Atomic Razor

The Nature of the Firm

I don't know what publication that person on Bluesky works for that has a house style that deprecates the use of the term firm except for firms that are actually firms, i.e. some kind of partnership, but that notion is completely contradicted by the theory of the firm, i.e. it's called the theory of the firm, not the theory of the business or the company or corporation or enterprise or organization. So, one might wonder how useful that distinction is the readers of the publication.

At the Museum of Flight in Seattle they have a reconstruction of the Boeing factory from the mif-1910s. They have something similar in the Science Museum, a reconstruction of a Victorian workshop. I have wondered whether this is relevant to Coase. So, that the owner of the works hires worker on a daily or weekly basis as needed. The worker might go and work at another works as suits them and as the demand for labour varies between shops.Furthermore, a worker might contract with a shop to use its specialist as each worker is also an entrepreneur in their own right. So, in the 1930s, it might have seemed odd that to organise industrial production on the basis of permanently employed workers. The production line was still quite new at the time, so the changing nature of employment practices might have demanded an explanation or even appeared to be a paradox. Of course, large scale industrial organisations have existed for thousands of years and there might be an explanation that a high degree of casualisation (i.e. that people were hired on short periods) existed in their practices, although that can't have been absolute. Of course, a lot of modern business practices arise through the railways, so it would be interesting to understand how they spread from there throughout the whole of the economy.

Generations Numbered

Currently nearly 8.2 million generations in. So still over 85 million to go. Not even 10% of the way there. But the streams are getting closer. They are going to interact... eventually. And the bottom region is somehow still active. How is that possible? We have an irregular region before it becomes period, so the puffer is synthesised by that region. But there has to a glider gun in there somewhere that has a very long period. Which isn't something one would expect, especially if we the period is irregular. 

But, yes, I would like a grunty computer. I remember Jum Emerson in ~1990 talking one of the pacademics in the Theoretical Physics group at QMW who said he wanted to have the most powerful computer in the college in his: "I used to think he was joking, but then I realised he wasn't!" Of course, how powerful a computer could you have got in your office in 1990? But a $2000 Dell would be pretty grunty in 2024. It will have to wait though.

Move Fast and Break Things

My third Thanksgiving in the US and my first as a permanent resident alien. What is there to be thankful for? Well, not the geopolitical situation, which is the gravest it has been since 1933. The point about regulation is that it's there to try and make things safer. Yes, you might get a degree of capture by the bureaucratic-regulatory state. The point of a regulator is to regulate and so it will regulate. Sometimes. Many regulators are toothless and compromised. Americans don't like bureaucracy because they are bad at bureaucracy. It's possible that that nuclear power station hasn't been built because of regulations or that drug approved because of regulation or that rocket launch. Getting rid of might mean that some of these things get done, which might be beneficial to some people, the only possible ray of sunshine as we tap dance towards the abyss, but the thing is that a lot of people are also going to die in the mean time.

Nature’s Bounds on Transmitted Information: Defining The Limits of Observational Astronomy

Martin Harwit is still alive! He's 93 and was giving seminars on Nature’s Bounds on Transmitted Information: Defining The Limits of Observational Astronomy in 2021 or possibly even later, which is pretty good going. I bring this up because I saw on Bluesky a link to a Google DeepMind paper on ""A new golden age of discovery: Seizing the AI for Science opportunity", which is, of course, what I want, but then I remembered that Harwit in one of his books, presumably Cosmic Discovery, which I probably in QMW library, talks about, well, the limits of observational astronomy, that there's only so many things you can know. Of course, in astronomy, there's a limit to the resolution or sensitivity of your instruments in any given region of the electromagnetic spectrum (or, indeed, these days, other modalities), so there's some kind of information-theoretic limit on what you can get out and thus, perhaps, in some ways, on the number of laws of physics we can know. 

