> ## Content Index > Fetch the complete content index at: https://labnotes.org/llms.txt > Use this file to discover other available public pages before exploring further. # Weekend Reading — At the POST office trying to GET my mail - URL: https://labnotes.org/weekend-reading-at-the-post-office-trying-to-get-my-mail/ - Published: 2026-09-05T19:25:27.000Z - Updated: 2026-09-05T19:25:27.000Z - Author: Assaf Arkin [**A nosy polar bear and a lucky squirrel: Wildlife Photographer of the Year 2026 shortlist**](https://www.bbc.com/news/articles/cg7m5e3mylpo?ref=labnotes.org) Some wildlife photos for your viewing pleasure. --- The world is changing. This weekend we got a post about the software industry emulating fast fashion, because AI makes software so cheap to produce. We'll also cover new businesses forming away from city centers, in part thanks to AI. And watch out, because some of the anti-AI sentiment you're seeing is coming from Effective Altruism and similar AI proponents. And have we reached a major milestone? AGI has been announced this week! We'll explore a wisecrack article to prove that AGI is at best thin marketing hype. --- ## Tech Stuff [**Plainva**](https://plainva.com/?ref=labnotes.org) Local-first Markdown editor that opens your existing Obsidian vaults. Includes Notion-style databases (table, gallery, Kanban, etc.) Can manage tasks and calendaring, import from Notion, Logseq, Craft, TiddlyWiki, and friends, save to OneDrive, Google Drive, Dropbox, Nextcloud, etc. Quite a lot of features for an open-source project. ![](https://labnotes.org/content/images/2026/09/image-1788628311071.png) [**Relative to your interests, Friday**](https://cote.io/2026/09/04/relative-to-your-interests-friday-6.html?ref=labnotes.org) Quote of the day: "MCP is SOAP for Zoomers" [**Tabularis**](https://tabularis.dev/?ref=labnotes.org) Some of you are going to love this app. It's a database GUI that can handle your typical SQL databases — MySQL, PostgreSQL, SQLite, etc — and can also handle CSV, ElasticSearch, MongoDB, Redis, and a few more. Beyond connecting to the database, editing values, running SQL queries, exporting data, and such, it can also visualize the DB schema, backup your database, explain the query plan, and has an AI assistant for writing queries. It supports MCP so you can access the database from your favorite AI app. I'm not 100% sold on the UI, but I do love the extensive feature set. ![](https://labnotes.org/content/images/2026/09/image-1788627746401.png) [**Unsubscribe links without a login**](https://belderbos.dev/blog/unsubscribe-without-login-django-signing/?ref=labnotes.org) I remember once talking to the PM of an app whose unsubscribe link insisted you sign in before you can confirm your unsubscribe. Something along the lines of "people accidentally clicking on email links". Yeah right. Anyway, here's to the opposite of a dark pattern: > Single-use is different, it needs memory. The signature can't tell you a token was already spent, so you'd have to record each accepted token and check that list. That's the state you're forced to add. > > An unsubscribe link is idempotent, clicking it twice does the same safe thing, so it can stay stateless. A login link can't: reusing it opens the account again. [**How I Use AI and Why I Think We Need to Learn to Grow With It**](https://www.david-duval.com/how-i-use-ai-and-why-we-need-to-grow-with-it/?ref=labnotes.org) > So my confidence hasn't really come from deciding that AI can't make mistakes. It's almost the opposite. I know it can make mistakes, and because of that I've tried to build systems where a mistake is less likely to become a disaster. > > That distinction becomes especially important when other people are using AI. ![](https://labnotes.org/content/images/2026/09/image-1788628415542.png) [**You have to beat the models at something**](https://seangoedecke.com/you-have-to-beat-the-models-at-something/?ref=labnotes.org) Yes. > This is a scary thought. But you're not doing yourself any favors by pretending that LLMs can't actually write code and it's all just a scam, or that LLM-written code is inherently so bad as to cause companies using it to collapse next year. **We are not going to wake up in 2027 to find that the AI craze is over and everyone is writing code by hand again.