Full Width 
Shortcuts 
Sign Up 
Log In 
Articles 
Reading List More from Jorge Arango
Do you ever catch yourself avoiding things you need to do? Sure you, do: we all do it. In episode 9 of Traction Heroes, Harry and I discuss what to do about it. The conversation took off when Harry read a fragment from Oliver Burkeman’s book, Meditations for Mortals. I won’t cite the entire passage here, but this gives you a taste: It can be alarming to realize just how much of life gets shaped by what we’re actively trying to avoid. We talk about not getting around to things as if it were merely a failure of organization or a will. But often the truth is that we invest plenty of energy in making sure that we never get around to them. … The more you organize your life around not addressing things that make you anxious, the more likely they are to develop into serious problems. And even if they don’t, the longer you fail to confront them, the more unhappy time you spend being scared of what might be lurking in the places you don’t want to go. The irony, of course, is that we put off uncomfortable tasks because they make us anxious. But putting them off ultimately makes us more anxious. As Harry reminded us, “bad news doesn’t get better over time.” He also proposed a helpful framing: that facts are friendly. That is, even though knowing the truth might make us uncomfortable, knowing is better than not knowing. We discussed practical steps to gain traction: Ask yourself, what am I pretending not to know? Deep down, you know there’s more to the situation than you’ve let on; acknowledging the elephant in the room to move forward. Plan around the last responsible moment. Some events have fixed time windows; understand by when you must decide. Rewrite the narrative using the non-violent communication lens: separate your observations from interpretations, feelings, and needs. As always, I got lots of value from this conversation with Harry. But this one you can’t think about; it’s about doing. And doing is hard when the mind doesn’t want to face facts. Traction Heroes episode 9: Procrastination
In week 18 of the humanities crash course, I read five stories from One Thousand and One Nights, a collection of Middle Eastern folktales that have influenced lots of other stories. Keeping with the theme, I also saw one of the most influential movies based on these stories. Readings An influential collection of Middle Eastern folk tales compiled during the Islamic Golden Age. The framing device is brutally misogynistic: a sultan learns that his wife is unfaithful, so he executes her. He decides all women are the same, so he marries a new bride every day and has her executed the following day. Sheherazade asks her father, the vizier, to offer her in marriage to the sultan. The vizier is reluctant: they both know the wives’ fate. But Sheherazade has a clever plan: she starts a new story for the sultan every night but leaves it in a cliffhanger. Curious for the outcome, the sultan stays her execution to the next day. In this way, Sheherazade spares the lives of other maidens of the land. Of the many stories in the book, I read five recommended by Gioia: The Fisherman and the Jinni: a poor fisherman unwittingly unleashes a murderous jinni from a bottle, but tricks him back into the bottle by outwitting him. The Three Apples: an ancient murder mystery (again, centered on the murder of an innocent woman); the “solution” involves more unjust death (at least by our standards.) Sinbad the Sailor: a series of seven fantastical voyages involving monsters, magic, and stolen treasures; one of the voyages closely parallels the Cyclops episode from the Odyssey. Ali Baba and the Forty Thieves: another story of murder and ill-gotten treasure; a poor man discovers where a band of thieve stashes their loot and steals from them. Aladdin: a poor boy discovers a magic lamp that makes him wealthy and powerful, allowing him to marry a princess. These have been re-told in numerous guises. As often happens in these cases, the originals are much darker and bloodier than their spawn. These aren’t Disney versions, for sure. Audiovisual Music: Highlights from Tchaikovsky’s famous ballets plus Rimsky-Korsakov’s Sheherazade. I’d heard the ballets, but not the Rimsky-Korsakov. This piece reminded me of Paul Smith’s music for Disney’s 20,000 LEAGUES UNDER THE SEA (1954). Arts: Gioia recommended aboriginal Australian art. I’d seen works in this style, but hadn’t paid attention. This tradtion has a particular (and gorgeous) style that expresses strong connections to the land. I was surprised to learn about recent developments in this tradition. Cinema: Alexander Korba’s THE THIEF OF BAGDAD (1940), one of the many films ispired by the One Thousand and One Nights. While it now looks dated, this film was a special effects breakthrough. As an early example of Technicolor, it also features an over-the-top palette, much like it’s near-contemporary, THE WIZARD OF OZ. Reflections One can’t do justice to One Thousand and One Nights by only reading five stories. But the ones I read dealt with poor people being unfairly granted wealth and