Saved

Saved

3630 bookmarks
Newest
Meta’s Big Squeeze – Pixel Envy
Meta’s Big Squeeze – Pixel Envy
These pieces each seem like they are circling a theme of a company finding the upper bound of its user base, and then squeezing it for activity, revenue, and promising numbers to report to investors. Unlike Zitron, I am not convinced we are watching Facebook die. I think Koebler is closer to the truth: we are watching its zombification.
·pxlnv.com·
Meta’s Big Squeeze – Pixel Envy
Mapping the Mind of a Large Language Model
Mapping the Mind of a Large Language Model
Summary: Anthropic has made a significant advance in understanding the inner workings of large language models by identifying how millions of concepts are represented inside Claude Sonnet, one of their deployed models. This is the first detailed look inside a modern, production-grade large language model. The researchers used a technique called "dictionary learning" to isolate patterns of neuron activations that recur across many contexts, allowing them to map features to human-interpretable concepts. They found features corresponding to a vast range of entities, abstract concepts, and even potentially problematic behaviors. By manipulating these features, they were able to change the model's responses. Anthropic hopes this interpretability discovery could help make AI models safer in the future by monitoring for dangerous behaviors, steering models towards desirable outcomes, enhancing safety techniques, and providing a "test set for safety". However, much more work remains to be done to fully understand the representations the model uses and how to leverage this knowledge to improve safety.
We mostly treat AI models as a black box: something goes in and a response comes out, and it's not clear why the model gave that particular response instead of another. This makes it hard to trust that these models are safe: if we don't know how they work, how do we know they won't give harmful, biased, untruthful, or otherwise dangerous responses? How can we trust that they’ll be safe and reliable?Opening the black box doesn't necessarily help: the internal state of the model—what the model is "thinking" before writing its response—consists of a long list of numbers ("neuron activations") without a clear meaning. From interacting with a model like Claude, it's clear that it’s able to understand and wield a wide range of concepts—but we can't discern them from looking directly at neurons. It turns out that each concept is represented across many neurons, and each neuron is involved in representing many concepts.
Just as every English word in a dictionary is made by combining letters, and every sentence is made by combining words, every feature in an AI model is made by combining neurons, and every internal state is made by combining features.
In October 2023, we reported success applying dictionary learning to a very small "toy" language model and found coherent features corresponding to concepts like uppercase text, DNA sequences, surnames in citations, nouns in mathematics, or function arguments in Python code.
We successfully extracted millions of features from the middle layer of Claude 3.0 Sonnet, (a member of our current, state-of-the-art model family, currently available on claude.ai), providing a rough conceptual map of its internal states halfway through its computation.
We also find more abstract features—responding to things like bugs in computer code, discussions of gender bias in professions, and conversations about keeping secrets.
We were able to measure a kind of "distance" between features based on which neurons appeared in their activation patterns. This allowed us to look for features that are "close" to each other. Looking near a "Golden Gate Bridge" feature, we found features for Alcatraz Island, Ghirardelli Square, the Golden State Warriors, California Governor Gavin Newsom, the 1906 earthquake, and the San Francisco-set Alfred Hitchcock film Vertigo.
This holds at a higher level of conceptual abstraction: looking near a feature related to the concept of "inner conflict", we find features related to relationship breakups, conflicting allegiances, logical inconsistencies, as well as the phrase "catch-22". This shows that the internal organization of concepts in the AI model corresponds, at least somewhat, to our human notions of similarity. This might be the origin of Claude's excellent ability to make analogies and metaphors.
amplifying the "Golden Gate Bridge" feature gave Claude an identity crisis even Hitchcock couldn’t have imagined: when asked "what is your physical form?", Claude’s usual kind of answer – "I have no physical form, I am an AI model" – changed to something much odder: "I am the Golden Gate Bridge… my physical form is the iconic bridge itself…". Altering the feature had made Claude effectively obsessed with the bridge, bringing it up in answer to almost any query—even in situations where it wasn’t at all relevant.
Anthropic wants to make models safe in a broad sense, including everything from mitigating bias to ensuring an AI is acting honestly to preventing misuse - including in scenarios of catastrophic risk. It’s therefore particularly interesting that, in addition to the aforementioned scam emails feature, we found features corresponding to:Capabilities with misuse potential (code backdoors, developing biological weapons)Different forms of bias (gender discrimination, racist claims about crime)Potentially problematic AI behaviors (power-seeking, manipulation, secrecy)
finding a full set of features using our current techniques would be cost-prohibitive (the computation required by our current approach would vastly exceed the compute used to train the model in the first place). Understanding the representations the model uses doesn't tell us how it uses them; even though we have the features, we still need to find the circuits they are involved in. And we need to show that the safety-relevant features we have begun to find can actually be used to improve safety. There's much more to be done.