** You ought to put some serious thought into what you can do better than the models in the medium and long term. [**Token Monitor**](https://javis-ai.com/token-monitor/?ref=labnotes.org) Watch your token usage across time, models, apps, sessions, etc. This macOS menu bar app has a lot of display options you can tweak, I'm sure you'll find the combination that best suites your needs. ![](https://labnotes.org/content/images/2026/09/image-1788627636215.png) [**How to Favicon in 2026: Three files that fit most needs**](https://evilmartians.com/chronicles/how-to-favicon-in-2021-six-files-that-fit-most-needs?ref=labnotes.org) TL;DR prefer SVG over PNG, trust browsers to downscale, drop obscure formats. An exhaustive guide to favicons for modern web browsers. [**Software Is Fast Fashion.**](https://medium.com/google-cloud/software-is-fast-fashion-44bb892deb1a?ref=labnotes.org) The software industry is going to change in a big way. Not everybody is going to like that, yet more people will purchase fast fashion than not: > The illusion of software permanence is over. > > The winners of this era won’t be the ones trying to preserve code that was destined to expire. They will be the builders who recognize that **code is just cheap fabric on a seasonal rack.** > > Have the courage to ship the bold idea first, the wisdom to let your code be completely disposable, and the velocity to keep evolving long after the copycats run out of things to clone. ![](https://labnotes.org/content/images/2026/09/image-1788624621918.png) [**Typebase**](https://typebase.io/?ref=labnotes.org) Deploy a fully typed server your frontend calls like a local function. I have no idea if this project is useful or not, but I enjoyed reading the industry-shaking mock testimonials on its website: > “We replaced 14 microservices with one folder. The DevOps team threw me a party. The party was a meeting. The meeting was about layoffs.” > > — Anonymous — For obvious reasons > “I’ve told four separate therapists about Typebase. Two stopped taking me as a client. The other two are now shipping an app with it.” > > — Clementine Ryu — Engineer, between therapists [**Indie Developer Simulator**](https://indiesimulator.com/?ref=labnotes.org) Perfection. Covers everything from Apple charging you $99, to building the 1000th todo app, to proper names being all taken, to useless coding AI assistants, to … (via [Casey Liss)](https://mastodon.social/users/caseyliss/statuses/117218037769599509?ref=labnotes.org) ![](https://labnotes.org/content/images/2026/09/image-1788624100976.png) --- ## Eye for Design [**Why Spliit Doesn't Ask You to Create an Account**](https://spliit.app/blog/why-spliit-doesnt-ask-you-to-create-an-account?ref=labnotes.org) There's a high chance your app doesn't need users to sign in at all: > Spliit's promise fits in one line, and it's been on the homepage from the start: *no account, no limitation, no problem*. You shouldn't have to hand over your identity to figure out who owes what for pizza. > > Ten seconds from link to expense. That's the product. **The missing login screen is not a gap in it — it's the point.** ![](https://labnotes.org/content/images/2026/09/image-1788627474300.png) [**Flock of Cats**](https://famichiki.jp/users/FlockOfCats/statuses/117205231766164369?ref=labnotes.org) > Modern software annoyance No. 37361: apps that display text but don’t let you easily select a part of the text. > > I feel like this is especially egregious in messaging apps. > > “Copy message” is nice to have but please also let me highlight a word to look up, copy, etc. > > And I don’t wanna hunt through the menu to find “select text” to make it selectable. It’s text — it should always be selectable! [**Nevermind**](https://github.com/lamzhonghang/nevermind?ref=labnotes.org) A charming font if you ask me. ![](https://labnotes.org/content/images/2026/09/image-1788627660312.png) --- ## Peoples [**Specs Don't Drive Anything**](https://leadership.garden/spec-driven-delusion/?ref=labnotes.org) The DRI: > Give validation a name. One person per initiative, accountable for the question "Was this worth building," with the standing to answer no and kill it. A named individual, not a team or a pod or a triad, because shared accountability for that question has never once produced someone willing to say the word no in a room with a director in it. [**Paul Cantrell**](https://hachyderm.io/@inthehands?ref=labnotes.org) I think this is true more often than we'd like to admit: > (2) *Identity problem:* developers