power. Escapist fantasies tend to stand the test of time. The “heroes” in the stories deserved as much comeuppance as the “villains.” For example, in Ali Baba and the Forty Thieves), one of the heroes commits a mass killing of the “bad guys” while they were unable to react. Not only does this go unpunished; it’s celebrated. The people who told these stories had moral standards different from our own. I also learned several stories — including some of the most famous, such as Ali Baba and the Forty Thieves and Aladdin — were not part of the original collection. Instead, they were added by a French translator in the 18th Century. This was frustrating, as they weren’t present in the collection I bought; I had to seek them out separately. So, this week, I’ve been pondering questions of authorship and derivation. We don’t know who originated these stories. Like the aboriginal Australian art, the stories in the One Thousand and One Nights emerged — and belong to — a people more than an individual author or artist. And yet, they’ve inspired other works, such as THE THIEF OF BAGDAD — which inspired Disney’s ALADDIN. (The latter “borrows” liberally from the former.) Is it any wonder I heard Rimsky-Korsakov in the 20k score? At this point, I assume at least some cross-pollination — after all, Rimsky-Korsakov himself was inspired by the One Thousand and One Nights. This is how art has always evolved: artists build on what’s come before. In some cases, the inspiration is obvious. In others, it’s more nebulous. Did Odysseus inspire Simbad? Or did they both retell older stories? The process changed in the 20th Century. With strong copyright laws, stories become intellectual property. Disney may build on the One Thousand and One Nights stories, but we can’t build on Disney’s stories. And it’s changing again with large language models. It will be interesting to see how these new tools allow us to retell old stories in new ways. At a minimum, they’re causing us to reevaluate our approach to IP. Notes on Note-taking A realization: my Obsidian knowledge repository is better suited to reflecting on text than other media. I can try to write down my impressions of the beautiful aboriginal art and Rimsky-Korsakov’s music. But words articulate concepts, not feelings — even when trying to articulate feelings. So I end up reflecting on abstract ideas such as authorship and derivation rather than the nature of the works. It’s a limitation of my current note-taking system, and one I can’t do much about. Perhaps ChatGPT can help by letting me riff on pictures and sounds? But there, too, communication happens through language. Up Next Gioia recommends the Bhagavad Gita, the Rule of St. Benedict, and the first two books of Saint Augustine’s Confessions. This will be my first time with any of them. Again, there’s a YouTube playlist for the videos I’m sharing here. I’m also sharing these posts via Substack if you’d like to subscribe and comment. See you next week!
The dream is running GraphRAG with locally-hosted LLMs. And at least for now, the dream is on hold for me. In case you missed it, GraphRAG is a way of getting more useful results with LLMs by working with data you provide (in addition to whatever they’ve trained on.) The system uses LLMs to build a knowledge graph from documents you provide and then uses those graphs to power RAG queries. This opens lots of possibilities. For information architecture work, it lets you ask useful questions of your own content. I’ve written about my experiments in that scenario. In that case, I used OpenAI’s models to power Microsoft’s GraphRAG application. But I’m especially excited about the possibilities for personal knowledge management. Imagine an LLM tuned to and focused on your personal notes, journals, calendars, etc. That’s primarily why I’m dreaming of GraphRAG powered by local models. There are several reasons why local models would be preferable. For one, there’s the cost: GraphRAG indexing runs are expensive. There’s also a privacy angle. Yes, I’ve told OpenAI I don’t want them to train their models using my data, but some of this stuff is extremely personal and I’m not comfortable with it leaving my computer at all. But an even larger concern is dependency. I’m building a lifelong thinking assistant. (An amanuensis, as I outlined in Duly Noted.) It’s risky to delegate such a central part of this system to a party that could turn off the spigot at any time. So I’ve been experimenting with graphrag using local models. There are good news and bad news. Before I tell you about them, let me explain my setup. I’m using a 16” 2023 M2 Max MacBook Pro with 32GB of RAM. It’s not an entry-level machine, but not a monster either. I’m using ollama to run local models. I’ve tried around half a dozen at this point and have successfully set up one automated (non-GraphRAG) workflow using mistral-small3.1. GraphRAG is extremely flexible. There are dozens of parameters to configure, including different LLMs for each step in the process. Off-the-shelf, its prompts are optimized specifically for GPT-4-turbo; other models require tweaking. Indexing runs (where the