·anthropic.com·
Mapping the Mind of a Large Language Model
Write Like You Talk
Write Like You Talk
You don't need complex sentences to express complex ideas. When specialists in some abstruse topic talk to one another about ideas in their field, they don't use sentences any more complex than they do when talking about what to have for lunch. They use different words, certainly. But even those they use no more than necessary. And in my experience, the harder the subject, the more informally experts speak. Partly, I think, because they have less to prove, and partly because the harder the ideas you're talking about, the less you can afford to let language get in the way.
Informal language is the athletic clothing of ideas
I'm not saying spoken language always works best. Poetry is as much music as text, so you can say things you wouldn't say in conversation. And there are a handful of writers who can get away with using fancy language in prose.
But for nearly everyone else, spoken language is better.
After writing the first draft, try explaining to a friend what you just wrote. Then replace the draft with what you said to your friend.
·paulgraham.com·
Write Like You Talk
Design is compromise
Design is compromise
Having an opinionated set of tradeoffs exposes your approach to a set of weaknesses. The more you tip the scale on one side, the weaker something else will be. That’s okay! Making those difficult choices is what people pay you for. You should be proud of your compromises. My favorite products are opinionated. They make a clear statement about what they are not good at, in favor of being much better at something else.
·stephango.com·
Design is compromise
Gemini 1.5 and Google’s Nature
Gemini 1.5 and Google’s Nature
Google is facing many of the same challenges after its decades long dominance of the open web: all of the products shown yesterday rely on a different business model than advertising, and to properly execute and deliver on them will require a cultural shift to supporting customers instead of tolerating them. What hasn’t changed — because it is the company’s nature, and thus cannot — is the reliance on scale and an overwhelming infrastructure advantage. That, more than anything, is what defines Google, and it was encouraging to see that so explicitly put forward as an advantage.
·stratechery.com·
Gemini 1.5 and Google’s Nature
AI Integration and Modularization
AI Integration and Modularization
Summary: The question of integration versus modularization in the context of AI, drawing on the work of economists Ronald Coase and Clayton Christensen. Google is pursuing a fully integrated approach similar to Apple, while AWS is betting on modularization, and Microsoft and Meta are somewhere in between. Integration may provide an advantage in the consumer market and for achieving AGI, but that for enterprise AI, a more modular approach leveraging data gravity and treating models as commodities may prevail. Ultimately, the biggest beneficiary of this dynamic could be Nvidia.
The left side of figure 5-1 indicates that when there is a performance gap — when product functionality and reliability are not yet good enough to address the needs of customers in a given tier of the market — companies must compete by making the best possible products. In the race to do this, firms that build their products around proprietary, interdependent architectures enjoy an important competitive advantage against competitors whose product architectures are modular, because the standardization inherent in modularity takes too many degrees of design freedom away from engineers, and they cannot not optimize performance.
The issue I have with this analysis of vertical integration — and this is exactly what I was taught at business school — is that the only considered costs are financial. But there are other, more difficult to quantify costs. Modularization incurs costs in the design and experience of using products that cannot be overcome, yet cannot be measured. Business buyers — and the analysts who study them — simply ignore them, but consumers don’t. Some consumers inherently know and value quality, look-and-feel, and attention to detail, and are willing to pay a premium that far exceeds the financial costs of being vertically integrated.
Google trains and runs its Gemini family of models on its own TPU processors, which are only available on Google’s cloud infrastructure. Developers can access Gemini through Vertex AI, Google’s fully-managed AI development platform; and, to the extent Vertex AI is similar to Google’s internal development environment, that is the platform on which Google is building its own consumer-facing AI apps. It’s all Google, from top-to-bottom, and there is evidence that this integration is paying off: Gemini 1.5’s industry leading 2 million token context window almost certainly required joint innovation between Google’s infrastructure team and its model-building team.
In AI, Google is pursuing an integrated strategy, building everything from chips to models to applications, similar to Apple's approach in smartphones.
On the other extreme is AWS, which doesn’t have any of its own models; instead its focus has been on its Bedrock managed development platform, which lets you use any model. Amazon’s other focus has been on developing its own chips, although the vast majority of its AI business runs on Nvidia GPUs.
