Latest Posts (20 found)
Jim Nielsen 1 weeks ago

Book Notes: “Poor Charlie’s Almanack”

I’ve been slowly listening to Poor Charlie’s Almanack: The Essential Wit and Wisdom of Charles T. Munger . I like his practicality. He’s never trying to be overly academic, as if he needs to prove how smart he is. He says Berkshire’s success doesn’t come from them solving hard problems, but from spending their time knowing what a simple solution looks like — and acting on it when they see it! We’ve succeeded by making the world easy for us, not by solving the world’s hard problems. Munger analogizes their approach to investing like jumping a fence. They don’t spend all their time trying to figure out how to jump a seven-foot tall fence. Instead, they find a spot where the fence is only a foot tall, jump it, and take the reward on the other side. The approach he articulates for investing, in fact, seems broadly applicable to any kind of problem solving: Whenever people ask him for advice (as if somehow he could bestow upon them some kind of knowledge that will save them the pain and hardship of experience) he seems anathema to the idea that you can live life without making lots of mistakes. To paraphrase Charlie: “I don’t want you to think that we have a method of learning that will prevent you from making mistakes. The best you can do is learn to make fewer mistakes than others. And then, when you inevitably do make mistakes, learn to acknowledge them and fix them quickly.” Straightforward. Practical. No bullshit. No ego. (Basically the opposite of everything I see on social platforms.) I quite enjoyed his perspective. Reply via: Email · Mastodon · Bluesky Quickly eliminate the universe of what not to do. Follow up with a multi-disciplinary attack on what remains. Act decisively when — and only when — the right circumstances appear.

career Career books Books
0 views
Jim Nielsen 1 weeks ago

Something’s Rotten in the State of macOS Icon Design

This is an iconic observation : If you put the Apple icons in reverse it looks like the portfolio of someone getting really really good at icon design This isn’t, however, just the story of Apple’s Creator Studio icons. It’s the unfolding story of icon design across the entire macOS platform. For example, take a look at some of Apple’s other apps like iMovie : Or Remote Desktop : Apple sets the standard (and the rules) for how icons look on the Mac. Wherever they go, so goes the ecosystem — and they’re taking the entire ecosystem along down with them. It’s fast becoming the case that if you put any Mac app’s icons in reverse, it looks like the portfolio of someone getting really, really good at icon design. Even Microsoft — not exactly a bastion of design — starts to look pretty decent with their icons the further back you go. For example, with OneNote , the app icon’s progression looks like it went something like this: Some 3rd-party apps continue to fight a good fight, even as Apple’s definition of what an icon should be — or what’s even possible — shrinks all around them. Apps like Capo (remember, these are reverse chronological ): Or BBEdit : Or Fantastical : Or Cot Editor : Everyone’s being put in a box squircle. The imposition is real. I don’t blame any of the 3rd-party app makers. Their designs have to play by Apple’s rules (or end up in icon jail ). World-class designers like Matthew Skiles or The Iconfactory are still out there striving for excellence, even as they’re hamstrung by the Mac’s latest rules. When it comes to icon design on the Mac, the sky is no longer the limit: Apple’s icon design sensibilities are. They set the examples of what world-class icon design should look like, but what do you do when the examples are no longer exemplary? Reply via: Email · Mastodon · Bluesky “I made this with AI” “I tried to make the AI one, but by hand myself” “I don’t need to be constrained by this squircle” “Hey, I’m getting better at this”