model converts texts to knowledge graphs) can take a long time. So tweaks are time-consuming. I’ve had a go at it several times, but given up after a bit. I don’t have much free time these days, and most experiments have unsuccessfully ended with failed (and long!) indexing runs. But a few things have changed in recent weeks: GraphRAG itself keeps evolving There are now more powerful small local models that run better within my machine’s limitations ChatGPT o3 came out That last one may sound like a non-sequitur. Aren’t I trying to get away from cloud-hosted models for this use case? Well, yes — but in this case, I’m not using o3 to power GraphRAG. Instead, I’m using it to help me debug failed runs. While certainly nothing like AGI, as some have claimed, o3 has proven to be excellent for dealing with the sort of tech-related issues that would’ve sent me off to Stack Overflow in the past. Debugging GraphRAG runs is one such task. I’ve been feeding o3 logfiles after each run, and it’s recommended helpful tweaks. It’s the single most important factor in my recent progress. Yes, there’s been some progress: yesterday, after many tries, I finally got two local models to successfully complete an indexing run. Mind you, that doesn’t mean I can yet successfully query GraphRAG. But finishing the indexing run without issues is progress. That’s the good news. Alas, the indexing run took around thirty-six hours to process nineteen relatively short Markdown files. To put that in perspective, the same indexing run using cloud-hosted models would likely have taken under ten minutes. My machine also ran at full throttle the whole time. (It’s the first time I’ve felt an M-series Mac get hot.) The reduced processing speed isn’t just because the models themselves are slower: it’s also due to my machine’s limitations. After analyzing the log files, ChatGPT suggested reducing the number of concurrent API calls. The successful run specified just one call at a time for both models. The upshot is that even though the indexing run finished successfully, this process is impractical for real-world use. My PKM has thousands of Markdown files. ChatGPT keeps suggesting further tweaks, but progress is frustratingly slow when cycles are measured in days. I’ve considered upgrading to a MBP with more RAM or increasing the number of concurrent processes to find the upper threshold for my machine. But based on these results, I suspect improvements will be marginal given the amount of data I’m looking to process. So that’s the bad news. For now, I’ll keep working with local models for other uses (such as OCRing handwritten notes; the workflow I alluded to above. More on that soon!) And of course, I’ll continue experimenting with cloud-based models for other use cases. In any case, I’ll share what I learn here.
In week 17 of the humanities crash course, I read a book that was completely new to me: Apuleius’s Metamorphoses, better known as The Golden Ass. I also watched a movie with a similar story (but with different aims.) Readings The Golden Ass was written by Apuleius around the second century CE. The only complete Latin novel to survive, it tells the story of Lucius, a man whose reckless curiositas leads him to accidentally be transformed into an ass. (What is curiositas, you ask? Read on…) As a donkey, Lucius goes from owner to owner, exposing him to dangers, adventure, and gossip. Characters tell several sub-stories, mostly about crime, infidelity, and magic. The most famous is the story of Cupid and Psyche, a cautionary allegory that echoes the themes and structures of the novel as a whole. Throughout his wanderings, Lucius is treated brutally. At one point a woman falls in love with him and treats him as a sex object. Eventually, the goddess Isis brings him back to human form after an initiation into her cult. He becomes an acolyte, making the story a metaphor for religious conversion. The final section of the book, where Lucius undergoes his spiritual transformation, is one of several surprising tone shifts: the book is sometimes drama, horror, fairy tale, and bawdy farce. Overall, it gives an entertaining picture of moral codes in second century Europe. Audiovisual Music: Scott Joplin. Again, a composer whose work was familiar to me. Rather than the usual piano solo versions, I listened to a recording of his works featuring Andre Previn on piano and Itzhak Perlman on violin. Arts: van Gogh, who, like Joplin, is overly familiar. This lecture from The National Gallery helped put his work in context: I hadn’t realized the degree to which van Gogh’s paintings are the result of a tech innovation: synthetic pigments in the newly invented roll-up tubes. As always, understanding context is essential. Cinema: Jerzy Skolimowski’s EO, a road picture that follows a donkey as he drifts through the Polish and Italian countrysides. Like Lucius, he’s exposed to humanity’s moral failings (and a tiny bit of tenderness.) While visually and aurally stunning, I found the movie overbearingly preachy. Reflections As usual, I entered my reflections on the book into ChatGPT to ask for what I might have missed or