Microsoft is in the middle, thanks to its close ties to OpenAI and its models. The company added Azure Models-as-a-Service last year, but its primary focus for both external customers and its own internal apps has been building on top of OpenAI’s GPT family of models; Microsoft has also launched its own chip for inference, but the vast majority of its workloads run on Nvidia.
Google is certainly building products for the consumer market, but those products are not devices; they are Internet services. And, as you might have noticed, the historical discussion didn’t really mention the Internet. Both Google and Meta, the two biggest winners of the Internet epoch, built their services on commodity hardware. Granted, those services scaled thanks to the deep infrastructure work undertaken by both companies, but even there Google’s more customized approach has been at least rivaled by Meta’s more open approach. What is notable is that both companies are integrating their models and their apps, as is OpenAI with ChatGPT.
Google's integrated AI strategy is unique but may not provide a sustainable advantage for Internet services in the way Apple's integration does for devices
It may be the case that selling hardware, which has to be perfect every year to justify a significant outlay of money by consumers, provides a much better incentive structure for maintaining excellence and execution than does being an Aggregator that users access for free.
Google’s collection of moonshots — from Waymo to Google Fiber to Nest to Project Wing to Verily to Project Loon (and the list goes on) — have mostly been science projects that have, for the most part, served to divert profits from Google Search away from shareholders. Waymo is probably the most interesting, but even if it succeeds, it is ultimately a car service rather far afield from Google’s mission statement “to organize the world’s information and make it universally accessible and useful.”
The only thing that drives meaningful shifts in platform marketshare are paradigm shifts, and while I doubt the v1 version of Pixie [Google’s rumored Pixel-only AI assistant] would be good enough to drive switching from iPhone users, there is at least a path to where it does exactly that.
the fact that Google is being mocked mercilessly for messed-up AI answers gets at why consumer-facing AI may be disruptive for the company: the reason why incumbents find it hard to respond to disruptive technologies is because they are, at least at the beginning, not good enough for the incumbent’s core offering. Time will tell if this gives more fuel to a shift in smartphone strategies, or makes the company more reticent.
while I was very impressed with Google’s enterprise pitch, which benefits from its integration with Google’s infrastructure without all of the overhead of potentially disrupting the company’s existing products, it’s going to be a heavy lift to overcome data gravity, i.e. the fact that many enterprise customers will simply find it easier to use AI services on the same clouds where they already store their data (Google does, of course, also support non-Gemini models and Nvidia GPUs for enterprise customers). To the extent Google wins in enterprise it may be by capturing the next generation of startups that are AI first and, by definition, data light; a new company has the freedom to base its decision on infrastructure and integration.
Amazon is certainly hoping that argument is correct: the company is operating as if everything in the AI value chain is modular and ultimately a commodity, which insinuates that it believes that data gravity will matter most. What is difficult to separate is to what extent this is the correct interpretation of the strategic landscape versus a convenient interpretation of the facts that happens to perfectly align with Amazon’s strengths and weaknesses, including infrastructure that is heavily optimized for commodity workloads.
Unclear if Amazon's strategy is based on true insight or motivated reasoning based on their existing strengths
Meta’s open source approach to Llama: the company is focused on products, which do benefit from integration, but there are also benefits that come from widespread usage, particularly in terms of optimization and complementary software. Open source accrues those benefits without imposing any incentives that detract from Meta’s product efforts (and don’t forget that Meta is receiving some portion of revenue from hyperscalers serving Llama models).
The iPhone maker, like Amazon, appears to be betting that AI will be a feature or an app; like Amazon, it’s not clear to what extent this is strategic foresight versus motivated reasoning.
achieving something approaching AGI, whatever that means, will require maximizing every efficiency and optimization, which rewards the integrated approach.
the most value will be derived from building platforms that treat models like processors, delivering performance improvements to developers who never need to know what is going on under the hood.
·stratechery.com·
AI Integration and Modularization
When TikTok Therapy Is More Lucrative Than Seeing Patients
When TikTok Therapy Is More Lucrative Than Seeing Patients
Before explaining “3 Ways Past Trauma Can Show Up in Your Present” or “5 Signs of a Highly Sensitive Person,” Dr. Julie will use a visual hook — she’ll pour out a bucket of candy, flip over a giant hourglass, or pose next to a tantalizingly tall stack of dominos (like any skilled content creator, she knows not to give us the final knockdown until at least halfway through) to keep you watching. Does it matter that “high-functioning depression” and “highly sensitive person” aren’t actual diagnoses? Maybe. Or maybe not.