design Design
0 views
Jim Nielsen 2 weeks ago

Building Software Requires Digestion

Here’s Scott Jenson in his insightful piece “The Ma of a New Machine” : the chatbot interface [makes us] feel like deep cognitive work is happening. But the interface is fundamentally reactive. It spits complex text at you, you skim it quickly, and you immediately type a reaction to keep the momentum going. My hypothesis is that the very structure of the chatbot interface (type, read, type again) actively discourages reflection. When you are moving too fast, you get stuck in a groove. You literally need to take a break, step back, and basically step out of this groove so you can view the problem from a new angle. We’ve all walked away from a tough problem only to have the solution arrive unbidden into our thoughts later in the day. In my decades+ experience designing and developing software, I can’t count the number of times I’ve stepped away from a problem at the computer only to return and find the problem magically resolved in my brain. But the human-computer interaction of prompting doesn’t encourage the use of that skill in our subconscious. In fact, I think it actively discourages it (our tools shape us). Scott talks about this Japanese concept called “Ma” which is about deliberately creating pauses between things. He quotes Studio Ghibli director Hayao Miyazaki who says “if you just have non-stop action with no breathing space at all, it’s just busyness.” Here’s Scott (emphasis mine): Ma provides a framework for understanding that a pause is not a lack of work As humans we need pauses. We need space to breathe. We need time to digest. Pausing, breathing, synthesizing, digesting — these are all necessary work . “Digestion” is an interesting word here. Putting food in your body is merely the beginning of feeding yourself. Our bodies must digest that food, break it down, absorb it, and get rid of the waste. But that’s all happening mostly without our attentive oversight, so I guess it’s not “real” work — right? Building good, healthy software requires digestion. Reply via: Email · Mastodon · Bluesky

career Career programming Programming design Design
0 views
Jim Nielsen 2 weeks ago

Out With the JS, In With the HTML

I’ve been posting about how you can make lots of HTML pages and leverage navigations over in-page, JS-dependent interactions . Now I’m gonna post another example. On my icon sites, I have a little widget that allows you to resize the icons you’re looking at. Previously, I implemented this functionality as a web component that looked something like this: The attribute corresponded to an enumeration like which mapped to actual pixel dimensions like 64×64 or 512×512. When the little widget was clicked to render icons at a different size, JavaScript changed the attribute on the custom element. From there, the web component’s JS took over changing the dimensions of the children elements, their attributes, etc. It all worked pretty well. However, because that was a client-side solution to my otherwise entirely pre-rendered static site, it required some templating logic and data be duplicated and sent over the wire to every client. I didn’t love that for various reasons — like “Crap, I updated this one small part of how my icon list renders on the server, but forgot to tweak it on the client, so things are slightly broken now.” Then one day the thought hit me: instead of relying on JS to make that interaction work (click, execute JS, modify in-page DOM to a new list), what if I just made that interaction a navigation? Click, navigate to a new list. Instead of “every list of icons ships with some JS that allows them to re-render at four different sizes” I could do “every list of icons ships in four different sizes”. So I tried it. And guess what? Once I added some code to support CSS view transitions, I got a cool effect amongst the icons for free — that’s right, by removing code! Works nice on mobile too! I know I’m not doing anything particularly novel here, but as we continue to get new, powerful primitives on the web — like CSS view transitions — I find it really interesting to revisit basic patterns and explore what’s possible now that wasn’t previously. It’s fun to ask yourself: “Could I remove some client-side JS and get a better overall experience?” If the answer is yes, I’ll bet you the development experience (and maintenance burden) is much improved too! Reply via: Email · Mastodon · Bluesky Previously: one page, like , with JS to re-render the icon list based on user interactions. Idea: four pages, like , each a different icon list size.

javascript JavaScript css CSS web-development Web Development html HTML frontend Frontend
0 views
Jim Nielsen 3 weeks ago

Reminder: You Can Stitch Together Lots of Little HTML Pages With Navigations For Interactions