There's a couple of counters. We aren't going to run out of physics to do in the near future because there's lots of things to investigate. One could build big accelerators, in principle, or other detectors and that leaves aside areas like molecular nanotechnology or turbulence before we even get to exotic systems we have thought of yet (exotic here just ordinary weird physics, like phases of matter that haven't yet been constructed, not anything really exotic and novel). This goes ten- or a thousand-fold or more for biological systems (at all levels) or, indeed, cybernetic and complex systems in general. We still don't have a comprehensive catalogue of all the bodies in the Solar System over 10km in diameter, much less one for every stellar system in the galaxy. After a while you are stamp-collecting, but that doesn't mean there's still not much of interest to be found in both specifics and generalities. You can over come some of Harwit's limits by going to a neutron star or a black hole, although there would be diminishing returns eventually. Of course, what of cataloguing pseudo-arthropods on distant planets, but he would say that's not a law of nature, perhaps. We might talk of the law of evolution rather than just the theory and it would be fascinating to have thousands or billions of examples of biological evolution from which, doubtless, some high-level laws about universals and parochials could be deduced.

I am reading Dario Amodei's essay "Machines of Loving Grace", of which more when I have finished it, which is still pretty sf even if he explicitly tries to step away from the sf-ness of the implications of "powerful" AI, but I suppose he's gesturing towards those of us who want a Singularity that gives us our personal FTL starships by the end of 2026. Harwit would doubtless say that until then the best we can do is to find microbes in the atmosphere of Venus or Martian aquifers or the oceans of Ceres, Europa and Enceladus and look for biosignatures in the atmospheres of exoplanets. It's not what we really want, but it's something and it will still keep us occupied for a while yet.   

Through the Looking Glass

A new nuclear arms race seems near inevitable. Even with the US likely to be little more than a puppet of Russia in the years to come, and the possibility of an Axis of Steel including China - after all, really, what's not to like about an authoritarian dictatorship? - it's very unlikely that China is not going to build up its nuclear forces substantially over the next few years and the US is almost inevitably, especially with the end of the strategic arms limitation agreements, going to modernise its nuclear arsenal. That's good for some people, there's going to plenty of jobs building nuclear weapons and missiles and all the ancillary systems that go with them. That's the thing,, there are almost fascinating to read about as long as you don't actually think about things too deeply. As Stanley Kubrick said back in the early 1960s at the time of Dr. Strangelove, the strategic systems might be 99.9% reliably fail-safe in any given year, but if you multiple a small probability enough times the probability of a failure tends to unity disturbingly quickly. Which was why he was so pessimistic about the prospect of avoiding nuclear war. And yet he have over 60 years since DS, but hear we are now making things more dangerous in a word in which the possibilities for more actors making their own WMDs become daily more likely.

Obviously, in terms of nuclear logic, nothing is as it seems. Operation Looking Glass is (or was) "the code name for an airborne command and control center operated by the United States". This part of a vast array of historical, current and future systems. Several other countries must have at least have some parts of that functionality. But is Russia still using their equivalent of 8-inch floppies as part of their "command and control system for intercontinental ballistic missiles, nuclear-armed long-range bombers, and ballistic missile submarines"? What state is Dead Hand in right now?  Does it, through that nuclear logic, actually makes the world safer by taking humans out of the loop? Or is a real-life version of the Doomsday Machine in DS? It would, of course, have been fascinating to have become a defence analyst, but that was another path not taken. But until I read the Dead Hand article a few days ago, I was unaware of command missiles. Dead Hand supposedly uses one based on the SS-17 Spanker ICBM (but still or has that been replaced by a current missile, probably not the RS-28 Sarmat)? The  US had the Emergency Rocket Communication System util 1991. The Minimum Essential Emergency Communications Network subsequently substituted for it. There's always more and we have to hope it is enough each time we roll the dice.

 

Large-Scale Natural Engineering in Conway's Game of Life

We are nearly 2.6 million generations in with well over 90 million to go, so there's no this is going to run to termination given that it's been at it for 30 hours or so already. Still, we might be able to see something interesting. At some point in the not too distant future, the glider stream from the right bottom, which seems to still be emitting some gliders is going to start interacting with the wake glider stream from the head. 

This is, I think, technically, a puffer, or at least some kind of variant on a a puffer. The head and the wake are moving, so any interaction between the glider streams that leads to a glider propagating back along the wake can be eventually arrive back at the head and start causing chaos. So, we will get some kind of novelty when the stream collide and then later when the glider gets back to the head, I am guessing.I think the bottom is still producing gliders, but I need to understand how. If it is it's (very) long period and how could a very long period gun arise naturally. Well, of course, it can because any period of gun is possible from 14, but I would have thought that it's much more likely that short-period guns will arise than long ones. I have increased the speed so now the grid only updates every 1000 generations, but that means the time between updates is longer and I still feel it's like to run out of memory or the machine will be rebooted before we get to the end, but I will try again with hyperspeed from the beginning.