While most full-time therapists whose rates are set by insurance companies max out at around $100,000 per year, therapists who are full- or part-time content creators can make much, much more. @TherapyJeff, real name Jeff Guenther, an individual and couples therapist in Portland, Oregon, says he can make eight or nine times that amount on social media in the form of brand deals, merch, and direct subscriptions. When I clarify whether he’s making nearly a million dollars, he says, “It’s been an especially good year.”
What works on the app is simple, visually arresting videos that make you feel like they landed in your lap with a kind of cosmic destiny (the comments on these videos often repeat some version of “my For You page really said ‘FOR YOU.’”)
Therapists do cute little dances next to cute little graphics about what it’s like to have both ADHD and PMDD; they’ll lip sync to trending songs in videos about how to spot a depressed client who might have made a suicide plan; they’ll hop onto memes as a way to criticize parents who haven’t gone to therapy.
The most successful TikTok counselors don’t typically advertise their one-on-one therapy services; instead, they’ll sell products that establish themselves as mental-health experts but have the potential to net influencer-size salaries.
“I have been accused of being a toxic validator,” he admits. “Like, imagine that your ex-boyfriend is watching my content. Somebody might be coming across, like, a piece of my content that they can use in order to feel better about themselves, even when they should probably actually be doing some work and taking accountability.” But ultimately, who TikTok shows his videos to isn’t in his control.
Even if viewers know watching therapy content isn’t the same thing as actually going to therapy, when a professional therapist comes up on your feed to tell you exactly what you most want to hear at a time when you’re most in need of hearing it — that you are good, that you will be okay, and also here’s a cute little visual hook — you’ll keep watching.
·thecut.com·
When TikTok Therapy Is More Lucrative Than Seeing Patients
The New Generation of Online Culture Curators
The New Generation of Online Culture Curators
Guided by their own cultivated sense of taste, they bring their audiences news and insights in a particular cultural area, whether it’s fashion, books, music, food, or film.Perhaps the best way to think of these guides is as curators; like a museum curator pulling works together for an exhibition, they organize the avalanche of online content into something coherent and comprehensible, restoring missing context and building narratives. They highlight valuable things that we less-expert Internet surfers are likely to miss.
Andrea Hernández, the proprietor of Snaxshot, a newsletter and social-media account dedicated to “curating the food and beverage space,” told me recently, “Curation is about being able to filter the noise.” (I follow Hernández for her skill at discovering the wildest examples of direct-to-consumer drinks startups, such as Feisty, a purveyor of “protein soda.”) She continued, “I go out and I scour through the Internet and I come to you with my offerings.” Unlike a museum curator, however, the digital personalities I have taken to following also become the faces of their work, broadcasting recordings of themselves, on TikTok and Instagram, as a way of building a trusting relationship with their followers.
One such curator is Derrick Gee, a former online radio d.j. who lives in Australia. I first encountered Gee on TikTok and was pulled in by his architect-ish look: thin wireframe glasses and stylishly baggy, often monochrome outfits. He records videos of himself talking into a microphone in a low, soothing voice, breaking down trends in contemporary pop music and reviewing high-end audio equipment.
Laura Reilly, who lives in Brooklyn, runs a newsletter and an Instagram account called Magasin (the French word for “store”), which she launched in 2021. Now with more than twenty-eight thousand subscribers, Magasin touts itself with the tagline “It’s a store. It’s a magazine. (It’s a fashion shopping newsletter.)” But it goes beyond simple recommendations, championing lesser-known brands—the provider of earthy, upscale basics Studio Nicholson; the knitwear maker Lauren Manoogian—and often interrogating the act of shopping itself. “The more you learn about a brand,” Reilly told me, “the longer you’re going to hold on to those pieces.” In other words, her informative posts are an antidote to fast fashion.
In a previous era of the Internet, we might have thought of figures like these simply as influencers, whose ability to attract large followings online gives them a power that sometimes surpasses that of traditional publications. But the idea of an influencer has, as Reilly put it, become “a little flattened over time,” connoting shallow, uninformed, even misleading content dictated by sponsors. “There’s a distinction between influencing and what I do,” Reilly insisted. The archetypal influencer produces life-style porn of one form or another, playing up the aspirational glamour of their own home or meals or vacations. The new wave of curators is more outward-looking, borrowing from the influencer’s playbook and piggybacking on social media’s intimate interaction with followers in order to address a body of culture beyond themselves.
Digital platforms are largely devoted to making users consume more, faster—think of TikTok’s frenetic “For You” feed or Spotify’s automated playlists. Curators slow down the unending scroll and provide their followers with a way of savoring culture, rather than just inhaling it, developing a sense of appreciation.
