terriblesoftware.org
What Actually Makes You Senior
I feel like takes like this are often reductive, but this one really resonates—not just with my own experience as a senior engineer but with what I see in senior+ engineers that I look up to.
Links that I find interesting
This is what is known among IndieWeb writers as a linkblog . It's a collection of links that I find interesting.
Most of the links I share are of a technical nature, but I also occasionally share links relating to UI/UX design or behavioral science.
Learn more here. You can also subscribe to this feed via RSS or JSON.
terriblesoftware.org
I feel like takes like this are often reductive, but this one really resonates—not just with my own experience as a senior engineer but with what I see in senior+ engineers that I look up to.
pavi2410.com
Today I learned there’s a JS+CSS API for applying custom formatting to specific text ranges inside a TextNode.
primer.style
Really interesting read about the accessibility downsides of toasts and some recommended alternatives.
millercenter.org
My wife (a history enthusiast) found several cool presidential speeches on this website from the Miller Center at UVA. This one from George Washington seemed timely for the beginning of November, with Thanksgiving coming up:
I do recommend and assign Thursday the 26th day of November next to be devoted by the People of these States to the service of that great and glorious Being, who is the beneficent Author of all the good that was, that is, or that will be. That we may then all unite in rendering unto him our sincere and humble thanks, for his kind care and protection of the People of this country previous to their becoming a Nation, for the signal and manifold mercies, and the favorable interpositions of his providence, … and in general for all the great and various favors which he hath been pleased to confer upon us.
This next part seemed particularly apropos this year. (Emphasis mine.)
And also that we may then unite in most humbly offering our prayers and supplications to the great Lord and Ruler of Nations and beseech him to pardon our national and other transgressions, to enable us all, whether in public or private stations, to perform our several and relative duties properly and punctually, to render our national government a blessing to all the People…
developers.cloudflare.com
This looks really interesting—a speech recognition model that detects when users are done speaking (rather than detecting pauses for that purpose, which isn’t the best UX). It’d be fun to cobble together a voice agent using this and astro-cloudflare-websocket.
blog.cloudflare.com
Wow, this seems extremely cool.
www.theregister.com
I really like these thoughts; they align closely with my personal experience. AI can amplify our abilities, but we shouldn’t delegate our thinking to it wholesale, or we shortchange ourselves in the long run. Reminds me of a quote from Walden:
Where is this division of labor to end? and what object does it finally serve? No doubt another may also think for me; but it is not therefore desirable that he should do so to the exclusion of my thinking for myself.
hyperclay.com
An app platform built around webpages with a script that gives them persistent DOM. So anything you change on the page (including via script) is persisted across refresh and for other viewers.
futurism.com
This is really interesting—apparently commonly-touted fixes for social media (chronological feeds, hiding social statistics, intentionally boosting diverse viewpoints, etc.) don’t prevent polarization in AI-simulated networks. Another piece of evidence for my hypothesis that it is the hyperconnectedness itself that is the core problem. A less dense social graph puts more negative pressure on anti-social viewpoints because people can’t as easily find people with whom they agree. (Polarization and the normalization of fringe viewpoints aren’t exactly the same, but they seem close.)
www.recraft.ai
Whoa neat. I think this (a canvas editor with natural language support) is a really cool direction for image editing to go in. Some things are so tedious to do manually.
thegrowtheq.com
This post makes an interesting case for how key measures of economic growth are actually contributing to decreased perceptions of personal mastery and meaning—making us “the tools of our tools” (as Thoreau puts it). I do think, however, that this growth likely brings with it more opportunity for meaning and mastery. It’s just that the default path for most people has become less masterful. Which is interesting to think about.
www.vaultrice.com
This is pretty neat looking—a BaaS with a simple key value API. Calling it NonLocalStorage is pretty cheeky 😆
docs.astro.build
Oooooooo neat.
