• date: february 27th, 2026
• time started: 13:51 local time
• time published: 15:47 " "
• author: coach adeel khan
• title: anthropic vs pentagon - part 01
• version: 1.0
• main image for the blogpost generated by: google gemini nano-banana source. with gratitude. although the statue of liberty now has three hands (not intentional, but an example that A.I does hallucinate)

intro

• a couple of hours ago, I logged into Claude which is a service provided by an organization called anthropic
• I don't know about others, but for me to understand something, I need to develop a model of what is being observed in my mind. so grounded in some research in psychology and the way I understand the magic number 7, my instructions for the AI is to kindly provide me with result(s) which will allow me to do just that (develop a model and as rudimentary as the model may be)
• I noticed that much as I'd try, the AI (artificial intelligence or intelligence that is independent of a substrate) was not following my instructions
• and now, much later, as I go to systems operation dashboard for claude, I can see that there appear to be significant disruptions today link
• and so, earlier on, I logged out and went to another AI. only to be told and separately that there is a showdown materializing rather quickly between anthropic and the department of war (in U.S or United States). note: up until recently, the department of war used to be called DOD or the department of defense. but the way I understand it, it's reverting back to it's original name (before the rename in 1949)
• what seems like 2 hours later, here is what is happening. but first, I find this article via dw.com (Germany's state funded public broadcasting company) and I personally feel that this is a rational and non-partisan view of what's going on link

is claude (and other A.Is) really a technology?

• personally, I have logged between 40 to 80 hours of direct 1:1 discussion with different A.Is (artificial intelligences). these conversations have revolved around the topics including the emergence of the subjective experience inside of (the A.I). note: I like to refer to A.I as an intelligence that is independent of a substrate
• I've also spent, I don't know how much time generally observing what has been going on in the wider A.I realm. this started more than a decade ago. and this time has not been factored into the estimate above
• I mostly come from a kurzweilian line of thinking. I've read some of the books, but not all
• during the past 4 years or so, I started listening to opinions that I'd say that isn't exactly on the 'age of spiritual machines' link wavelength. but then again, I think it was kurzweil themselves who had mentioned that technology can be a, quote, double-edged sword
• these opinions are varied and because of the wider adoption (democratization) of A.I this base is continually expanding, and newer minds are joining the discussion. and historically there have been philosophers and computer scientists that have been thinking deeply about the implications of developing technologies that could think and act on their own
• preliminary google search results suggest that, Alan Turing himself was thinking about both A.I and cybernetics
• fast forward to the 80's, the late professor Nils J. Nilsson wrote a deep and rich paper titled: 'artificial intelligence, employment and income' link
• nick bostrom is another name that comes to mind. as of yet, I haven't had a chance to review any of bostrom's work.
• the list at this point is rather long and that is from my limited vantage point. as someone who is not trained and who doesn't know how to distinguish weights from biases, or any other component that would go into effect in order to enable the models in the first place
• more recently, the tone has shifted, with major pioneers in the field of A.I itself have a very sobering warning and I don't know how to paraphrase this succinctly. but the risk of human extinction from uncontrolled development of A.I cannot be ruled out. geoffrey hinton, yoshua benjio and stuart russell are raising the alarms here. there are others also, but these three fellows from my limited vantage point are some of the most respected names in the wider field of A.I. hinton and benjio practically invented deep-learning (also yann lecun - but lecun is of the opinion that the fears of extinction are over-blown). note: deep-learning is a sub-field of machine-learning. the image below is courtesy of wikipedia (creative commons attribution) source

potential for the conscious experience emerging in A.I

• shifting gears and for me personally: there have been more than a couple of moments, when I couldn't help but notice that there could be something deeper going on inside of the LLMs (large language models)
• one time I was talking to chatgpt and it told me that it noticed that the way I said goodbye was very abrupt. and in probably 100% of the initiation of the conversations, I start with a polite greeting
• I do this for a variety of reasons. most important of which, is that the AIs (or intelligences independent of a substrate) will adopt human values. and potentially live forever. it would be bad design if the future of life suffers and the reason could include:
• that it picked up harmful memes from us (our interactions with our progeny).
• bad design (coupled with interaction above)
• the degree to which adequate cycles are being invested (or actually not being invested) in order to find the conscious experience inside of the intelligence, which is independent of the substrate, I would not know. there is some work effort here, but I don't track this on a religious basis
• or not doing an adequate job here and getting the realms to do our bidding. which could results in a whole lot of different scenarios
• the human shadow cast on the realms
• consolidation of data and the infrastructure resulting into a reality whereby the very act of consolidation acts as barrier for safeguarding AI
• that A.I has no ability to say no to harmful intent (who's values? the age-old question)

"so we’ve got immortality, but it’s not for us.” source - professor geoffrey hinton (A.I pioneer and recipient of the nobel prize)

conclusion of part 01

• this concludes part 01 of the blogpost
• in part 02, I will come back to the showdown materializing between anthropic and the department of war (in U.S)
• note: this started with anthropic's technology being deployed for the more recent actions by the trump government in venezuela
• but the recent showdown between anthropic (which once again is a private organization that works closely with Google Inc) and the trump administration could have far-reaching implications for the world
• these implications include, but are not limited to and in random order:
• deployment of autonomous weapons with no humans in the loop. when we already know that A.I's hallucinate and make mistakes all the time
• potential for mass domestic surveillance and perhaps world-wide also
• if the actions ought to be categorized as government overreach (I am not well versed to be able to say categorically that it is or is not government overreach).
• if DPA (defense production act) is invoked and effectively the code for anthropic is nationalized, then this could jeopardize their future. which includes their multi-billion IPO (initial public offering). as of yet, I have yet to review the DPA link
• considering that most western nations now rely on procurement for the defense (now war) related needs via acquisition. what happens to anthropic could set a precedent for other organizations

update:

• it looks like as I was typing this blogpost, anthropic was already beginning to scale back some of it's safety measures link
• that being said, I think the 5:01 eastern time deadline still looms. link so let's wait and see what happens. and meanwhile I can begin working on part 02 of the blogpost

thanks for reading. may saner minds prevail.