gotten wrong. My notes said Lucius’s curiosity about witchcraft led him to be transformed into an ass. ChatGPT corrected me: it wasn’t curiosity but curiositas. I asked for clarification, since the two terms are so similar. As I now understand it, curiositas refers to “an immoderate appetite for forbidden or frivolous knowledge that distracts from real duties” — i.e., wasting time on B.S. of the sort one finds in tabloids or chasing after forbidden knowledge. ChatGPT suggested as contemporary equivalents clickbait and doomscrolling, gossip culture (think the Kardashians), and “risk-blind experimentation” — i.e., the “move fast and break things” ethos — as the LLM put it, a “reckless desire to test the limits without counting the costs.” In other words, Lucius wasn’t punished (and ultimately disciplined) because he was curious. Instead, he “messed around and found out” — literally making an ass out of himself. For the ancients, the healthy opposite was studiositas, a “disciplined study in service of truth.” We’ll spend time with Thomas Aquinas later in the course; ChatGPT suggests he makes much of this distinction. Notes on Note-taking Last week, I said I’d return to ChatGPT 4o for its responsiveness. I haven’t; the o3 model’s results are better enough that the slightly longer wait is worth it. That said, I remain disappointed with o3’s preference for tables. One good sign: at one point, ChatGPT presented me with a brief A/B test where it asked me to pick between a table-based result and one with more traditional prose. Of course, I picked the latter. I hope they do away with the tables, or at least make them much less frequent. Up Next Gioia recommends selected readings from The Arabian Nights. While I’ve never read the original, several of these stories (Aladdin, Sinbad) are familiar through reinterpretations. I’m looking forward to reading the originals. Again, there’s a YouTube playlist for the videos I’m sharing here. I’m also sharing these posts via Substack if you’d like to subscribe and comment. See you next week!
There’s a lot of turbulence in the world. What is the source of the turbulence? And how can we navigate skillfully? These questions were on my mind as I met with Harry to record episode 8 of the Traction Heroes podcast. My (at least partial) answer to the first question is that there’s a general lack of systems literacy in the world. Most people aren’t aware of the high degree of complexity that characterizes highly intertwingled systems such as modern economies. As a result, they opt for simplistic interventions that often do more harm than good. At least that was my hypothesis. I was keen to hear Harry’s thoughts — and he didn’t disappoint. My prompt was the following passage from Donella Meadows’s classic Thinking in Systems: A Primer (emphasis in the original): Ever since the Industrial Revolution, Western society has benefited from science, logic, and reductionism over intuition and holism. Psychologically and politically we would much rather assume that the cause of a problem is “out there,” rather than “in here.” It’s almost irresistible to blame something or someone else, to shift responsibility away from ourselves, and to look for the control knob, the product, the pill, the technical fix that will make a problem go away. Serious problems have been solved by focusing on external agents—preventing smallpox, increasing food production, moving large weights and many people rapidly over long distances. Because they are embedded in larger systems, however, some of our “solutions” have created further problems. And some problems, those most rooted in the internal structure of complex systems, the real messes, have refused to go away. Hunger, poverty, environmental degradation, economic instability, unemployment, chronic disease, drug addiction, and war, for example, persist in spite of the analytical ability and technical brilliance that have been directed toward eradicating them. No one deliberately creates those problems, no one wants them to persist, but they persist nonetheless. That is because they are intrinsically systems problems—undesirable behaviors characteristic of the system structures that produce them. They will yield only as we reclaim our intuition, stop casting blame, see the system as the source of its own problems, and find the courage and wisdom to restructure it. Of course, the broader context was (and is) on my mind. But we’re all enmeshed in complex systems in our day-to-day lives. It behooves us to ponder whether the causes of problems are really “out there” — or whether, as Harry suggested, we need to be more introspective. Traction Heroes ep. 8: Quagmires
More in technology
There are a handful of instruments that are staples of modern music, like guitars and pianos. And then there are hundreds of other instruments that were invented throughout history and then fell into obscurity without much notice. The Luminaphone, invented by Harry Grindell Matthews and unveiled in 1925, is a particularly bizarre example. Few people […] The post Recreating a bizarre century-old electronic instrument appeared first on Arduino Blog.