This feature allows you to edit files directly in Chrome DevTools and have those changes reflected in your local file system via a connected workspace folder. This is useful for applying edits such as adjusting CSS values without leaving your browser tab.
yosefk.com
I really like this post. I’ve noticed this in my own usage of AI—it has no actual knowledge of reality. It just knows what realistic text sounds like.
mattwie.se
This is bonkers. 😂 I think I’ve hit Claude’s limits maybe once in the week or so I’ve had it so far. But from what I hear it works better in certain codebases than in others; in Lucid’s it’s been hit or miss so far.
developers.cloudflare.com
Whoa neat. Looks like git support is only for public repos but still super cool.
tsx.is
This is really nice for writing scripts in an Astro project. Need to update this site’s scripts to use it (right now they are just node scripts…).
every.to
A list of people who would benefit from Claude code:
Senior engineers tired of implementation details, tech leads who want to multiply their impact, anyone maintaining software who wishes they had three more of themselves, and non-technical founders held back by not knowing how to build.
I fit two and a half of those four descriptions 😂 so I’m pretty hyped to get access to Claude Code at work soon 🙌
hashbrown.dev
This looks interesting—a framework for generative UI.
www.raptitude.com
This is an interesting technique for managing moments of overwhelm or mental shutdown.
docs.astro.build
TIL there’s an Astro Docs MCP server. And it can be used with Copilot in Agent Mode. Neat.
www.anthropic.com
Yikes. The identity crisis bit is crazy. Further confirms something I increasingly believe: AI has no concept of what is real. It only knows what realistic text is like. Realistic text can be true to reality one line, then untrue to it the next—as in historical fiction, for example. AI has no knowledge of the difference there.
astro.build
Astro shipped experimental header-based CSP for static sites!
We will soon ship support for Cloudflare too, which will use the _headers file.
🙌
www.mostlymetrics.com
Not gonna lie, I wouldn’t even be that mad if SEO died. The whole industry is pretty slimy (and there are better ways of finding stuff online, such as curating your social graph and seeking word-of-mouth recommendations). Replacing SEO with word smoothies doesn’t seem great though.
www.tandfonline.com
The findings of the study indicated that the inclusion of the “Artificial Intelligence” term in descriptions of products and services decreases purchase intention, and that emotional trust mediates this relationship.
Completely unsurprising, but it’s nice to have an empirical source for it.
github.com
An interesting proposal for a native DOM templating API in JavaScript.
astro.build
Live Collections in Astro opens up some neat possibilities, especially if you combine them with Server Islands. Like comments for a blog. (It’d be neat if JamComments would natively support this! Although I guess a much simpler solution would be to make my JamComments wrapper component a Server Island and call it good… I might do that 😅 )
george.mand.is
Apparently you can speed up audio before transcribing it using OpenAI’s APIs to cut down the cost of transcription.
developers.cloudflare.com
Neat! A way to run AI-generated code from a Cloudflare Worker
calnewport.com
[I]t’s … now clear that this field is no longer actually divided on the question of whether, generally speaking, smartphones and social media are bad for kids. In this new study, almost every major claim about this idea generated at least majority support, with many being accepted by over 90% of the experts surveyed. There were close to no major claims for which more than a very small percentage of experts felt that there was contradictory evidence.
In social psychology, this might be as clear a conclusion as we’re likely to achieve. Combine these results with the strong self-reports from children and parents decrying these technologies and their negative impacts, and I think there’s no longer an excuse not to act.
web.archive.org
Apparently the Minecraft modding group Hypixel formed a game company and has been working on a Minecraft-like RPG for the last ten years, only to shut down yesterday without releasing anything at all. Must suck to have put ten years of work into something only to have it go literally nowhere.
webkit.org
Some neat stuff in the latest Safari, including scroll timelines, color-contrast, and this interesting change to the add-to-home-screen behavior on iOS:
By default, every website added to the Home Screen opens as a web app. If the user prefers to add a bookmark that opens in their default browser, they can turn off “Open as Web App”, even if the site is configured to be a web app.
open.spotify.com
The first song from CHAPPY’s debut album was released today! I contributed to the Kickstarter for the album back in 2020 after finding his only song at that point on Spotify and loving it. (Hard to believe that was five years ago!)