As I said the space is truly enormous even with just 16x16 soups and changing just one cell will usually completely change the behaviour, in this case back to a typical 16x16 soup that probably a few hundred generations. The run might have frozen, so I will wait a while and restart if necessary.   

Not Even the End of the Beginning...

I have two screens, so if you take a screenshot you get both screens. I guess there's a way of avoiding that, but have a gander at a bit of the Mandelbrot set. I am sure there is some profound connection to cellular automata and Conway's Game of Life at certain levels. The 95-million-generation 16x16 random soup does not appear, oddly, to have a common name. Here we see it at generation 230,634, so not even 0.3% of its way through its non-boring life. What we have here is a naturally occuring glider gun, which is emitting gliders from its head in three beams as it slowly crawls to the north-east. I am sure there is a term of art for this kind of structure.

The structure is quasi-periodic, which, I suppose, is not surprising for a pattern that is going to survive for such a long time. That suggests a connection to fractals because the pattern can't be completely periodic because if it were  it wouldn't eventually become boring or rather it would be boring from the beginning (for certain values of "boring", of course).I am guessing that at the end it's no longer emitting gliders, so things are going to have to change. The extraordinary thing is that the deterministic randomness that must be in there doesn't kill the pattern's evolution much sooner. But obviously when you have 2²⁵⁶ patterns some of them are going to get very lucky indeed and there are doubtless much luckier patterns than this out there. I read that there's an exponential falloff (I guess as opposed to a power law, but you can still curve fit) in the lifetime of patterns, so it would be interesting to see what the longest-lived patterns are anticipated to be for asymmetric soups, symmetric ones much more easily being able to produce immortal patterns of different kinds. 

"You and I will never know"

The Weasel reminded me of an exchange of emails we had back in 2009 during the swine flu pandemic in which I sent him to a link to the FoAK article on a writer called Breece D'J Pancake.

"After Pancake's death, author Kurt Vonnegut wrote in a letter to John Casey, "I give you my word of honor that he is merely the best writer, the most sincere writer I've ever read. What I suspect is that it hurt too much, was no fun at all to be that good. You and I will never know."

I had completely forgotten about Pancake despite his name and the most fulsome praise of a writer I admire greatly. I wonder why I didn't buy his single collection of 12 short stories then. It has been called the American Dubliners. I have remedied that omission now. There is a Library of America edition that contains the twelve stories plus some fragments and letters. So, I shall see for myself what it was that such an impression on Kurt back in the 1980s.

Random Soup

As I said to the Weasel, one cannot fault the energy and enterprise of Dr Adam P. Goucher. He even works for OpenAI! and claims to have been an Associate Member of the Emmanuel College Monarchist League. This might be a joke, but what isn't a joke is his apgsearch program. As of right about now, 2024-11-23 00:49:50 UTC, it has run 311 459 888 046 103 asymmetric 16-by-16 soups (with a density of 0.5) in B3/S23 (i.e. Conway's Game of Life) with a total of 6 803 790 810 204 363 objects of 536 399 distinct types being found. Of course, even 311 trillion random soups is only the merest fraction of a fraction of a fraction of the 2^256 possible soups (OK, some of those are duplicates). The main run is focussed on asymmetric soups. There are 2^128 bilaterally symmetric soups, but obviously the chances of stumbling on a symmetric soup by chance are effectively 0, but perhaps it's worth checking. Symmetric soups tend to produce larger symmetric objects, so they are slightly artificial. 

The vertiguous thing is that the exploration of the space of behaviour of the just a 16x16 random soup has barely scratched the surface (although obviously we know have a good idea of what typical asshes are like statistically.  And this at a density of 0.5. Thinner or thicker soups would have different behaviours. And this is just 16x16 soups. Even 7x7 soups would be inexhaustible, although could quantum computers help perhaps. Yes, this is yet another topic for Little, Big. But the really fascinating question is what lurks out there. This 16x16 random soup takes more than 95 million generations to become boring

bbbbbbbobooboobo$

bbbboboobbobbobb$

bbobbboobboobboo$

booobbbooooooobb$

oobobbooobobbboo$

bobobbobobooobbb$

obobbobboobbobob$

boobbobbobbooobo$

bbbobboboooobobb$

ooboobbbobobbbob$

obbbobobboobbbbo$

bbobbbobbooooobb$

oobooooobboooooo$

booooobbooobobob$

boooobooobboooob$

oboobbobooboobbb!