Sometimes I think I should pivot my career to home automation critic, because I have many opinions on the state of the home automation industry---and they're pretty much all critical. Virtually every time I bring up home automation, someone says something about the superiority of the light switch. Controlling lights is one of the most obvious applications of home automation, and there is a roughly century long history of developments in light control---yet, paradoxically, it is an area where consumer home automation continues to struggle. An analysis of how and why billion-dollar tech companies fail to master the simple toggling of lights in response to human input will have to wait for a future article, because I will have a hard time writing one without descending into incoherent sobbing about the principles of scene control and the interests of capital. Instead, I want to just dip a toe into the troubled waters of "smart lighting" by looking at one of its earliest precedents: low-voltage lighting control. A source I generally trust, the venerable "old internet" website Inspectapedia, says that low-voltage lighting control systems date back to about 1946. The earliest conclusive evidence I can find of these systems is a newspaper ad from 1948, but let's be honest, it's a holiday and I'm only making a half effort on the research. In any case, the post-war timing is not a coincidence. The late 1940s were a period of both rapid (sub)urban expansion and high copper prices, and the original impetus for relay systems seems to have been the confluence of these two. But let's step back and explain what a relay or low-voltage lighting control system is. First, I am not referring to "low voltage lighting" meaning lights that run on 12 or 24 volts DC or AC, as was common in landscape lighting and is increasingly common today for integrated LED lighting. Low-voltage lighting control systems are used for conventional 120VAC lights. In the most traditional construction, e.g. in the 1940s, lights would be served by a "hot" wire that passed through a wall box containing a switch. In many cases the neutral (likely shared with other fixtures) went directly from the light back to the panel, bypassing the switch... running both the hot and neutral through the switch box did not become conventional until fairly recently, to the chagrin of anyone installing switches that require a neutral for their own power, like timers or "smart" switches. The problem with this is that it lengthens the wiring runs. If you have a ceiling fixture with two different switches in a three-way arrangement, say in a hallway in a larger house, you could be adding nearly 100' in additional wire to get the hot to the switches and the runner between them. The cost of that wiring, in the mid-century, was quite substantial. Considering how difficult it is to find an employee to unlock the Romex cage at Lowes these days, I'm not sure that's changed that much. There are different ways of dealing with this. In the UK, the "ring main" served in part to reduce the gauge (and thus cost) of outlet wiring, but we never picked up that particular eccentricity in the US (for good reason). In commercial buildings, it's not unusual for lighting to run on 240v for similar reasons, but 240v is discouraged in US residential wiring. Besides, the mid-century was an age of optimism and ambition in electrical technology, the days of Total Electric Living. Perhaps the technology of the relay, refined by so many innovations of WWII, could offer a solution. Switch wiring also had to run through wall cavities, an irritating requirement in single-floor houses where much of the lighting wiring could be contained to the attic. The wiring of four-way and other multi-switch arrangements could become complex and require a lot more wall runs, discouraging builders providing switches in the most convenient places. What if relays also made multiple switches significantly easier to install and relocate? You probably get the idea. In a typical low-voltage lighting control system, a transformer provides a low voltage like 24VAC, much the same as used by doorbells. The light switches simply toggle the 24VAC control power to the coils of relays. Some (generally older) systems powered the relay continuously, but most used latching relays. In this case, all light switches are momentary, with an "on" side and an "off" side. This could be a paddle that you push up or down (much like a conventional light switch), a bar that you push the left or right sides of, or a pair of two push buttons. In most installations, all of the relays were installed together in a single enclosure, usually in the attic where the high-voltage wiring to the actual lights would be fairly short. The 24VAC cabling to the switches was much smaller gauge, and depending on the jurisdiction might not require any sort of license to install. Many systems had enclosures with separate high voltage and low voltage components, or mounted the relays on the outside of an enclosure such that the high voltage wiring was inside and low voltage outside. Both arrangements helped to meet code requirements for isolating high and low voltage systems and provided a margin of safety in the low voltage wiring. That provided additional cost savings as well; low voltage wiring was usually installed without any kind of conduit or sheathed cable. By 1950, relay