Oh
Oh the things I don’t know
Stack against my soul
Feeling comes and it goes
Can somebody tell me,
Is my biggest defeat that I’m still asleep?
Someone wake me
This album seems like it will be well worth the wait.
www.nytimes.com
“What does a human slowly going insane look like to a corporation?” Mr. Yudkowsky asked in an interview. “It looks like an additional monthly user.”
rwsdk.com
This looks neat. Redwood JS is pivoting to a deeply integrated Cloudflare framework.
remix.run
Seems very interesting. Reminds me of Taylor Hunt’s statement that the only thing he didn’t like about Remix is React. Maybe now he’ll like Remix?
trailbase.io
Neat—a self-hosted single executable Supabase alternative
vercel.com
This is neat. I wonder if it’s possible to do something similar with Cloudflare.
motion.dev
TIL Framer Motion is now just called Motion, and supports vanilla JS and Vue in addition to React.
kentcdodds.com
I liked this take:
Applying for a job now often means competing against a flood of AI-generated resumes and hoping your application makes it past automated filters. And on the flip side, hiring managers face the Herculean task of finding real candidates among piles of AI-crafted applications. And they’re increasingly using AI to do this 😬
That’s the kind of competition we’re dealing with. And yet, a personal connection—a real conversation—cuts through all of that.
www.youtube.com
A neat approach to paywalling a static site. Might eventually do this for Innerhelm, if I end up taking it in a Substack kind of direction to monetize it. (Lol maybe I can eventually create and sell a Substack clone Astro template.)
x.com
Whoa, Bolt let’s you open an existing GitHub repo now. That’s cool
www.youtube.com
A neat talk about how AI could enable different levels of abstraction within an interface. I wish more companies would do this.
claytonwramsey.com
I liked these thoughts.
If it’s not worth writing, it’s not worth reading.
www.newyorker.com
An extremely good essay on AI by a professor of humanities
blog.extensiontotal.com
[A]ny Chrome extension can exploit this. No special permissions required. If there’s a vulnerable MCP server running on a host machine, that’s it. We’ve already found vulnerable MCP servers tied to services like filesystem access, Slack, WhatsApp, and more. This isn’t just a theoretical risk anymore, it’s real, and the impact could be devastating.
Yikes.
dbushell.com
This guy’s personal site has an alternate for-LLMs version of each post that is just nonsense. Amazing. 😂
github.com
This is neat.
I made my AI think harder by making it argue with itself repeatedly. It works stupidly well.
open-web-advocacy.org
We [propose] the following six potential targeted changes:
- Cap Google’s default search deals to 50% per browser, excluding Apple whose contract should be canceled entirely.
- Introduce a carve-out for smaller browsers.
- Mandate 90% reinvestment of Google search revenues into web platform and browser development.
- Restructure Chrome as an independent subsidiary within Alphabet.
- Limit Chrome’s default search deal with Google to 50% of users and auction the remaining defaults to rival search engines.
- Enforce transparency and fair revenue share terms across all search deals.
These adjustments would still achieve the DOJ’s core goal: restoring meaningful competition in general search. Conservatively, we estimate these adjusted remedies would reduce Google’s U.S. search market share to below 50%, the threshold for presumed monopoly power.
Critically, though, rather than collapsing platform funding, these adjusted remedies would likely increase web platform investment by 150%, creating a healthier, more competitive, and more innovative internet ecosystem.
Forcing Google to sell Chrome is such a red herring of a solution. The DOJ can do much better.
zero.rocicorp.dev
Rocicorp’s sync engine Zero is in alpha now. Looks like it only supports Postgres, and only fully supports bare-bones Postgres hosts for now (meaning Neon, Supabase, and Fly.io all have limitations). Still looks pretty cool though.
www.nytimes.com
An interesting and very good read. I really like Sonuga-Barke’s proposed alternative model of ADHD as an environment-individual mismatch.
