Monday, February 8, 2016

Korean rocket and Navy LENR illustrate pathways to NUC/AI extinction

Posted to scientific advisory board of Lifeboat Foundation, discussing possibilities for extinction of human species:


There are actually many specific pathways or credible scenarios or real time-tracks leading from where we are to the clusters of human species extinction which I label as "NUC/AI" for short. (H2S is also still in play, but two items are enough to analyze in one post.)

Probably you already know about the new North Korean rocket launch, and its link to their recent "H bomb" test. I recently received several inputs on the new Navy LENR activities, which are just as important; one of the inputs was the link:

The Korean rocket does relate to various "NUC" type extinction possibilities, but it calls out for some discussion here and now on this list because of how it relates to the bad-AI issue we have discussed here a bit. The discussion was a bit fuzzy, and it helps to anchor a bit in something concrete.

There are ways for the US to react to the Korean rocket which would play directly into the AI extinction scenario, following the very serious Terminator script. Since some of you really demand evidence (which is a very good thing), I would ask you to please scan the simple colorful slides at:

which in turn are substantiated by the citations in the accompanying paper, for which the link and background are posted at the top of

One of these slides illustrates the work of S.N. Balakrishnan of Missouri, who implemented one of the RLADP neural network algorithms I developed in the 1970s (DHP), and showed that it reduced error in hit-to-kill missile interception applications by more than an order of magnitude. It was a long and entertaining path, and I will resist the temptation to elaborate. Just two stories: 

(1) in 2009, I attended a briefing at the Marshall Institute for Congressional staff (which I was) on missile defense, and couldn't help smiling when the technical guy with the Lockheed representative said: "What you folks need to understand is how much more real the opportunities for missile interception are now. Yes, back in the days of the Patriot missile, we only hit about one or two in a hundred, but thanks to this guy S.N. Balakrishnan, we can do it much more reliably. His mathematics is a bit weird, and we are struggling to understand it, but a whole new world has opened up to protect us...

(2) Many people believe it is a top security priority not to let the world know, and thus not discuss the very limited surface details I am posting here today. But ALSO a few years ago... the Chinese arranged two major speaking tours, one by Balakrishnan and one by me, curiously all at the same places, one of them being Harbin (the main center of hard military aerospace activities in China of they types they don't mind Russians knowing about). No way they don't know. They can do web searches as well as any of us. I came on official NSF travel, after full approvals and checking on the rules, and had no security clearance; however, there were amusing little cameos where a couple of people just assumed I must have some CIA connection, and I work actively to forget some bits of that. China is also certainly not the only country which knows. (Balakrishnan has bent over backwards to hold back stuff, respect important security concerns, and even offer a few primrose paths to the web searchers.) When we go to nutty extremes in holding back, we hurt ourselves and our own consciousness more than we reduce proliferation; some of you think of the very serious nonsense about Clinton emails, but I think more about the historical introduction chapter in The Skyrme Model by Makhankov, Rybakov and Sanyuk, who recount how classification in the UK made certain nuclear information widely disseminated in Russia but almost unknown in the West.  

I will refrain from boring you with discussions near SPAWAR linking to Iron Dome, Russians and a friend in blue uniform who really looked like a character from the Terminator movies, and really had an option to make the whole thing come true.

All that being true, one obvious option with Korea's rocket launch would be to use it as a test of antimissile interception capabilities. With a previous Korean launch, that was discussed quite actively in the press, but not this time. Should we seek a UN security council resolution authorizing the US to intercept any Korean launches until and unless  they stop being such serious NUC outlaws (and violators of strong past commitments) that they no longer add to the NUC extinction pathway (not directly but opening up a path to an avalanche)?

In fact, Iron Dome has already displayed major capabilities, already saving many many lives in a situation where lack of such a defense might already have led us into a nuclear war. 

But: (1) Is the Bala/Iron-Dome level of missile defense technology really safe (versus the AI risk)?; (2) If we start to need even better defense, as larger NUC risks grow in the world, just how much more is both possible and safe along the same pathway?

In my view, the technology deployed so far is reasonably safe. It is an example of what I call "vector intelligence," which can handle much more complexity than the earlier reinforcement learning schemes of folks like Minsky/Selfridge and Barto/Sutton/Anderson, but is still limited, and can't optimise more than about one vehicle with about a dozen state variables in a highly nonlinear environment. (And yes, we have stability proofs. One even in Automatica this past year, albeit for a level weaker than DHP.) 

On the other hand, with more missiles launched (or with the task of coordinating whole fleets of drones), more complexity comes into it. "Theater missile defense," exactly as in the highly realistic Skynet scenario, is what I once viewed as "a beautiful testbed demonstrating the value of rising to a higher quantum level of intelligence." It is, and we know how to do it. But I am very grateful to the guy from Boeing, who in their Executive Dining Room in 1991, strongly urged me to go see Terminator II: "You may not like that kind of movie, but you really have to. It is your absolute duty, given what  you are doing and what you are doing with us right now." More of the story is at 
But no, I do not intend to go there. I do at times worry whether other people might someday catch up... have we just bought a little time, and are we out of the woods yet?