lighting controls were making common appearances in real estate listings. A feature piece on the "Melody House," a builder's model home, in the Tacoma News Tribune reads thus: Newest features in the house are the low voltage touch plate and relay system lighting controls, with wide plates instead of snap buttons---operated like the stops of a pipe organ, with the merest flick of a finger. The comparison to a pipe organ is interesting, first in its assumption that many readers were familiar with typical organ stops. Pipe organs were, increasingly, one of the technological marvels of the era: while the concept of the pipe organ is very old, this same era saw electrical control systems (replete with relays!) significantly reduce the cost and complexity of organ consoles. What's more, the tonewheel electric organ had become well-developed and started to find its way into homes. The comparison is also interesting because of its deficiencies. The Touch-Plate system described used wide bars, which you pressed the left or right side of---you could call them momentary SPDT rocker switches if you wanted. There were organs with similar rocker stops but I do not think they were common in 1950. My experience is that such rocker switch stops usually indicate a fully digital control system, where they make momentary action unobtrusive and avoid state synchronization problems. I am far from an expert on organs, though, which is why I haven't yet written about them. If you have a guess at which type of pipe organ console our journalist was familiar with, do let me know. Touch-Plate seems to have been one of the first manufacturers of these systems, although I can't say for sure that they invented them. Interestingly, Touch-Plate is still around today, but their badly broken WordPress site ("Welcome to the new touch-plate.com" despite it actually being touchplate.com) suggests they may not do much business. After a few pageloads their WordPress plugin WAF blocked me for "exceed[ing] the maximum number of page not found errors per minute for humans." This might be related to my frustration that none of the product images load. It seems that the Touch-Plate company has mostly pivoted to reselling imported LED lighting (touchplateled.com), so I suppose the controls business is withering on the vine. The 1950s saw a proliferation of relay lighting control brands, with GE introducing a particularly popular system with several generations of fixtures. Kyle Switch Plates, who sell replacement switch plates (what else?), list options for Remcon, Sierra, Bryant, Pyramid, Douglas, and Enercon systems in addition to the two brands we have met so far. As someone who pays a little too much attention to light switches, I have personally seen four of these brands, three of them still in use and one apparently abandoned in place. Now, you might be thinking that simply economizing wiring by relocating the switches does not constitute "home automation," but there are other features to consider. For one, low-voltage light control systems made it feasible to install a lot more switches. Houses originally built with them often go a little wild with the n-way switching, every room providing lightswitches at every door. But there is also the possibility of relay logic. From the same article: The necessary switches are found in every room, but in the master bedroom there is a master control panel above the bed, from where the house and yard may be flooded with instant light in case of night emergency. Such "master control panels" were a big attraction for relay lighting, and the finest homes of the 1950s and 1960s often displayed either a grid of buttons near the head of the master bed, or even better, a GE "Master Selector" with a curious system of rotary switches. On later systems, timers often served as auxiliary switches, so you could schedule exterior lights. With a creative installer, "scenes" were even possible by wiring switches to arbitrary sets of relays (this required DC or half-wave rectified control power and diodes to isolate the switches from each other). Many of these relay control systems are still in use today. While they are quite outdated in a certain sense, the design is robust and the simple components mean that it's usually not difficult to find replacement parts when something does fail. The most popular system is the one offered by GE, using their RR series relays (RR3, RR4, etc., to the modern RR9). That said, GE suggests a modernization path to their LightSweep system, which is really a 0-10v analog dimming controller that has the add-on ability to operate relays. The failure modes are mostly what you would expect: low voltage wiring can chafe and short, or the switches can become stuck. This tends to cause the lights to stick on or off, and the continuous current through the relay coil often burns it out. The fix requires finding the stuck switch or short and correcting it, and then replacing the relay. One upside of these systems that persists today is density: the low voltage switches are small, so with most systems you can fit 3 per gang. Another is that they still make N-way switching easier. There is arguably a safety benefit, considering the reduction in mains-voltage wire runs. Yet we rarely see such a thing installed in homes newer than around the '80s. I don't know that I can give a definitive explanation of the decline of relay