First, the new model more accurately reflects the latest scientific understanding of A.D.H.D. And second, it gives children a vision of their future in which things might actually improve — not because their brains are chemically refashioned in a way that makes them better able to fit into the world, but because they find a way to make the world fit better around their complicated and distinctive brains.
tsdown.dev
This looks nifty. Might use it to bundle my utilities repo as a package so I don’t need to set it up as a git submodule in all my projects. (I tried creating a TS package for it before, but the setup proved to be too much of a hassle to be worth it.)
hakojs.com
Whoa this is neat—a faster and more modern QuickJS alternative.
github.com
Cool—a React useState hook with an undo/redo stack.
youtu.be
SpacetimeDB looks really neat. 👀 I think I came across their website before, but from what this video says, it’s a lot cooler than I noticed before. It seems almost like Convex, but built for high scalability use cases like MMOs. Like a Convex-meets-Durable-Objects kind of thing.
fonts.google.com
A neat font—look at the W’s middle vertex. Saw this on the website for Bitcraft, an MMORPG powered by SpacetimeDB.
astro.build
Astro 5.7 includes an experimental Fonts API that generates optimized fallback fonts automatically. And it works with Fontsource! Nifty.
nerdy.dev
Yikes!! Adam Argyle was a huge part of Google contributing positively to the web platform. Alongside the imminent and ill-considered possibility of Google having to sell Chrome, this really does not portend a bright future for the Web in the next few years.
posthuman.blog
This was one heck of a ride, as was the sequel.
blog.cloudflare.com
Some super cool stuff here—new models, faster inference, improved LoRA fine-tuning support, and a nifty batch mode.
chef.convex.dev
Whoa. Pretty impressive if it can actually create a Slack clone from a single prompt. (Does that include billing, organizations, etc? I wonder.)
developers.cloudflare.com
Cloudflare launched a beta of read replication in their D1 database offering! It includes a slick Sessions API to guarantee sequential consistency among queries in a given session.
The speed at which Cloudflare ships cool stuff is really impressive. Their changelog RSS feed is almost too noisy for my reader app!
www.drawdb.app
This is super neat—a tool for modifying a DB schema as a diagram, which you can then export as a DDL script. I think I came across it before when it was in beta; it seems pretty full-featured now. And it’s free!
I’m pleasantly surprised to see that it works with MS SQL Server—a lot of FOSS database tools don’t support SQL Server, in my experience.
developers.cloudflare.com
Cloudflare Durable Objects are usable on the Workers free tier now! I’m not sure I’ll have reason to use it anytime soon, but as someone who spins up lots of little hobby projects, this is great news.
developers.cloudflare.com
Cloudflare supports source maps now! Super cool. I’ll have to update my projects, and my Astro Cloudflare starter, to include that setting in their configuration.
astro.build
Server side logic on Astro + Cloudflare (and on Cloudflare generally) just got so much cleaner to write! No need to do hacky stuff like instantiating DB clients in middleware (which is what I ended up needing to do when trying to use Better Auth with Cloudflare back in January).
easingwizard.com
This is a nifty tool for creating CSS easing curves. Unfortunately, CSS easing is one of those things I do just infrequently enough that I’ll probably forget about this by the time I need it next. Hopefully sharing it here will make it stick in my memory until then (or at least make it easier to find again).
sqlsync.dev
This approach to a syncing offline data store seems very interesting. Dealing with data migrations still seems like it would be a headache, though. (I accidentally mistyped “migrations” as “migraines”… Not inaccurate!)
github.com
CSP support in Astro got moved to a Stage 2 proposal! Excited to see the work being done here.
gitdiagram.com
This is a neat concept—an AI tool that creates a diagram of a GitHub repository, which you can use by replacing hub with diagram in the repository URL. But every time I try it on my personal site’s repo I encounter an unresolvable Mermaid syntax error, which is annoying. They probably should have (1) included more-resilient error handling and/or caching between each step of the process and (2) included more hard-logic steps, such as creating an import graph of the code, instead of relying solely on AI to be the brains of their product.