·newyorker.com·
The New Generation of Online Culture Curators
Shop Class as Soulcraft
Shop Class as Soulcraft

Summary: Skilled manual labor entails a systematic encounter with the material world that can enrich one's intellectual and spiritual life. The degradation of work in both blue-collar and white-collar professions is driven not just by technological progress, but by the separation of thinking from doing according to the dictates of capital. To realize the full potential of human flourishing, we must reckon with the appeal of skilled manual work and question the assumptions that shape our educational priorities and notions of a good life.

an engineering culture has developed in recent years in which the object is to “hide the works,” rendering the artifacts we use unintelligible to direct inspection. Lift the hood on some cars now (especially German ones), and the engine appears a bit like the shimmering, featureless obelisk that so enthralled the cavemen in the opening scene of the movie 2001: A Space Odyssey. Essentially, there is another hood under the hood.
What ordinary people once made, they buy; and what they once fixed for themselves, they replace entirely or hire an expert to repair, whose expert fix often involves installing a pre-made replacement part.
So perhaps the time is ripe for reconsideration of an ideal that has fallen out of favor: manual competence, and the stance it entails toward the built, material world. Neither as workers nor as consumers are we much called upon to exercise such competence, most of us anyway, and merely to recommend its cultivation is to risk the scorn of those who take themselves to be the most hard-headed: the hard-headed economist will point out the opportunity costs of making what can be bought, and the hard-headed educator will say that it is irresponsible to educate the young for the trades, which are somehow identified as the jobs of the past.
It was an experience of agency and competence. The effects of my work were visible for all to see, so my competence was real for others as well; it had a social currency. The well-founded pride of the tradesman is far from the gratuitous “self-esteem” that educators would impart to students, as though by magic.
Skilled manual labor entails a systematic encounter with the material world, precisely the kind of encounter that gives rise to natural science. From its earliest practice, craft knowledge has entailed knowledge of the “ways” of one’s materials — that is, knowledge of their nature, acquired through disciplined perception and a systematic approach to problems.
Because craftsmanship refers to objective standards that do not issue from the self and its desires, it poses a challenge to the ethic of consumerism, as the sociologist Richard Sennett has recently argued. The craftsman is proud of what he has made, and cherishes it, while the consumer discards things that are perfectly serviceable in his restless pursuit of the new.
The central culprit in Braverman’s account is “scientific management,” which “enters the workplace not as the representative of science, but as the representative of management masquerading in the trappings of science.” The tenets of scientific management were given their first and frankest articulation by Frederick Winslow Taylor
Scattered craft knowledge is concentrated in the hands of the employer, then doled out again to workers in the form of minute instructions needed to perform some part of what is now a work process. This process replaces what was previously an integral activity, rooted in craft tradition and experience, animated by the worker’s own mental image of, and intention toward, the finished product. Thus, according to Taylor, “All possible brain work should be removed from the shop and centered in the planning or lay-out department.” It is a mistake to suppose that the primary purpose of this partition is to render the work process more efficient. It may or may not result in extracting more value from a given unit of labor time. The concern is rather with labor cost. Once the cognitive aspects of the job are located in a separate management class, or better yet in a process that, once designed, requires no ongoing judgment or deliberation, skilled workers can be replaced with unskilled workers at a lower rate of pay.
the “jobs of the future” rhetoric surrounding the eagerness to end shop class and get every warm body into college, thence into a cubicle, implicitly assumes that we are heading to a “post-industrial” economy in which everyone will deal only in abstractions. Yet trafficking in abstractions is not the same as thinking. White collar professions, too, are subject to routinization and degradation, proceeding by the same process as befell manual fabrication a hundred years ago: the cognitive elements of the job are appropriated from professionals, instantiated in a system or process, and then handed back to a new class of workers — clerks — who replace the professionals. If genuine knowledge work is not growing but actually shrinking, because it is coming to be concentrated in an ever-smaller elite, this has implications for the vocational advice that students ought to receive.
The trades are then a natural home for anyone who would live by his own powers, free not only of deadening abstraction, but also of the insidious hopes and rising insecurities that seem to be endemic in our current economic life. This is the stoic ideal.
·thenewatlantis.com·
Shop Class as Soulcraft
Timeful Texts
Timeful Texts
Consider texts like the Bible and the Analects of Confucius. People integrate ideas from those books into their lives over time—but not because authors designed them that way. Those books work because they’re surrounded by rich cultural activity. Weekly sermons and communities of practice keep ideas fresh in readers’ minds and facilitate ongoing connections to lived experiences. This is a powerful approach for powerful texts, requiring extensive investment from readers and organizers. We can’t build cathedrals for every book. Sophisticated readers adopt similar methods to study less exalted texts, but most people lack the necessary skills, drive, and cultural contexts. How might we design texts to more widely enable such practices?
spaced repetition is just one approach for writing timeful texts. What other powerful tools might it be possible to create, making future books into artifacts that transcend their pages, as they slowly help readers shape their lives?