lighting control, but reduced prices for copper wiring were probably a main factor. The relays added a failure point, which might lead to a perception of unreliability, and the declining familiarity of electricians means that installing a relay system could be expensive and frustrating today. What really interests me about relay systems is that they weren't really replaced... the idea just went away. It's not like modern homes are providing a master control panel in the bedroom using some alternative technology. I mean, some do, those with prices in the eight digits, but you'll hardly ever see it. That gets us to the tension between residential lighting and architectural lighting control systems. In higher-end commercial buildings, and in environments like conference rooms and lecture halls, there's a well established industry building digital lighting control systems. Today, DALI is a common standard for the actual lighting control, but if you look at a range of existing buildings you will find everything from completely proprietary digital distributed dimming to 0-10v analog dimming to central dimmer racks (similar to traditional theatrical lighting). Relay lighting systems were, in a way, a nascent version of residential architectural lighting control. And the architectural lighting control industry continues to evolve. If there is a modern equivalent to relay lighting, it's something like Lutron QSX. That's a proprietary digital lighting (and shade) control system, marketed for both residential and commercial use. QSX offers a wide range of attractive wall controls, tight integration to Lutron's HomeSense home automation platform, and a price tag that'll make your eyes water. Lutron has produced many generations of these systems, and you could make an argument that they trace their heritage back to the relay systems of the 1940s. But they're just priced way beyond the middle-class home. And, well, I suppose that requires an argument based on economics. Prices have gone up. Despite tract construction being a much older idea than people often realize, it seems clear that today's new construction homes have been "value engineered" to significantly lower feature and quality levels than those of the mid-century---but they're a lot bigger. There is a sort of maxim that today's home buyers don't care about anything but square footage, and if you've seen what Pulte or D. R. Horton are putting up... well, I never knew that 3,000 sq ft could come so cheap, and look it too. Modern new-construction homes just don't come with the gizmos that older ones did, especially in the '60s and '70s. Looking at the sales brochure for a new development in my own Albuquerque ("Estates at La Cuentista"), besides 21st century suburbanization (Gated Community! "East Access to Paseo del Norte" as if that's a good thing!) most of the advertised features are "big." I'm serious! If you look at the "More Innovation Built In" section, the "innovations" are a home office (more square footage), storage (more square footage), indoor and outdoor gathering spaces (to be fair, only the indoor ones are square footage), "dedicated learning areas" for kids (more square footage), and a "basement or bigger garage" for a home gym (more square footage). The only thing in the entire innovation section that I would call a "technical" feature is water filtration. You can scroll down for more details, and you get to things like "space for a movie room" and a finished basement described eight different ways. Things were different during the peak of relay lighting in the '60s. A house might only be 1,600 sq ft, but the builder would deck it out with an intercom (including multi-room audio of a primitive sort), burglar alarm, and yes, relay lighting. All of these technologies were a lot newer and people were more excited about them; I bring up Total Electric Living a lot because of an aesthetic obsession but it was a large-scale advertising and partnership campaign by the electrical industry (particularly Westinghouse) that gave builders additional cross-promotion if they included all of these bells and whistles. Remember, that was when people were watching those old videos about the "kitchen of the future." What would a 2025 "Kitchen of the Future" promotional film emphasize? An island bigger than my living room and a nook for every meal, I assume. Features like intercoms and even burglar alarms have become far less common in new construction, and even if they were present I don't think most buyers would use them. But that might seem a little odd, right, given the push towards home automation? Well, built-in home automation options have existed for longer than any of today's consumer solutions, but "built in" is a liability for a technology product. There are practical reasons, in that built-in equipment is harder to replace, but there's also a lamer commercial reason. Consumer technology companies want to sell their products like consumer technology, so they've recontextualized lighting control as "IoT" and "smart" and "AI" rather than something an electrician would hook up. While I was looking into relay lighting control systems, I ran into an interesting example. The Lutron Lu Master Lumi 5. What a name! Lutron loves naming things like this. The Lumi 5 is a 1980s era product with essentially the same features as a relay system, but architected in a much stranger way. It