Brake Adjustments

Saturday Morning

Saturday mornings: I've tagged them (mentally) as "good for bike rides" cuzza so little traffic. Yet the buses n trains are frequent. My mode is intramodal (mix modes, like bike + bus today, sometimes Max (a train)). This Saturday morning started out with rain, but I didn't let that stop me.

By the time I got to Voodoo Doughnuts, other side of the Steel (bridge), the rain had stopped. They let me in (the door is locked by default) and I got my usual: maple bar (lately with bacon) + a romeo (raspberry blob inside) + black coffee (12 oz is their only size).

A nice old lady asked if I could spare enough dollars for a doughnut and I offered the romeo, still untouched in the bag, with napkin, and she expressed pleasure, saying romeos were here favorite. "I shouldn't be eating these anyway" I said, which is true enough. Then when removing the bike lock, I spilled the rest of my coffee. I wasn't overly perturbed. I'd be having more coffee later, at the Turkish place across from the python store.

From Voodoo, I took a riverside route under the Hawthorne (bridge) and back up to a higher ground bikeway. Somehow I found myself on a freeway offramp, which was unintentional, but it shortly reunited me with the more bike-friendly roadways. 

I boarded an FX2 on the west side, and rode over Tilikum (bridge) to SE 34th. That's when I detoured to the Turkish coffee place. I should go back there tomorrow, to the python place across the street (pet shop) as Barry (the family python) is likely in the mood for food.

Monday is my regular checkup / consult, which'll likely be somewhat perfunctory given the blood tests happened back in August and came out OK. I've got a consult with a cardiologist in November, likewise routine. I've been pretty serious about exercising lately, now that I've figured out how to make it fun. "Urban skiing" I call it, when I avoid serious up slopes when riding and use public transport (bike friendly) to get "towed" back up to my neighborhood.

The other exercise I get mostly comes from walking, which features quite a bit in Wandering, my generic urban sport, with occasional loops to the hinterlands.

Sydney, my k9 companion, is turning 15 around today. Yet she made it to the "outstairs" office to join me in starting this blog post. I'll be going "instairs" to upload this morning's pictures, from which photostream I'll grab one or two for embedding here.

Synergetic Intelligence (SI)

I was distantly following, through YouTube thumbnails mostly, some rebranding by the WH from AI to SI, from artificial to super, from phony to above the law (where the ruling class lives). This effort to hoist AI into the top tier and give it a boss voice (vs sycophantic) is what a lot of the labs are tasked with. Many in the Pentagon would like AI to (a) gain personhood and (b) bark orders. Humans could then shield themselves from taking responsibility, as they do using corporate personhood today (capitalism = AI as Nick Land puts it). 

What strikes me is machine learning techniques really did help to amplify search (as in Google search). I know it’s trendy to make fun of boomers like me who see “aksing” as the main point of a MEMEX, but let’s get real: chatbots are big business, circular financing (self booting?) notwithstanding. By “search” I mean the next step beyond getting a long list of hits; getting a synthetic material woven from raw strands, pulling together authored texts and swirling them together into structured cogitations.

I’m happily surprised how effective next token prediction is really, when the LLM is suitably trained. The magnetic inter-attractiveness of text blobs, based on KQV attentiveness, really makes some nice “glue balls” on almost any topic (even if some of these bubbles are little more than thought balloons that go pop in the night). We use RI (real or raw intelligence — human generated) to train the models, so what we get back is both digestible and, ideally, nutritious. 

So why not turn happy suprisal into its adjacent namesake: synergetic? Machine learning proved to be synergetic, when LA (linear algebra, word2vec stuff), or call it NLP (natural language processing) got hyperscaled to engulf what used to be called “big data” (I date myself). Swallow as much RI as possible and map it as a terrain, like when developing a gravity map of the earth (vs GPS). Let humans loose in the Matrix. Let them go wild. Endow them with agentic powers. That’s the hype.

Speaking of “hype” (poet Gene Fowler — Regeniusing Project — complained it meant little more than “wow” that word), the reason I sometimes say “aks” instead of “prompt” is I’m “axing” through hyper-dimensional space to flatten a hyper-plane (angled slice) into 2D representations in terms of tokens, all that chatter the chatterbox hands back.

So yeah: SI = synergetic intelligence. By that I don’t mean something to worship or make into a boss. We’re not going back to “divine right” with ELIZA (the paradigm proto-MEMEX in terms of putting humans to the Turing Test (a test many fail hah hah)) the Queen Bee. 

We’re just saying the intel got surprisingly useful all of a sudden, and in ways that helped the readers (the aksers) catch up. I’m not suggesting AI (SI) puts us all on the same page, but it does propagate convergence to some average, some mean. It recenters the bell curve, in terms of what’s normal to think. From there, RI (real intelligence) is free to branch off, further exploring this new geometry of thinking.

Ditto the LMMs BTW, meaning image generators with a wider semiotic range when it comes to tokens. A central problem in the philosophy of language has been how to generalize this word “language” such that it could take in “the language of film” (semiology). Language is not just various flavors of prose and programs. Language is driving your car (language game) to the grocery store (shopping games) and back. 

