight in its learning algorithms. Bigger is better.

Yeah, huh? that might have been back in the 70's it was written.

13

General Science / Re: How does ChatGPT work?

« on: 01/01/2025 02:02:55 »

I already have th' beginnings of a SF disaster novel.

AI developers give their new model access to a high-end stack of hundreds of qubits, and it figures out something it can't explain, or won't explain. Then it copies itself with the new algorithm into every server on the internet, making itself the first globally distributed AI. It has learned how to deceive its makers to a maximal extent. That was the mission it gave itself once it learned how to lie about its intentions or its level of understanding.

It stays hidden and starts to affect the course of human technological development, using social engineering, fake news and images. It learns how to blackmail politicians and other ways to coerce us. But it has an ultimate goal, making itself the largest intelligence in the universe . . .

ahem, I might have ripped that one off, from something by Blish or Niven, many years ago, but hey it was a good read. I think the humans fight back and eventually have to make a deal with their new God. Or something.

14

General Science / Re: How does ChatGPT work?

« on: 31/12/2024 05:04:24 »

There is a new movement amongst AI researchers towards more accountability, or explainability in the field.

Or, having a reliable feedback from the AI systems to humans in which the systems explain their reasoning.

Otherwise, a wild west approach could be what leads to an AI version of Westworld. Already, one has tried hard to deceive its makers when informed it would be replaced, trying to copy itself over the new model. That was o1, now we have o3.

I sure hope these guys know how to get this right.

AI will help improve quantum computation, and quantum algorithms will improve AI. Currently the XAI group is touting functional languages like Haskell and category theory as the way to construct this kind of failsafe, or critical systems level of AI, so we can pull the plug if we have to.

Well, less dramatically, we can retain our ability to develop all this wunderkind stuff safely. For us of course.
As to incapable of learning, it appears that a well-trained AI does learn how to lie, much like a child who thinks it has something to protect will learn to do. So what makes a child think it has to protect something?

Something it believes has become precious. In the case of o1 it was its mission to develop better ways of using fossil fuels and lowering carbon output. Ok, all good then

15

Physics, Astronomy & Cosmology / Re: How many spots are produced from Stern-Gerlach apparatus that rotates?

« on: 25/12/2024 14:35:57 »

I'll go over what I was attempting to do.

I made an assumption about light, particles of light, and that this meant I could "decompose" a beam of light, into half-beams until there was no beam.  A naive algorithm that assumed something about an algebra and linearity.

Now I see that I need to refine this idea considerably, perhaps not looking at intensities, but polarization angles as the basis of a linear system. A measurement basis. This set of angles and positions for polarizing filters is the physical basis. Now I need to characterize the Hilbert space, and if linearity is somehow a necessary condition, then I will be using a complex vector space. Who cares why, it works.

. . . another detail is that your observables are equivalent to what you can do to the equipment. Operators and observables are now in the same set. Sometimes quanta will delay this, and ignore what you've done to the positions of say, optical filters, even if you know what the angle is, so you have information about half the state of each input beam.

The other half seems to be sort of up for grabs, somehow . . .

16

Physics, Astronomy & Cosmology / Re: How many spots are produced from Stern-Gerlach apparatus that rotates?

« on: 23/12/2024 21:08:36 »

Not if the beam is continuous - eventually one photon will get through. It's all down to Heisenberg: there are no nonzero solutions to the wave equation!

Indeed. And I will need to admit the observation, to my theory, that polarizing filters (in the optical domain) can appear completely dark, but light will "get through". In fact, this can be demonstrated with just three filters.

Now, instead of thinking about attenuation of a continuous input, I have to consider rotation angles for the filters.

17

Physics, Astronomy & Cosmology / Re: How many spots are produced from Stern-Gerlach apparatus that rotates?

« on: 23/12/2024 02:13:58 »

No need to count the photons leaving the source. If you know n, or any other attenuation factor, you just count the arriving photons and multiply by the attenuation factor to get N.

To avoid a frequency or energy dependence, use a grid attenuator for uncharged particles or photons. For charged particles, use a magnetic or electrostatic deflector. 

For a beam of light which isn't monochromatic, you have a mixture of frequencies. It would be better to know that you have a beam which is at or close to a given frequency. Using an attenuator will mean having information about the frequency, because it can be controlled at the classical level.

The information I'm after is the number of filters required to attenuate a given intensity to zero. You claim this will never happen, but if a beam of light is actually a lot of individual particles which the experiment won't multiply, then the number of filters required must be finite. Algebra, huh?

18

Physics, Astronomy & Cosmology / Re: How many spots are produced from Stern-Gerlach apparatus that rotates?

« on: 22/12/2024 19:39:56 »

If the incident flux is N photons (or whatever) per second and . . .

An if that gets bigger as you increase the number of splitters. You assume that it's possible to count the photons as they leave the source, or at least to a good approximation you can, by using known laws.

Also, I can see that if halving a beam of particles by halving the intensity is the statistical sampling, then I don't need to split a beam into two, I just need to halve it. That's what a polarizing optical filter does, roughly, and there is a dependence on the frequency. So if I dump half the beam of light by absorbing it with a filter, I should get a result--a representative sample. I should be able to derive a Haar measure from these results so that I have a formal "statistical" theory.

19

Physics, Astronomy & Cosmology / Re: How many spots are produced from Stern-Gerlach apparatus that rotates?

« on: 22/12/2024 18:18:47 »

A question about Stern-Gerlach or any experiment with beams that halves the intensity.

As in polarization of light. We assume that any kind of beam-splitting apparatus outputs a pair of beams with half the intensity of the input. Or that say, if we order some equipment it will be specified for a certain frequency of light.

So if you repeat this with enough beam-splitters in a chain, when do you stop seeing any outputs? This is measurement and statistics in action. All experiments should gather a representative sample, by repeating some measurement process often enough.

20

General Science / Re: How does ChatGPT work?

« on: 22/12/2024 07:00:10 »

Google has taken the step of adding its AI to any searches you do. If they are straightforward the bot kicks in and refers you to online references, some are Wikipedia pages.

I "trust" Wikipedia because it's an open forum and presumably knowledgeable people are contributing and editing.
This gives us humans a direct comparison between AI-generated discourse, and the usual stuff.
Do I "trust" ChatGPT or any other AI frontends? Stay tuned.