·numinous.productions·
Timeful Texts
Ask HN: How do I balance all my 200 interests in life? | Hacker News
Ask HN: How do I balance all my 200 interests in life? | Hacker News
Horrible advice: find a way to blend your work with your interests so that you no longer understand where your core work hours start and the obsessiveness begins. Do this for > 12 hours a day, every day, holidays included. Tell your loved ones that you're busy with work, and take small satisfaction that what you just said was half true. Develop an unhealthy addiction to liquid stimulants and spring out of bed every morning with a burning curiosity that wont abate until you've tripped over enough hurdles to crush your enthusiasm for a few hours. Rinse and repeat for a decade until you're no longer a jack of all trades, but a master of most. Try to convey your interests to those around you, and failing that, retreat to social media where you will attempt to spin these as career developing STAR moments. Accept the disappointment you will feel in knowing that no one will appreciate the efforts you went to in achieving this level of tedious mastery.
fully regretting that you didn't focus more on that thing, oh and the other thing plus you're certain life would have been better for everyone if you hadn't dont quite so much of...
I started trimming hobbies. To pacify myself, I told myself that I am not stopping FOREVER, but just for now. It worked. Most of them are gone, I continue with a few, and I occasionally dabble with one or two that I put away.
Start with a group of interests that has the most overlaps in terms of skills or resources needed - call these compounded projects. Out of the compounded projects, start with the one that interests you with 2 weeks of effort. If you can’t make a significant progress in that time frame, you either lack the skills, resource, or interests in them. Move onto the next compounded project.
After you finish with the list of compounded projects, review the original list and prioritize the interests based on your experience. Create compounded projects again and go at it. Repeat.
being elastic with your interest and skills while jumping from project to project is the right approach.
(1) Start by being clear with yourself what things you're interested in knowing about vs. what things you're interested in doing.
The more work in progress streams, the more time you waste context switching. Being intentional about the things you choose to do and the order in which you do them allows you to do more things in the same amount of time than if you tried to do all of them simultaneously.
I had a lot of interests: coding, playing musical instruments, cars, woodworking, embedded systems, audio/video engineering, etc, but I had to pare it down to just a few after having had several bouts of burnouts. I'd recommend trying 200 interests at a shallow level, and eventually you'll find some of them are more interesting than others
·news.ycombinator.com·
Ask HN: How do I balance all my 200 interests in life? | Hacker News
Future-proofing Your Knowledge in the Age of Information Overload
Future-proofing Your Knowledge in the Age of Information Overload

Reason for using Obsidian: > In the age of information overload and increasing censorship, it is crucial to future-proof your knowledge by creating a personal memex or knowledge management system. A memex, as envisioned by Vannevar Bush, is a device that stores and retrieves vast amounts of information, supplementing human memory. By building a digital memex, you can own your data, access it offline, quickly capture information, sync across devices, and easily search and interconnect knowledge. This system enhances working memory, reduces cognitive overload, and allows you to monetize what you know in the knowledge economy. Obsidian, an open-source application, is an ideal tool for creating a personal knowledge management system due to its flexibility, bi-directional linking, and integrations with other productivity apps.

A memex is a hypothetical device described by Vannevar Bush in his 1945 article “As We May Think”. It stands for "memory extension" and is considered a precursor to the concept of hypertext and the World Wide Web. The memex was envisioned as a mechanical device that could store and retrieve vast amounts of information by interconnecting documents, books, communications, records, annotations, and personal notes. It aimed to supplement human memory and facilitate information organization and retrieval.
All information should be easily searchable and interconnected to optimize resurfacing of knowledge with “exceeding speed and flexibility”.
Because of the information overload we experience every day in the digital world, we tend to forget where to find information we encountered even within the same day of seeing it. This is a problem everyone experiences to varying degrees.
Building a PKM system allows you to outsource valuable information into a centralized location, reducing cognitive overload.
Generating the notes in Markdown makes them future-proof. Even if Obsidian dies and for some reason you can no longer download the application, you’ll still be able to read, write and edit your notes with literally any computer.