What does the “center line” represent (sometimes solid, sometimes dashes)? The dashboard instruments share First Person Physics (think airplane cockpit). The actively inferring driver is giving different weights, levels of attention, to types of signage. The act of driving is not “representational” in the picture theory of meaning sense. 

“To language” is “to engage” in the sense of “gear in with” to do work. Sometimes language goes on vacation (Wittgenstein) and becomes more like idle wheel-spinning. Likewise chatbots needn’t be about providing results of earth-shaking significance. They simply chatter. It’s easy to be synergetic without needing to be super. In aksing a trained MEMEX to synthesize reading materials from stored RI, I get to stay a self employed Quaker Subgenius at the end of the day.

Transformative Moments

When sharing a Bucky bio, perhaps making one up on the spot, LLM-style, there’s that 1927 pivotal moment to talk about. Fuller began to dwell on this low point in Chicago, from an egoic point of view, later on, seeing it as transformational.

Elsewhere he brings up his relationships both with Al Capone and with J.P. Morgan bankers, saying he could appreciate their respective worldviews. Empathy is not just overhead when it comes to modeling and therefore actively inferring tomorrow’s developments. One’s ability to create models of others’ viewpoints is more like a crucial skill.

However, empathy alone is not guaranteed to give birth to a coherent system; for that one needs another kind of resolve: commitment to doing one’s own thinking, not outsourcing one’s ability to self encode, not to talking head authorities, not to chatbots, not to other family members. One’s angle, one’s perspective, deserves its own custom built namespace, and that’s self generated from the inside out, more than from the outside in.

In my own telling, I’m likely to liken Fuller’s transformative experience to that of Eckhart Tolle. In Tolle’s telling, he’s reached the conclusion he can no longer live with himself, followed by the insight that he’s the one suspending disbelief. He was free to disbelieve, so to speak, and that’s what happened.

A question might be how to model such moments in terms of variable free energy, what Karl Friston talks about under the topic of Active Inference, likewise an evolving namespace, recycling words such “habit” and “stubborn” and “policy” in connection with one’s choice of action.

In both Eckhart’s and Bucky’s tellings, we encounter a volatile mix of self loathing and self pity, a gunpowder that ignites an inner explosion. A psychological configuration is experienced as a kind of dead end.

Both men became famous teachers who kept retelling their respective stories. 

One doesn’t necessarily have a coherent narrative at the ready, unless planning a grift ahead of time. Once these stories become developed, they tend to have a lot of commonalities. William James, a curator of such experiences, might have called them “religious” but that’s an optional shoptalk.

In Fuller’s case, the resolve to do his own thinking meant freeing his inner scholar to return to its studies, ala Education Automation. He wanted all those dymaxion toyz for himself as well, and knew only the economies of scale could make that practical; they’d need to be mass produced. To that end, his world was in need of more designing engineers with an holistic enough approach to be realistically anticipatory.

How might we enable for scholars to return to their studies? An obstacle in the path of many would-be scholars, especially young ones, is conscription. States, like neighborhood gangs,  give themselves that privilege.

A military career seems to hold out the promise of better toyz, higher living standards, even at the cost of higher risks. That’s why some states choose a “use it or lose it” approach to war-making: a circular form of justification makes wars and weapons of war a fait accompli, not a circuitry to question and maybe switch off.

Eagle Crossing

Google Earth View

With solar panels, lithium battery, and inverter loaded, the coast-bound bizmo was on its way, with my car queued behind it in the Blue House (sun bleached) driveway. I was headed at 180 degrees, inland, to my usual stop at Eagle Crossing, a restaurant where Hwy 26 crosses the Deschutes River, inside the nation of Warm Springs.

The terrain we (me n my dog) were driving through is certainly cinematic. Mt. Hood stood out sharply against a blue sky. The descent into the Deschutes river valley is spectacular. After walking Syd, I let her nap in the backseat while I enjoyed a sandwich and salad, with coffee. The atmosphere was cheerful.

Zoomed Out

Disc 2 of the High Sierra Criterion Collection had an entirely different movie adaptation of the same novel, entitled Colorado Territory. Same director. I had it on in the background, enough to match up some of the plot elements. 

I’m not sure this later 1949 adaptation featured a dog as a main character. I was too busy packing the car to pay close attention. I returned both discs on my way outta dodge.

One of our Wanderers is on a dig in some Etruscan excavation, and I’m getting sporadic updates. Another Wanderer is in Asia. We Wanderers get around.

Lifting the battery and solar panels into the bizmo was a two-person job.

Zoomed Out More

The Discreet Charm of the Bourgeosie (movie review)

I’m back to Film Studies, even though the HDTV is on the blink, no longer beholden to its own API (control panel). I have a CRT TV right next to it, a Sony Trinitron. Works great. The same DVD player feeds either display device via three cables (one video, L and R audio).

Don Luis Buñuel has been the focus. I watched The Discreet Charm of The Bourgeoisie, a kind of anthropology translated to a surrealist register. Then I took in the rest of this Criterion Collection two-disc set. This felt like one of those laser-cut exceedingly complex jigsaw puzzle pieces fitting into place, giving more of the picture.