is, essentially, five three way switches in a box with remote controls. That means that each of the actual light switches in the house (which could also be dimmers) need mains-voltage wiring, including runner, back to the Lumi 5 "interface." Pressing a button on one of the Lutron wall panels toggles the state of the relay in the "interface" cabinet, toggling the light. But, since it's all wired as a three-way switch, toggling the physical switch at the light does the same thing. As is typical when combining n-way switches and dimming, the Lumi 5 has no control over dimmers. You can only dim a light up or down at the actual local control, the Lumi 5 can just toggle the dimmer on and off using the 3-way runner. The architecture also means that you have two fundamentally different types of wall panels in your house: local switches or dimmers wired to each light, and the Lu Master panels with their five buttons for the five circuits, along with "all on" and "all off." The Lumi 5 "interface" uses simple relay logic to implement a few more features. Five mains-voltage-level inputs can be wired to time clocks, so that you can schedule any combination(s) of the circuits to turn on and off. The manual recommends models including one with an astronomical clock for sunrise/sunset. An additional input causes all five circuits to turn on; it's suggested for connection to an auxiliary relay on a burglar alarm to turn all of the lights on should the alarm be triggered. The whole thing is strange and fascinating. It is basically a relay lighting control system, like so many before it, but using a distinctly different wiring convention. I think the main reason for the odd wiring was to accommodate dimmers, an increasingly popular option in the 1980s that relay systems could never really contend with. It doesn't have the cost advantages of relay systems at all, it will definitely be more expensive! But it adds some features over the fancy Lutron switches and dimmers you were going to install anyway. The Lu Master is the transitional stage between relay lighting systems and later architectural lighting controls, and it straddled too the end of relay light control in homes. It gives an idea of where relay light control in homes would have evolved, had the whole technology not been doomed to the niche zone of conference centers and universities. If you think about it, the Lu Master fills the most fundamental roles of home automation in lighting: control over multiple lights in a convenient place, scheduling and triggers, and an emergency function. It only lacks scenes, which I think we can excuse considering that the simple technology it uses does not allow it to adjust dimmers. And all of that with no Node-RED in sight! Maybe that conveys what most frustrates me about the "home automation" industry: it is constantly reinventing the wheel, an oligopoly of tech companies trying to drag people's homes into their "ecosystem." They do so by leveraging the buzzword of the moment, IoT to voice assistants to, I guess now AI?, to solve a basic set of problems that were pretty well solved at least as early as 1948. That's not to deny that modern home automation platforms have features that old ones don't. They are capable of incredibly sophisticated things! But realistically, most of their users want only very basic functionality: control in convenient places, basic automation, scenes. It wouldn't sting so much if all these whiz-bang general purpose computers were good at those tasks, but they aren't. For the very most basic tasks, things like turning on and off a group of lights, major tech ecosystems like HomeKit provide a user experience that is significantly worse than the model home of 1950. You could install a Lutron system, and it would solve those fundamental tasks much better... for a much higher price. But it's not like Lutron uses all that money to be an absolute technical powerhouse, a center of innovation at the cutting edge. No, even the latest Lutron products are really very simple, technically. The technical leaders here, Google, Apple, are the companies that can't figure out how to make a damn light switch. The problem with modern home automation platforms is that they are too ambitious. They are trying to apply enormously complex systems to very simple tasks, and thus contaminating the simplest of electrical systems with all the convenience and ease of a Smart TV. Sometimes that's what it feels like this whole industry is doing: adding complexity while the core decays. From automatic programming to AI coding agents, video terminals to Electron, the scope of the possible expands while the fundamentals become more and more irritating. But back to the real point, I hope you learned about some cool light switches. Check out the Kyle Switch Plates reference and you'll start seeing these buildings and homes, at least if you live in an area that built up during the era that they were common (1950s to the 1970s).
Is there anything more irritating than living with a partner who procrastinates on their share of the chores? Even if it isn’t malicious, it sure is annoying. Taking out the trash is YouTuber CircuitCindy’s boyfriend’s responsibility, but he often fails to do the task in a timely manner. That forced Cindy to implement a sinister […] The post YouTuber builds robot to make boyfriend take out the trash appeared first on Arduino Blog.