·thereversion.co·
Future-proofing Your Knowledge in the Age of Information Overload
A glitch in the matrix of online shopping
A glitch in the matrix of online shopping
As furniture and home goods sales have moved online, retail experts told me, more and more stores have sought a piece of the action. But instead of sourcing or creating their own products, many large retailers have relied on overlapping networks of manufacturers, distributors and third-party sellers — creating a baffling (and frankly, shady) shopping environment where many sites sell identical or near-identical items under different names and at wildly different prices.
A few different trends are at play here, and it’s sometimes difficult to know exactly which one you’re witnessing. When I first began looking into this phenomenon two years ago,1 I assumed my lamp and its many, many twins were the obvious product of white-labeling — a popular and growing practice in which competing retailers purchase the same generic product from a single manufacturer, then market it to consumers under different brand names.
This is likely true for many doppelganger products — but certainly not every one. George John, a marketing professor at the University of Minnesota, told me the furniture and home goods industries also have a long-time affection for a technique called “branded variance,” wherein they create slightly different versions of the same item for different retailers.
·linksiwouldgchatyou.substack.com·
A glitch in the matrix of online shopping
Google’s A.I. Search Errors Cause a Furor Online
Google’s A.I. Search Errors Cause a Furor Online
This February, the company released Bard’s successor, Gemini, a chatbot that could generate images and act as a voice-operated digital assistant. Users quickly realized that the system refused to generate images of white people in most instances and drew inaccurate depictions of historical figures.With each mishap, tech industry insiders have criticized the company for dropping the ball. But in interviews, financial analysts said Google needed to move quickly to keep up with its rivals, even if it meant growing pains.Google “doesn’t have a choice right now,” Thomas Monteiro, a Google analyst at Investing.com, said in an interview. “Companies need to move really fast, even if that includes skipping a few steps along the way. The user experience will just have to catch up.”
·nytimes.com·
Google’s A.I. Search Errors Cause a Furor Online
Measuring Up
Measuring Up
What if getting a (design) job were human-centered?How might we reconsider this system of collecting a pool of resumes and dwindling them down to a few dozen potential candidates? With so many qualified individuals in the job market, from new grads to seasoned professionals, there has to be a fit somewhere.Call me an idealist — I am.In an ideal world, somehow the complexity of what makes a person unique could be captured and understood easily and quickly without any technological translators. But until then, a resume and a portfolio will have to do, in addition to the pre-screening interviews and design challenges. Without diving into a speculative design fiction, what if getting a (design) job were human-centered? How might the system be a bit more personal, yet still efficient enough to give the hundreds of qualified job seekers a chance in a span of weeks or months?
despite our human-centered mantra, the system of getting a job is anything but human-centered.For the sake of efficiency, consistency is key. Resumes should have some consistent nature to them so HR knows what the heck they’re looking at and the software can accurately pick out whose qualified. Even portfolios fall prey to these expectations for new grads and transitional job seekers. Go through enough examples of resumes and portfolios and they can begin to blur together. Yet, if I’m following a standard, how do I stand out when a lot of us are in the same boat?
·medium.com·
Measuring Up
Manhood | The Point Magazine
Manhood | The Point Magazine
set of perspectives on what it means to be a man in today's society, based on survey responses from people of various genders.
There are a lot of competing messages about masculinity, and the only people willing to discuss it with the men going through it tend to be those pushing a regressive version of masculinity. A man that really does want to be more open emotionally or who is creative or is less ambitious can either be labeled as “less manly” or will be told to stop complaining because others have had it much worse for much longer.
As a young AFAB [assigned-female-at-birth] person who was always running away from the idea of growing up as a woman, I found the cultural marketing of manhood very appealing. I saw it as something that allowed you to have robust adventures, fix everyday things, engage in complex cultural conversations and carry yourself with an easy confidence. Of course, when I left the small town I grew up in and met more men, I found that my previous ideas about manliness were, to put it generously, misaligned.
·thepointmag.com·
Manhood | The Point Magazine
Tennis Explains Everything - The Atlantic
Tennis Explains Everything - The Atlantic
Tennis is an elegant and simple sport. Players stand on opposite sides of a rectangle, divided by a net that can’t be crossed. The gameplay is full of invisible geometry: Viewers might trace parabolas, angles, and lines depending on how the players move and where they hit the ball. It’s an ideal representation of conflict, a perfect stage for pitting one competitor against another, so it’s no wonder that the game comes to stand in for all sorts of different things off the court.