developer.chrome.com
In my opinion, alongside the HTML Popover and CSS Anchor Positioning APIs (both of which this is built upon), this is one of the best things to hit the web platform recently. So exciting!
github.com
NPM had a Cloudflare misconfiguration issue that prevented users from looking up or downloading packages with camel in the name. This affected a lot of utilities for changing text casing (which includes the two mentioned in the issue title). What a bug to have on April Fools’ day! 🐪
www.stepsecurity.io
Yikes! A secret-exfiltration attack was snuck into a GitHub Actions action I’ve used in some of my pipelines—three hours after I last used it.
Fortunately, while I’ve used it a lot in the past, the only place I still used it as of the time of the incident was my private repo for Innerhelm), and only public repos are affected, since the exfiltration mechanism was double-base64 encoded logging. (“There is no evidence that the leaked secrets were exfiltrated to any remote network destination.” Phew!)
I’ve removed that action from Innerhelm’s workflow now, but the real takeaway here is to pin your actions to specific commit hashes (which I have also done for the single remaining third-party action in my pipelines).
danmall.com
Some good advice here. I especially like the point about asking for a specific favor instead of being vague or broad; it seems to be an application of the “give three examples” principle I wrote about last week. Being concrete and specific makes your request much more actionable.
manifest.build
An interesting Backend-as-a-Service that uses a single config file to set up database entities, authentication, webhooks, and other common backend primitives. Feels a bit like a backend-only version of Wasp. The built-in admin panel functionality looks super cool.
github.com
This is neat—an MCP server for Cloudflare’s APIs, so you can use natural language commands to manage your Cloudflare resources via Claude Desktop.
kraftstories.com
An interesting example of taking a typically B2C idea (crafts and craft kits) and making it B2B (workshops for teams). My team at work is doing one of their workshops for a remote team activity in a couple weeks.
(The name makes me wince, though—using “craft” with a k sounds like a good way to manifest lawyers into your life.)
www.squibler.io
A web app that helps you write—by the looming threat of what you’ve written being deleted if you stop writing before a set timer is up. Maybe I should use this for my think-into-keyboard level 1 writing. 😆
(Before it was acquired by Squibler, it was an open-source project by Manu Ebert. The source code is still available on GitHub, and the original application has been helpfully preserved by GitHub user betaveros, who forked Manu’s repository.)
www.rippling.com
After Rippling’s security team detected the spy’s unusual systems activity, Rippling set a honeypot for Deel’s senior leadership to conclusively prove their involvement. Rippling crafted a letter that referenced an empty Slack channel in Rippling’s corporate Slack instance called “d-defectors,” and implied the Slack channel contained messages that would be of interest to Deel. The letter was sent to only three people – Phillipe Bouaziz, the chairman of Deel’s board, CFO, General Counsel, and the father of Deel CEO Alex Bouaziz; Spiros Komis, Deel’s Head of US Legal; and the company’s outside counsel at law firm. Within hours of sending the letter, Deel’s spy inside of Rippling searched – for the first time – for this empty and never-before-used Slack channel, proving that Deel’s top executives or its legal representatives were running the covert espionage operation.
Clever move on Rippling’s part.
the alleged spy, when confronted last Friday at Rippling’s Dublin office with a court order to hand over his phone, fled to the bathroom and locked the door. When repeatedly warned not to delete materials from his device and that his non-compliance could result in jail time, the spy responded: “I’m willing to take that risk,” and fled the premises.
The spy, on the other hand, …
Apparently this is not the first batch of drama between these two companies, either.
growsmethod.com
An interesting technique for incremental software development. Seems a bit difficult to build an end-to-end-functional first version of any non-trivial feature in just a day though, especially with unit tests. But I like the idea of focusing on getting the main path through the system built quickly instead of completing each layer in series. Been trying that with an epic I’m leading at work and it’s going pretty well so far.
anchoreum.com
This is pretty neat—a Flexbox-froggy-like webgame (made by the same people) for learning CSS anchor positioning.
danmall.com
A take on AI that I agreed with much of. I especially liked this bit:
As compelling as this match was, [designer] on Webflow vs. [non-designer] in Lovable was a great appetizer.