used to feel special because generating code was placed on a cultural pedestal (sign of a special third brain lobe or whatever), that perceived specialness was a pillar of our personal identity, and now our egos are bruised because machines can do it and we don’t feel special anymore. --- ## Business Side [**Nvidia’s investments grow to $99 billion as chip giant becomes major backer of AI companies**](https://www.cnbc.com/2026/09/04/nvidia-ai-investments-99-billion.html?ref=labnotes.org) Think about it for a second. You manufactur a powerful chip for a few dollars, then sell it for a few hundred, so your customer needs an investment to purchase your chips. You invest in your customer, trusting they use your money to buy your chips. Since the investment has an ROI, you make 10x when you invest your profit margins. Add to that, you're too big to fail, so worse that can happen is you get a bail out. [**Busyness Can Be a Form of Avoidance**](https://denver-frederick.com/2026/08/28/notes-114-the-future-often-arrives-looking-inferior/?ref=labnotes.org) So true! > In fact, excessive activity can protect us from the harder work of deciding what actually matters. When everything is important, prioritization disappears. When every request earns a meeting, the calendar begins determining the strategy rather than serving it. > > The deeper discipline of leadership may therefore be subtraction. [**Invisible Companies** ](https://colossus.com/article/invisible-companies/?ref=labnotes.org) An interesting article about companies making money without an obvious moat: > The businesses Ross bought and started did not have moats. Funeral parlors, parking lots, rental cars, and the rest are easy industries to compete in. You can tell because they are crowded with competitors. Yet in 1969, Ross bought Warner for roughly $400 million—about $3.5 billion today. Without any sort of moat, Ross should not have been able to accrue the wealth needed to buy one of the storied movie studios of the ages. But he did. [**AI and the city**](https://www.stripeeconomics.com/p/ai-and-the-city?ref=labnotes.org) According to Stripe, more new businesses are popping up in smaller places. Even in big cities, they're forming in outer suburbs instead of downtown. This trend started during COVID-19 and sped up since AI came along. ![](https://labnotes.org/content/images/2026/09/image-1788624701107.png) --- ## Machine Intelligence [**One CEO in eight got what AI was sold as, survey**](https://cote.io/2026/08/28/one-ceo-in-eight-got-what-ai-was-sold-as.html?ref=labnotes.org) A year of AI spend, and 42% of CEOs report that nothing moved: costs flat, revenue flat. 12% got both benefits at once. 13% came out worse. ![](https://labnotes.org/content/images/2026/09/image-1788627875415.png) [**OpenAI: GPT-6 is totally Artificial General Intelligence, guys**](https://pivot-to-ai.com/2026/09/04/openai-gpt-6-is-totally-artificial-general-intelligence-guys/?ref=labnotes.org) ChatGPT — the [latest ad platform](https://www.aitooldiscovery.com/guides/chatgpt-ads?ref=labnotes.org) provider — is now considering itself AGI, or as this article captures it so succccinctly: > The purpose of benchmarks is to generate headlines for press releases. So I’m looking through OpenAI’s announcement and the GPT-6 model card. And it’s all benchmarks. [**Your guide to astroturf fake anti-AI groups**](https://pivot-to-ai.com/2026/09/02/your-guide-to-astroturf-fake-anti-ai-groups/?ref=labnotes.org) Some people are mad about AI. So AI companies made fake groups that pretend to be against AI. These groups don't really want to stop AI, they just want to look like they care. One group is called Irreplaceable. They pay people to get students to walk out of class. Another group called Eko says "stop AI" but actually wants more AI. A group called Pause AI thinks robots will kill everyone someday. That's not real. Real anti-AI groups talk like normal people. Fake ones sound like ads or cults. If a group says "pause" instead of "stop," watch out. (I'm not saying being anti-AI is wrong, I'm just alerting you to watch where the anti-AI sentiment is coming from) ![](https://labnotes.org/content/images/2026/09/image-1788624804203.png) --- ## Insecurity [**FBI Probes Service Selling 153M+ Drivers Licenses**](https://krebsonsecurity.com/2026/09/fbi-probes-service-selling-153m-drivers-licenses/?ref=labnotes.org) Apparently IDscan.net