Fortunately, there is another way to substantially improve missile defense, well justified in today's world (and hopefully something we could work with Russians on as allies of a sort, as Reagan wanted to do in this area)... and that is good old space-based boost intercept. Just a few weeks ago, I was happy to see some really nice reviews out there on the web, which I bumped into when checking on high-energy and exotic lasers. Livermore really had a nice story. Some guys at Boeing really are excited about it. When I came back from Marshall's briefing, I asked a friend (a key guy on Senate defense appropriations subcommittee) their view; they said: "Story sounds great, but the numbers are wrong.  Cost per hour of protection is too high." Having once been a student of Tom Schelling (who advised McNamara on the economics of defense), I fully understood what he meant instantly. But I also understood ... since we DO have the technology and design (integrated by Ray Chase in a variety of key papers and sources) for RLVs with $500/kg-LEO (and with some GEO capabilities), we COULD bring that cost way down, and really afford to deploy space-based missile defense. In my view, that would be a great and safe thing, making all the world much safer, and providing really huge side benefits (which we list again and again in space policy papers). For example, it is the necessary enabler for 9-cents-per-kwh switchable space solar power, per the plan put together in the most recent book by John Mankins.

Unfortunately, the same bad guys that Hillary Clinton, Trump and Sanders complain about are also putting a hard freeze right now on our hopes of deploying or even continuing to have that RLV options. There are very scary specific stories about bad things done by really bad guys in DC. But "speak of the devil and he may appear." I won't go there either now. 


So... that's the Korean rocket... where it leads us... but what of the LENR case?

There are actually many specific, diverse clusters of pathways leading to extinction by misuse of nuclear technologies. One of them is the "lone wolf nuc" development. Since the Navy, US News and NASA have already let one of the cats out of the bag more than I am even capable of by any mundane channel, I chose to reply to the new information:


To X:

I have known that LENR works for some time. Here is some background, a link
and some extensions. Please do click to see the video...
taken when Lamar Smith's guy (Pramod Khargonekar, who told me that day that his deals with Smith bothered many people but "hey, they get us money") insisted that cold fusion be the main theme of the speech given on the occasion of my retirement:

(another email sent to a couple of friends on September 8, 2015)

Your probably know that a guy named Dennis Bushnell, Chief Scientist of NASA, has been very effective in getting the word out worldwide on cold fusion. 

I do worry about the nuclear proliferation aspects, but at this point it's my calculation that the underlying science is important enough and the cat is out of the bag enough, that's it time to let a bit more out of the bad. Of course, it helps that I retired from the US government on February 15 -- in part because of policy changes and personality conflicts in the NSF Engineering Directorate which led me to conclude it would not be a good use of my special abilities to stay there. 

I did not do anything with cold fusion since about 1990, but Kargonekar, being ever so friendly, decided to bring in a guy involved with the cold fusion story to make that the highlight of the parting speech for my retirement. But it didn't work out precisely as he imagined. Here is a 15 minute video
which left some of the NSF people open-jawed:

Somewhat more seriously, here is a folder of documents I scanned as part of compressing my files (throwing away lots of paper both at home and in the office) in my final months:

The "CF" cold fusion file includes such things as lab notes from the Pons/Schwinger collaboration, and the report of the NSF workshop of 1990 or so.
That stream of technology is actually just a first step towards a higher frequency coherent technology related to further steps which in my view are important to even more remarkable future technology. Wild as this is, it is the tip of an iceberg. 

By the way, this is all on the cloud - but life being what it is, and my life expectancy being finite, 
it might be reasonable to make backup copies not dependent on my machines and cloud areas.


..... (some further nuclear stuff) ..... 

I had good personal contact with the guy at Lockheed behind this LENR stuff, who collaborated with Macgregor of Livermore, but the "black wall" later got in the way. 

It is a nasty dilemma, that technology which humanity might need to survive long-term seems to pose ever more lethal risks short-term. Sometimes I think about the Babylon 5 episodes where they judge humanity is not ready yet for real longevity... more realistic than what I hear in the debates these days.  


Best of luck,



Here on this blog, maybe I should also add a little news n the H2S front.
One space guy asked: "if this is real, why don't I see it in the press?" I have posted the logic, but not tried to appear on reality TV. A life style choice. More seriously, when I discussed the issue last with the most serious ..
university specialist (much published in Science) I have worked with in this area... he said that I am late to the party. A whole lot of people realize the H2S extinction thing is likely to kill us, but, in view of the active censorship  driven by folks like Lamar Smith, Cheney and Koch brothers, penetrating much deeper than TV watchers realize... they are pretty much giving up. After all, Antarctica is the Big One, and we are still 40 years away from the start of the major oxygen crunch, according to best data... 

Yet I saw a new map yesterday from the ARCTIC which is also disturbing. No, even the Arctic and the North Atlantic combined would probably not produce and emit enough H2S to outso the smaller great extinction
(the paleocene discontinuity) which killed all proto-primates and other big mammals on earth but was not enough to kill the mice. The Big One (PT) could kill them all. But even so, the news... I think it was Science News...
clearly shows lowest ice ever in the Arctic, WITH the remaining ice mainly concentrated near the shores of 
Greenland and far north Canada... suggesting the kind of freeze of fresh water runoff we also see around the Antarctic. On first glance, it seems... Antarctic has ALREADY seen shutdown of the biggest "lung of the planet" (souther thermohaline current, THC), while the Arctic is smaller... and might not be cut off... yet... though the new map suggests ... maybe, maybe not. HOWEVER: once cutoff occurs (or occurred?) in the Arctic, 
the possibility for bad stuff happening is much quicker, due in part to depth of the water and direct runoff.
Could we smell it in our lifetimes? Maybe. Maybe that wouldn't be so bad, if we survive it, insofar as an early whiff might possibly wake a few folks up. Then again, people voting for oil/Cruz show that panic by itself is not enough...    

No comments:

Post a Comment