The main course I want is [designer] on Webflow vs. [designer] on Lovable.
It‘s not specifically about Webflow or Lovable. Or Framer. Or ChatGPT. Or Cursor. The real question isn’t, “Can AI tools help non-designers?” We’ve seen they can.
The question is: Can they make elite designers even better?
…
AI raises the floor for non-designers. But can it raise the ceiling for designers? I’m betting the answer is yes
I think similarly about AI vs. software engineers. AI might replace engineers in some cases, but I think the real victor will be the engineers that know how to use AI—and when not to—to achieve both speed and quality in their work.
Put another way, the question “Will AI replace software engineers?” is premised on a false dichotomy. It was never either-or.
superglue.cloud
This is really neat: an API transform layer that lets you query third-party APIs with your own expected response schema. It uses an LLM to generate a transformation between the API response and your desired schema, and can even handle pagination for you.
herman.bearblog.dev
I like these thoughts from Herman Martinus (the creator of Bear) a lot.
During the early days of lockdown, my girlfriend got deep into gardening. She would spend hours nurturing her tomato plants, weeding, watering, puttering about. She would spend afternoons reading articles on permaculture, figuring out which plants grew well alongside other plants. Lavender with tomatoes kept the pests away, slight shade for the granadilla vine, some water, but not too much for the root vegetables. It was lovely to watch her coax tasty and nutritious food individually out of the ground. I can’t imagine full-scale farming being nearly as enjoyable.
That’s what I want from my products. I want to putter about, feel connected to the process, and have fun doing so. I want to make things that don’t scale. To see people tuck into them and enjoy them as people, not as stats.
I’m not building any web app side projects at the moment, but if/when I pick that up again, I would want to do something like this.
emnudge.dev
I liked these thoughts, though I was disappointed the author doesn’t give more examples of funny Markov Chain output and unfunny ChatGPT output.
This bit is especially good:
Without a ton of “prompt engineering”, it’s actually quite easy to spot if some paragraph was LLM generated. It sounds soulless. It’s the most average thing you could have possibly produced given the context.
Asking an LLM for an “original thought” is almost oxymoronic, if not just moronic. It was built with the express purpose of not doing that.
I recently reworded the copy text on the landing page for Innerhelm. Most of it is written by hand, but there was part of the page I struggled with and used ChatGPT to generate suggestions for. The copy I ended up using for that part feels very LLM-ish to me. I wonder how noticeable it is to others.
github.com
This abstraction layer for building on top of the Cloudflare dev platform seems interesting. I wonder how it compares to Cloudflare’s already-pretty-great DX, and whether it would be possible to use with Astro. (I’d imagine you’d have to build an Astro integration for it, since it seems like EdgeKit has to wrap your worker module’s default export.)
nut.new
An interesting agentic AI-powered code editor that’s focused on debugging. At first glance I thought it looked a lot like Bolt; turns out it’s a fork of it. Which makes the name a top-tier pun—it evokes the phrases “nuts and bolts” and “a tough nut to crack” (from it’s focus on debugging).
www.asad.pw
I’ve wondered about this—whether it’s more expensive to use OpenAI’s API or ChatGPT. Turns out ChatGPT is cheaper by far; its $20/mo cost gives you $14,000 worth of model usage! That’s wild. I’m definitely not surprised that the API is more expensive—it’s a B2B product whereas ChatGPT is B2C—but I wasn’t expecting it to be 700 times more expensive.
newsletter.posthog.com
A good read about activation as an important metric for a software company. In essence, the goal is to measure when your potential user becomes a definite user—and this is usually not as simple as “when they sign up,” since a lot of users sign up but then drop off quickly.