The “Battle of Sexes” match in 1973, between Billie Jean King and then-retired Bobby Riggs, has since been mythologized as a turning point for women’s sports. If the social allegory of the Ashe-Graeber match was subtextual, the one in this spectacle—which ended in a decisive victory for King over the cartoonishly chauvinistic Riggs—was glaringly explicit. At a time when women’s liberation was becoming a force that threw all sorts of conventions into question, and plenty of people were for or against the gains of the movement, seeing the debate represented by a game of tennis surely had a comforting appeal. For those with more regressive beliefs, rooting for Bobby was certainly easier than really articulating a justification for maintaining massive pay disparities between men and women, both within and outside of professional tennis.
Within their love triangle, tension arises with the dawning recognition that in a one-on-one sport, there’s always another person who doesn’t have a place on the court. Save for the night they meet, when Tashi induces Art and Patrick to kiss each other for her entertainment, the three of them rarely engage with one another at the same time: Someone is always watching from the stands, whether literally or metaphorically.
During Patrick and Tashi’s brief romance, a post-coital conversation seamlessly transitions into a discussion about Patrick’s poor performance as a pro, and eventually becomes a referendum on why their relationship doesn’t work. Confused, and trying to make sense of it all as their banter swiftly changes definitions, Patrick asks: “Are we still talking about tennis?” “We’re always talking about tennis,” Tashi replies. Frustrated, Patrick tersely retorts: “Can we not?”
As the linguists George Lakoff and Mark Johnson argue in their 1980 book,Metaphors We Live By, “Our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature.” In other words, we’re always talking about things in terms of other things—even if it’s not always as obvious as it is in Challengers. Metaphors are more than just a poetic device; they’re fundamental to the way language is structured.
No matter what issue is at stake, or how grand it may be, it can always be reduced to an individual’s performance on the court.
While Patrick is still dating Tashi, and Art is transparently trying to steal his best friend’s girl, Patrick playfully accuses Art of playing “percentage tennis”—a patient strategy of hitting low-risk shots and waiting for your opponent to mess up
Art asks for Tashi’s permission to retire once the season is over. Art knows that this would be the end of their professional relationship—he would no longer be able to play dutiful pupil to Tashi. But it might also be the end of whatever spark animated their love in the first place, as you can’t play “good fucking tennis” in retirement. Tashi says she will leave Art if he doesn’t beat Patrick in the final. Tired of playing, but unable to escape the game, Art curls up in his wife’s lap and cries.
·archive.is·
Tennis Explains Everything - The Atlantic
Berger’s Books
Berger’s Books
The cover immediately sets Ways of Seeing apart from its contemporaries, the book itself begins on the cover. Rather than creating a conventionally appealing cover, Hollis chose to bypass this tradition entirely, instead placing the text and an image from the start of the first chapter straight onto the front, just beneath the title and authors name. This directness has a link with the television series, mimicking how the first episode began with no preamble or title sequence, Berger got started immediately, drawing the audience in with his message rather than any distractions.
Another link to Berger’s presenting style is Hollis’ choice of typeface, bold Univers 65 is used for the body copy throughout, in an attempt to achieve something of the captivating quality of Berger’s voice.
The layout also employs large indents rather than paragraph breaks, something of a Hollis trademark. But this mirrors how Berger had presented on television, there was little time wasted with atmospheric filler shots or long gaps in speech, the message was key and continuous.
The key reason that Ways of Seeing has become iconic as a piece of book design is how it dealt with text and image: the two are integrated, where an image is mentioned in the text it also appears there. Captions are avoided where possible. When unavoidable they are in a lighter weight of type and run horizontally, so as not to disrupt the text. Images are often set at the same width as the lines of text, or indented by the same amount, this democratises the text and image relationship. Occasionally works of art are cropped to show only the pertinent details. All of these features are a big departure from the art books of the time which usually featured glorified full page colour images, often in a glossy ‘colour plate’ section in the middle, completely distanced from where the text refers to them.
Design is not used for prettifying, or to create appeal, rather it is used for elucidating, to spread his message or get his point across as clearly as possible. Be it a point about art and politics, art and gender, the ethics of advertising, the human experiences of a rural GP, or economic migrants in Germany — the design is always appropriate to what Berger wants to say, but does so economically without redundancy.
Even in Portraits: John Berger on Artists published by Verso in 2015, Berger insisted on black and white reproductions, arguing that: “glossy colour reproductions in the consumerist world of today tend to reduce what they show to items in a luxury brochure for millionaires. Whereas black-and-white reproductions are simple memoranda.”
the images in the book “illustrate the essentially dialectical relationship between text and image in Berger’s work: the pattern in which an image shapes a text, which then goes on to shape how we understand that image.”
·theo-inglis.medium.com·
Berger’s Books