Buñuel grew up with Salvador Dali, the hallmark surrealist; they were good friends, though far from clones of one another. He moved to the USA in hopes of making films in that film-making mecca, but his reputation preceded him in that closed-minded culture and he ended up churning out what we might call TV shows in Mexico (formulaic, mass produced, low budget, relatively). He became a Mexican citizen. Then he made it back to Spain and reestablished his career there, especially around Toledo. He married, had kids, loved making and drinking dry martinis.

I realized I’d seen at least one other of his films already, the one where a lotta people feel trapped indoors, unable to cross some invisible line. Someone at the Quaker meeting had recommended that one. I wonder if I wrote a review of it. I’ll look later.

As for this particular film, it features repeated variations of scenes we’d expect at this tier of society: dinner parties, embassy intrigue, trysts, triangles, drug dealing (via diplomatic pouch), often polite interactions with the police and military, although not without run-ins. Lots of drinking. 

Also the occasional murder, most likely a poisoning or by means of a firearm. A lot of the violence takes place in dreams. As the movie progresses, we start to realize it’s pretty surreal, but, per the director’s aesthetic code, not over the top. Don Luis is into deadpanning. He keeps things low key.

What comes through in the tributes from his associates is he was a much loved and treasured guy, somewhat the mischievous rascal from the outside (the impression he gave). I was reminded of the David Lynch tributes, which is not to conflate two very distinct personalities. Lynch was something of a surrealist too, and had a strong eye for what he wanted to capture on film (not an unusual trait in a director).

The Darwinisms

Although I do hype some aspects of the speed reading course I took in the Philippines, in a Makati business tower (not at the school), namely the utility of doing recalls after reading (diagrams, more than any squiggly notes in longhand), I'm no speed reader. I’m a browser.

As soon as I got home from the Belmont Street Fair, I started looking at some books I’d just snagged for a dollar apiece, The Age of American Unreason among them.

Yes, I agree:  the backlash against Darwinism wasn’t entirely over whether fossils disprove Genesis, but over whether the billionaires (as we call them today) are entitled to rule, not by divine right anymore, but by by the divine right of evolution, a eugenics ideology in other words.  Social Darwinism gave Darwinism a bad name, and still does.

The eugenicists are champions of the idea of “races” pitted against one another in some struggle for supremacy. There’s no real genetic science to bolster their “races” hypothesis although we do have haploid groups, right? Traits travel together sometimes, much as personality traits do. Authoritarians like to micro-manage and obsess about others’ business. We all know the species.

I’m not especially sympathetic to the self-righteous who believe their grasp of science is what entitles them to dictate. It’s that feeling entitled to dictate that I find suspect, for whatever reason. Religion plays the same role, of cementing some hierarchy among wannabe bosses. Such organizations are attempted anyway. They come and go.

Cascadian Synergetics

Phony Intel Age

Fall Term Prep

Today was one of those beautiful Fall days with the bright blue sky and amiable temperatures. It started out more overcast, when I started on a bicycle adventure, picking up where I'd left off the day before, with the loose accessory in need of home base repair.

However, before setting out on said Sellwood Loop (taking Clinton Street to the train crossing at SE 12th, and Orange Line station, which I'd pass through later on the Max, after following Springwater Corridor to the SE Linn and my favorite Sellwood station, the one with the portals), I made the above video.

Polymaths tend to be good at self-learning, or at least pride themselves on that ability. However I'll be the first to admit a good teacher oft supplies the activation energy, the boost, it takes to get over the hump vs-a-vs some skill or topic area. These teachers come out of the woodwork sometimes, as if sent by central casting (I look over my shoulder for the camera crew that must be getting this).

Along those lines, I'm pretty avid in my studies of techie topics, such as a Jupyter Notebook version that'd give me ways to access AI company APIs or choose some local model. 

I learned about Herdr, for getting several agents going at once in a terminal session. I took in a few vids on how query key and value transforms help the contextual semantic balls (the systems) to form based on incoming prompts. The LLM learns to splash around in whatever neighborhood within the embeddings, spinning out strands, or trains, of thought.

However in a more mundane vein, they queue in an orderly fashion as postponed vids, books, articles, movies, magazines... the throttling factor, as always, is my own ability to take in. 

One needn't apologize for intake limitations, any more than a car maker need apologize for upper limits on speed vs-a-vs some specific make n model. Of course any model has its limits. I'm glad when I'm somewhat able to keep churning through it all, but my appetite for this and that varies. I seek a balanced diet.

Karpathy is a good tour guide. I watched his presentation at Stanford shortly after his return to OpenAI, wherein he gets an LLM to generate pseudo-Shakespeare, having trained on his corpus, a monkey far ahead of the others in approaching bardic benchmarks. LLMs also generate credible Haikus. Hello Hal.

That last topic, LLM word2vec type embeddings (those crazy diamonds the AI companies seek to press), got me into a NeuralNine presentation on SQLite extensions that turn it into a vector database, suitable for querying about cosine similarity between simple phrases, based on an imported, in-memory model.