It seems to relate closely to Krystal Higgins’ Better Onboarding: everything the user does in your product is onboarding (or un-onboarding), which means you have to draw a line somewhere to define what activation means. At what point are they onboarded enough that they’re likely to stick around? Higgins calls these points “key behaviors”—things the rest of your product is directing the user towards.
starlight-kbd.netlify.app
This is neat—an Astro Starlight plug-in for documenting keyboard shortcuts. It allows you to specify separate shortcuts for Mac and Windows and will shows an OS selector next to the theme selector in the navigation.
www.wearenotsaved.com
The intro to a book by J.W. Richey that intentionally leans into the “This book should have been a blog post” trope, by providing a “chapter zero” that is the blog post version, and then expounding the ideas further in subsequent blog-post-like chapters. I’ve seen this pattern in other books (Extreme Ownership, Indistractible, Essentialism, and Deploy Empathy all seem to fit the mold, to varying degrees). I think it usually works pretty well for nonfiction. And it seems like it could be a natural extension of my writing process if I ever decide to write a book.
flappybird.scd31.com
This is bonkers—it uses HTML streaming to update a page dynamically (by streaming in new CSS) instead of using JS.
www.gilesthomas.com
Some good thoughts on the value of blogging. I liked this bit:
Your GitHub profile shows your contributions to open source and lets people know how well you can code. But your blog shows your contributions to knowledge, and shows how well you can think. That’s valuable!
Another point I’d add is that blogging constitutes a decentralized, global conversation that can yield new connections—both between ideas and between people. As David Perrell (used to?) put it in “The Ultimate Guide to Writing Online”:
Content builds on itself. It multiplies and compounds. Day and night, your content searches the world for people and opportunities.
github.com
A neat jq-like CLI tool for querying Markdown files. Seems like it’d be useful with Obsidian. It’d be neat if Obsidian incorporated mdq’s query syntax into the in-app search.
github.com
This seems interesting. Although, ever since building Logwise, indie database projects that claim to make offline-first “easy” make me very suspicious—making such a claim probably means there’s some difficult part of decentralized state management that they’re overlooking.
newsletter.posthog.com
Some good stuff here, such as point #4:
Google is dumber than you think
[Y]ou only need to know two things:
- Google isn’t good at understanding content. When Google crawls your website, it’s creating a summary of what it’s about based on keywords, metadata, and links. But it doesn’t read and evaluate how good it is like a human.
- Google learns what’s good by observing its users. If you click on a result and then go straight back to Google, it was a bad result, and vice versa. Google uses this data to fine tune results so the results people like appears [sic] at the top.
readwise.io
Readwise’s streak-recovery feature is such an awesome solution to the downsides of streaks.
I’m reminded of a quote from James Clear’s Atomic Habits:
Missing once is a mistake. Missing twice is the start of a new habit.
Being able to recover a streak seems like a good way to disincentivize “missing twice.”
github.com
This is neat: a streaming data transfer format that supports things like Promises, Dates, etc.
readwise.io
An interesting and relatable read about someone’s experience being diagnosed with and learning about ADHD. This statistic about untreated ADHD’s impact on life expectancy was startling:
A 2015 Lancet study found that people with untreated ADHD die, on average, 9.5 years earlier than their peers. Not from the condition itself, but from its cascade of negative effects: accidents, impulsive decisions, and self-medication.
I recently read a book called ADHD 2.0 that provided a neat and accessible explanation of our modern understanding of ADHD and how it works. (Or, more accurately, I started reading it, only to get distracted and not finish it before the Libby loan I had ran out, which is painfully ironic. 😂) I highly recommend reading it for anyone who has ADHD themselves or spends a lot of time with someone who has it.
www.henrikkarlsson.xyz
I liked these thoughts by Henrik Karlsson (who I’ve quoted before). He writes about having an “anxious narrator” part of his mind, and explains how he deals with it. I won’t quote it here—it’s a handful of paragraphs long, and an excerpt won’t do the whole passage justice—but it’s worth a read.