was breached and driver licenses were exposed. This service is used by Hertz, Target, Fedex, marijuana dispensaries, Caesars and perform more than 21 million verifications a month, at more than 20,000 locations worldwide. (via [Julie Webgirl)](https://mstdn.social/users/juliewebgirl/statuses/117202087069031050?ref=labnotes.org) [**Anatomy of a scam campaign, from the point of view of a link shortener**](https://casparwre.de/blog/anatomy-of-a-scam-campaign/?ref=labnotes.org) Interesting read: > A scam ad operation pushed 89,826 clicks through Zip1.io in 48 hours. What it looked like from inside the shortener: how the links hid what they were doing, and why deleting them only helped the operator. ![](https://labnotes.org/content/images/2026/09/image-1788625225032.png) --- ## Everything Else **[Elissa](https://eigenmagic.net/users/vampiress/statuses/117182150851919630?ref=labnotes.org)** > To write this I busted out my iPad and my typewriter style keyboard and shoved on a vinyl because Sunday. > > The irony was I couldn’t find a decent iPad minimalist text writing app so I just ssh’d into my Linux server and wrote it in vim. ![](https://labnotes.org/content/images/2026/09/image-1788627968675.png) **[Tay](https://transfem.social/notes/aqie4r7w017w1963?ref=labnotes.org)** > at the POST office trying to GET my mail **[highly praised and little read](https://posts.rat.pictures/users/hannah/statuses/117215348695729147?ref=labnotes.org)** > Too bad its never like “our ai agent went rogue and deleted ten billion dollars of medical debt” [**The longer version is sometimes easier to read**](https://www.plainfinance.net/break-up-long-sentences-2/?ref=labnotes.org) Shorter does not mean more readable: > This is why the delete-key reflex so often disappoints. You cut and cut, the word count drops, and the writing still reads like treacle. Cutting words from a badly built sentence gives you a shorter badly built sentence. The fix is structural, and structure sometimes costs a few more words. > > You can go a long way in the other direction, where the setting allows it. The whole idea will compress to five words. ![](https://labnotes.org/content/images/2026/09/image-1788627400430.png) [**Lucas**](https://chaos.social/users/lucaslove/statuses/117215707861342311?ref=labnotes.org) > I forgot how much work it takes to make great software. > > Having fun. Just wish I had more time. [**Zalka Csenge Virág, PhD**](https://ohai.social/users/TarkabarkaHolgy/statuses/117184156420002977?ref=labnotes.org) > Bless the geologist going around the mineral show with a Geiger counter to see which crystals actually "radiate energy" 🤣 [**Ordinary Abundance**](https://ordinaryabundance.com/?ref=labnotes.org) Our modern life is full of abundance we take for granted because we can't imagine it any other way. This article explores many of these abundances. Here's one: > From Looking Backward, his utopian novel set in the year 2000\. Before recordings, listening to music meant having instruments or performers nearby, and favorite songs could go unheard for years at a time. ![](https://labnotes.org/content/images/2026/09/image-1788627904095.png) [**Client Challenge**](https://www.sfgate.com/local/article/sf-style-startup-tech-workers-22411062.php?ref=labnotes.org) Apparently Patagonia vests and Allbirds shoes are out of style … > “Most people I know find the SF dating scene very challenging,” she told me. “While that’s unfortunate on a personal level, it’s been good for our business because it’s given many men a reason to care about clothes.” [**The Singularity Is Not What It Seems**](https://www.theatlantic.com/technology/2026/09/ai-future-reckoning-singularity/688487/?ref=labnotes.org) This article is choke full of snarky yet wise quips — totally worth a read. It's also proof that we're no way near the singularity — name one LLM that can write an article like that? > Whether you believe that Silicon Valley is building a god or that large swaths of the economy are swept up in a mass delusion, some version of the AI future long promised by self-described "builders" in San Francisco has formed beneath our feet. Whatever it is, we're in it now. [**Angela**](https://mastodon.nz/users/seawall/statuses/117199181035272038?ref=labnotes.org) "Thank you, Leinster Chambers parking pay station, for this message of hope." ![](https://labnotes.org/content/images/2026/09/image-1788626959148.png)