I am fascinated to find out how that compares to a typical soup. But how long would the longest 16x16 last? Presumably, you could guess by plotting the number of soups versus the number of generations because it might follow a power law, but there's the possibility of getting patterns that just act in totally unexpected ways and thus are much longer lived than expected. These results are for open universes. But what about (large) closed universes (on tori?). How easy would be it be to get immortal patterns with small or large starter soups? A large ash that has emitted gliders that will eventually return is going to get distrupted, but might still flicker out in the end, the last few gliders ultimately annihilating one another to leave the universe empty and lifeless.  

 

Shelfie No. 1

Most of my books are still in my storage unit in Brentford, with some in Preston, but I have quite a few here. I am hoping to get them over next year and I will to work out how to organise them. I asked A about that, but she was non-committal, although there must be a body of best practice in organising "small" personal libraries.

Some notes:

1. A couple of self-published comic sf novels that Spencer gave me at Worldcon in Dublin.
2. A collection of essays by the well-regarded Australian writer, Gerald Murnane. Should probably go with the other ...And Other Books books.
3. Despite being a Picador, The Tolkien Companion isn't,  I understand, supposed to be very good or accurate. I have a vague recollection there's more Tyler's story than indicated in his FoAK article. Does Nicholas Royle discuss this in his book on Picadors?
4. I like books on writing books. I know, I know. Three books on writing mysteries and three random creative writing books. 
5. I have not seen The Corner mini-series, but I am a big fan of The Wire, of which this is one of the source texts.
6. 1491 is excellent. I have started 1493, which used to be here, but is now upstairs. For some bizarre reason, Amazon sent me two copies of each,weeks apart, so David got those for his birthday in 2022.
7. Four Jonathan Raban books. He was, of course, famously the Englishman in Seattle for many years and my ambition was to get to know him. Sadly, he died in 2023, so that is one drea, that is never going to come true. I have only read his Seattle novel, Waxwings, which is excellent, and his LRB and NYRB essays.
8. A random 1995 book on the future of television by a couple of Americans from OUP US.
9. A biography of the legendary architectural theorist, urban theorist and Ameriphile, Rayner Banham. I have read his essential Los Angeles: The Architecture of Four Ecologies and there's an early 1970s BC documentary by him on YouTube.
10. This white cover ...And Other Stories book should definitely go with the other white cover ...And Other Stories books.
11. Bought at the City Lights bookstore in San Francisco. The title sums it up. Looks interesting.
12. I need to. For many reasons. See this NYRB article: "It all began with a book review. Last year, I read an article by David Aaronovitch in The Times of London about Michael Pollan’s How to Change Your Mind.". Note that broadening your consciousness does not stop you becoming a professional T(E)RF, so bare that in "mind".
13. About building a toaster literally from scratch. i.e. from the elementary raw materials. Features Jan Cilliers from Imperial. I don't recall if I ever met him or worked with him, but he was oft-mentioned, so that that's other Small World  connection. Great fun and highly recommended.
14. About the underground (not the Underground).
15. A couple of books about skyscrapers.  
16. "How Capitalism Took Hold of Graphic Design and How to Escape It". Spotted at the Elliott Bay in the Interesting Books case just by the door. Dutch, although the text is in English. Looks interesting, but would probably outstay its welcome.
17. I was a big fan of this sitcom for obvious reasons. I might even have created the Wikipedia article. Interesting that this isa BBC-branded DVD whereas Joking Apart was done through an early kind of  crowdfunding. 
18. One of Those Books.
19. This was cheap on BookBub recently, but I already had it. Histories of America's geo-cultural regions.
20. Includes "Going on Robert Hodge's "Civil Wargasm", a week-long journey to various battle sites in Virginia and Maryland, remaining in authentic uniform and sleeping on the battlefields".
21. Is what it says.
22. From the days when you had to write to newspapers to have questions answered. The Past is a Foreign Country.