Synergy

I’m listening to a recording of Peter Meisen and company on the 52 Living Ideas channel.  Bear Island, Maine, duh, not Massachusetts. 

He is talking about the emergency situation in Ukraine, in the context of introducing the idea of Synergy. His focus is the energy budget, and the mix of resources, renewables versus coal.  Shrikant wants to get those already steeped in the Bucky stuff to expound upon these words “synergy” and “emergence”.

My focus these days has been Grunch dashboards, enterprise level instrumentation developed with tools such as dash and plotly, perhaps in tandem with pandas DataFrames.  We’re talking real time monitoring in many cases, of oil rigs for example.  I’m remembering a EuroPython talk.  The technology is a moving target.  I meandered over to Google Earth in one of my tabs, to talk about data visualization as seamless with these satellite painted global data displays.

Long time readers of this blog may know about my Global Data Corporation, which one might call science fictional.  The murky origins were not in Delaware, in “whiteman law” as I called it, but in native peoples (“tribes” is OK, I’m not at Stanford) learning the language of supranationalism, and hearing much of our nation-state talk as holdover Anglo imperialism.  I didn’t want Global Data in the lineage of the treaty breakers.

Thanks to the DSR (design science revolution), the core ideas and algorithms tend to percolate outward to the periphery.  Global Data is fringe in that way, over the horizon, outside the everyday jurisdiction of most business paradigms.  

Through computer science, linear algebra, machine learning and other STEM language games, we make a bridge to this alternative commercial sector shoptalk.  The language of open source interweaves.  As I was telling my students, there’s still plenty of room for private enterprise in the context of communally shared tools.  PayPal calls it “Inner Source” (bringing the communal way of doing things in-house and making it proprietary).  Cynics will cry “co-optation” but what’s free and open poses a non-trivial challenge, a form of competition that tends to cow the more cowardly capitalists.

The Anglos are used to exercising their right to contradict where foundational metaphysics is concerned.  To what extent is the “3Dness” of space a mere dogma?  Heresy.  

I talk about “hypercross dogmatics” and a quasi-religious streak in establishment science, in its embrace of Abbott’s Flatland and the mysteries of the tesseract (hyper cube). We boot a belief in the transcendent with the perpetuation of a priesthood, an elite, with a few from the laity, when we speak in hushed tones of multi-dimensionality.

Speaking of my confusion, regarding the timeline, the location of Bear Island, I'd also managed to scramble the order of the books we all talk about.  Tetrascroll came well before Critical Path, not after.

More Echo Chambers

We should revisit the image of "echo chamber", which I think is a good one, yet even more powerful if paired with other words such as "groupthink" and "choir", even "silo".  I'd make a great machine learner, or am one already, as Chelsea Manning suggested in Berlin that time.

Yes, I've been exploring the world (planet) of NLP, natural language processing.  The specialists have carved out a nifty niche, with a smooth vocab, kudos to University of Toronto.  

When I tell the story, I let Regular Polytopes hitch a ride on the ML bandwagon, because Graph Theory, because Linear Algebra.  I'm linking Regular Polytopes to Coxeter of course, also University of Toronto, and practitioner of hyper-dimensional Euclidean language games.

Did you know Coxeter lent his suite at Cambridge to the Wittgenstein posse, even after deciding what passed for hot philo was not his cup of tea.  Ludwig (LW) had entered his "second phase" by then ("language games"), if I'm not mistaken.  My source is the Siobhan Roberts Donald Coxeter bio:  The King of Infinite Space.  She did a later bio of Conway.

I'm not the expert in the room when it comes to NLP.  I do rub up against Machine Learning in my contemporary role on faculty, but more the way a high school teacher rubs against higher math, as an admirer and fan.  I'll lurk in on Andrius talking Sheffer Polynomials and even get deeper into his slides, but I'm not venturing as an equal into the ring.  I'm an apprentice at best in so many arenas.

My understanding is a breakthrough in the concepts of Attention and Self Attention (operationally defined) has enabled "next word predictors" to take more context into account, in addition to word order.  Would it be right to say a weighted semantic network develops internally, connecting higher attention words to their most frequent associates?  "Michelin" goes with "guide" "travel" and "tourist" as well as "tire".

Speaking of tourism and travel, at one point I was told, while gazing over the maze-like pattern of walls and chambers in the Colosseum in Rome, that I was seeing the ruins of a medieval castle that occupied the Roman stadium in a later century.  I believed that, but was recently talked into accepting these walls and chambers date to Roman times, while the castle story is also true (they're not mutually exclusive). 

Actually, my family, mother in particular, knows more about that castle than most, having taken an interest in Jacopa Frangipani, a luminary protagonist in her The Lions and The Lamb, an historical fiction novel, unpublished.  The Frangipanis owned the castle.  Jacopa was a friend and admirer of St. Francis.

You may have picked up on what I'm doing here:  showing off what "real intelligence" (vs the phony stuff) is still capable of, in contrast.  Everything may appear strung together haphazardly, but on deeper digging, everything checks out.  Eat your heart out GPT-3.