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
The shortest distance between your thoughts and the printed word.
It isn’t a secret that many kids find math to be boring and it is easy for them to develop an attitude of “when am I ever going to use this?” But math is incredibly useful in the real world, from blue-collar machinists using trigonometry to quantum physicists unveiling the secrets of our universe through […] The post This unique electronic toy helps children learn their shapes appeared first on Arduino Blog.
I have worked with a few software developers who made the switch to this industry in the middle of their careers. A major change like that can be scary and raise a lot of fears and doubts, but I can attest that this can work out well with the right personality traits and a supporting environment. Here’s what I’ve observed. To keep the writing concise, I’ll be using the phrase “senior junior”1 to describe those that have made such a career switch. Overcoming the fear Fear is a natural reaction to any major change in life, especially when there’s risk of taking a financial hit while you have a family to support and a home loan to pay. The best mitigation that I’ve heard is believing that you can make the change, successfully. It sounds like an oversimplification, sure, as all it does is that it removes a mental blocker and throws out the self-doubt. And yet it works unreasonably well. It also helps if you have at least some savings to help mitigate the financial risk. A years’ worth of expenses saved up can go a long way in providing a solid safety net. What makes them succeed A great software developer is not someone that simply slings some code over the wall and spends all of their day working only on the technical stuff, there are quite a few critical skills that one needs to succeed. This is not an exhaustive list, but I’ve personally observed that the following ones are the most critical: ability to work in a team great communication skills conflict resolution ability to make decisions in the context of product development and business goals maintaining an environment of psychological safety Those with more than a decade of experience in another role or industry will most likely have a lot of these skills covered already, and they can bring that skill set into a software development team while working with the team to build their technical skill set. Software development is not special, at the end of they day, you’re still interacting with humans and everything that comes with that, good or bad. After working with juniors that are fresh out of school and “senior juniors” who have more career experience than I do, I have concluded that the ones that end up being great software developers have one thing in common: the passion and drive to learn everything about the role and the work we do. One highlight that I often like to share in discussions is one software developer who used to work in manufacturing. At some point they got interested in learning how they can use software to make work more efficient. They started with an MVP solution involving a big TV and Google Sheets, then they started learning about web development for a solution in a different area of the business, and ended up building a basic inventory system for the warehouse. After 2-3 years of self-learning outside of work hours and deploying to production in the most literal sense, they ended up joining my team. They got up to speed very quickly and ended up being a very valuable contributor in the team. In another example, I have worked with someone who previously held a position as a technical draftsman and 3D designer in a ship building factory (professionals call it a shipyard), but after some twists and turns ended up at a course for those interested in making a career switch, which led to them eventually working in the same company I do. Now they ship builds with confidence while making sure that the critical system we are working on stays stable. That developer also kicks my ass in foosball about 99% of the time. The domain knowledge advantage The combination of industry experience and software development skills is an incredibly powerful one. When a software developer starts work in a project, they learn the business domain piece by piece, eventually reaching a state where they have a slight idea about how the business operates, but never the full picture. Speaking with their end users will help come a long way, but there are always some details that get lost in that process. Someone coming from the industry will have in-depth knowledge about the business, how it operates, where the money comes from, what are the main pain points and where are the opportunities for automation. They will know what problems need solving, and the basic technical know-how on how to try solving them. Like a product owner, but on steroids. Software developers often fall into the trap of creating a startup to scratch that itch they have for building new things, or trying out technologies that have for a very long time been on their to-do list. The technical problems are fun to solve, sure, but the focus should be on the actual problem that needs fixing. If I wanted to start a new startup with someone, I’d look for someone working in an industry that I’m interested in and who understands the software development basics. Or maybe I’m just looking for an excellent product owner. How to help them succeed If you have a “senior junior” software developer on your team, then there really isn’t anything special you’d need to do compared to any other new joiner. Do your best to foster a culture of psychological safety, have regular 1-1s with them, and make sure to pair them up with more experienced team members as often as possible. A little bit of encouragement in challenging environments or periods of self-doubt can also go a long way. Temporary setbacks are temporary, after all. What about “AI”? Don’t worry about all that “AI”2 hype, if it was as successful in replacing all software development jobs as a lof of people like to shout from the rooftops, then it would have already done so. At best, it’s a slight productivity boost3 at the cost of a huge negative impact on the environment. Closing thoughts If you’re someone that has thought about working as a software developer or who is simply excited about all the ways that software can be used to solve actual business problems and build something from nothing, then I definitely recommend giving it a go, assuming that you have the safety net and risk appetite to do so. For reference, my journey towards software development looked like this, plus a few stints of working as a newspaper seller or a grocery store worker. who do you call a “senior senior” developer, a senile developer? ↩︎ spicy autocomplete engines (also known as LLM-s) do not count as actual artificial intelligence. ↩︎ what fascinates me about all the arguments around “AI” (LLM-s) is the feeling of being more productive. But how do you actually measure developer productivity, and do you account for possible reduced velocity later on when you’ve mistaken code generation speed as velocity and introduced hard to catch bugs into the code base that need to be resolved when they inevitably become an issue? ↩︎
I uploaded YouTube videos from time to time, and a fun comment I often get is “Whoa, this is in 8K!”. Even better, I’ve had comments from the like, seven people with 8K TVs that the video looks awesome on their TV. And you guessed it, I don’t record my videos in 8K! I record them in 4K and upscale them to 8K after the fact. There’s no shortage of AI video upscaling tools today, but they’re of varying quality, and some are great but quite expensive. The legendary Finn Voorhees created a really cool too though, called fx-upscale, that smartly leverages Apple’s built-in MetalFX framework. For the unfamiliar, this library is an extensive of Apple’s Metal graphics library, and adds functionality similar to NVIDIA’s DLSS where it intelligently upscales video using machine learning (AI), so rather than just stretching an image, it uses a model to try to infer what the frame would look like at a higher resolution. It’s primarily geared toward video game use, but Finn’s library shows it does an excellent job for video too. I think this is a really killer utility, and use it for all my videos. I even have a license for Topaz Video AI, which arguably works better, but takes an order of magnitude longer. For instance my recent 38 minute, 4K video took about an hour to render to 8K via fx-upscale on my M1 Pro MacBook Pro, but would take over 24 hours with Topaz Video AI. # Install with homebrew brew install finnvoor/tools/fx-upscale # Outputs a file named my-video Upscaled.mov fx-upscale my-video.mov --width 7680 --codec h265 Anyway, just wanted to give a tip toward a really cool tool! Finn’s even got a [version in the Mac App Store called Unsqueeze](https://apps.apple.com/ca/app/unsqueeze/id6475134617 Unsqueeze) with an actual GUI that’s even easier to use, but I really like the command line version because you get a bit more control over the output. 8K is kinda overkill for most use cases, so to be clear you can go from like, 1080p to 4K as well if you’re so inclined. I just really like 8K for the future proofing of it all, in however many years when 8K TVs are more common I’ll be able to have some of my videos already able to take advantage of that. And it takes long enough to upscale that I’d be surprised to see TVs or YouTube offering that upscaling natively in a way that looks as good given the amount of compute required currently. Obviously very zoomed in to show the difference easier If you ask me, for indie creators, even when 8K displays are more common, the future of recording still probably won’t be in native 8K. 4K recording gives so much detail still that have more than enough details to allow AI to do a compelling upscale to 8K. I think for my next camera I’m going to aim for recording in 6K (so I can still reframe in post), and then continue to output the final result in 4K to be AI upscaled. I’m coming for you, Lumix S1ii.
Talks about the famous Dragon's Lair