I wrote about building websites with LLMs — (L)ots of (L)ittle ht(M)l page(s) — and I think it’s time for a post-mortem on that approach: I’ve tweaked a few things from that original post but the underlying idea is still the same, which I would describe as: Avoid in-page interactions that require JavaScript in favor of multi-page navigations that rely on HTML and are enhanced with CSS view transitions (and a dash of JS if/where prudent). As an example, on my blog I have a “Menu”. It doesn’t “expand” or “slide out” or “pop in” or whatever else you can do with JS. Instead, it navigates to an entirely-new page that is focused on just the menu options of my site. I say “navigates” because it’s just a link — — and it functions like a link, but the navigation interaction is enhanced by CSS view transitions. Have a newer device with a modern browser? Great, you get a nicer effect. Have an older device, or an older browser, or JS disabled, Et al.? It’ll still work. If you can follow a link — which is the most fundamental thing a browser can do — it will work. So how’s it all work under the hood? In essence, all the pages have a link to the menu (except the menu page). When you navigate to the menu, that link is changed to an “X” which “closes” the menu. The closing is still just a link (back to ) but it’s enhanced with JS to actually do a “back” in the browser history. This makes it so “opening/closing” the menu doesn’t add an entry to your browser history. As a simplified example, the code looks like this: The checks whether we came to this page as a navigation (mostly likely from within the blog itself) or via a direct visit (i.e. somebody typed it into the URL bar, unlikely but possible) which is how I suss out whether there’s a meaningful run or not. Here’s a video of how it all works, if that’s your thing: While this solution seems simplistic, it was not a simple thing to arrive at. It required me to spend time thinking about what was essential to navigation, how that interaction could work across multiple pages, and how I could ensure page size stayed small so the interaction was both fast and robust while remaining intuitive to use. In other words, the approach shaped the design. Turns out, if you have a website and you think of the browser as a way to navigate documents — rather than a runtime to execute arbitrary code and fetch, compile, and present them — things can be a lot simpler than our tools often prime us to make them. Reply via: Email · Mastodon · Bluesky

frontend Frontend web-development Web Development css CSS tutorial Tutorial html HTML
0 views
Jim Nielsen 1 months ago

Collective Speed Is Not the Summation of Individual Speed

I’ve been thinking about speed which is why Chris Coyier caught my attention in his latest piece discussing how AI might be 10✕ing the speed with which we code, but it’s not making our software 10✕ better: Faster individuals don’t make a fast company My mind immediately went to the 4✕100 relay at the Olympics. (Not sure which race that is? Watch the London 2012 one .) Imagine you were put in charge of winning the 4✕100 relay. All you gotta do is find the four faster sprinters in your country — right? I’m no track and field expert, but I doubt it’s that simple. In a relay race, the baton is arguably the most critical element. Passing it cleanly is vital because if you fumble it you’re easily behind a few meters or maybe even disqualified. So, one could argue, a sprinter’s ability to pass and receive the baton is more important than speed because all the speed in the world won’t help you overcome a dropped baton. (There are other considerations too, like which leg each runner takes, which sequence works best given individual pairings and rapport, and whether a slower veteran might perform better in the heat of the moment.) Faster runners won’t guarantee a faster team. And faster coders won’t guarantee a faster company. Like a relay race, it might be worth giving some thought to the relationships and interfaces between people. Reply via: Email · Mastodon · Bluesky

business Business
0 views
Jim Nielsen 1 months ago

Hook It Up to the Machine

In the early 2000’s, my parents took us on a road trip to Glacier National Park in Montana. We made the journey in our new (used) family van: a green Dodge Caravan whose reputation was soon to become “a lemon”. I was a teenager and didn’t pay a lot of attention to the details of what was happening around me, but I do remember how the van kept overheating. It ran fine on the interstate, but anything under 40MPH had the car’s temperature gauge rising into unsafe zones. I remember stopping in some small town in Montana to get it checked out by a mechanic. He checked it out, took it for a test drive, etc., and told my Dad the reason the car was overheating was because the idling fan wasn’t turning on. At higher speeds, like on the interstate, that was fine because there was enough airflow to keep the engine cool but at lower speeds the car would overheat. The mechanic said he didn’t know why the fan wasn’t turning on. There was nothing wrong mechanically from what he could see. But he couldn't fix it. He told my Dad that this was one of those increasingly common “computerized” cars that you have to hook up to another computer to diagnose the source of the issue. And he didn’t have one of those computers. So we continued on our way. The rest of the trip required my Dad taking “the long way around”, like back roads where he could keep up his speed in order to avoid the car overheating. It was all very amusing to us as kids, almost thrilling because Dad had a legitimate excuse to drive fast (suffice it to say, Mom did not like this). Once the trip was over and we returned home, my Dad was able to get the car in to a dealer where they hooked up the car’s computer to another computer to diagnose and fix the issue. I don’t really remember the specifics, but the issue was seemingly some failed digital sensor that prevented the idling fan from turning on. Once the sensor was replaced, things worked again. Computers talking to computers. Growing up in an era that shifted so many things from analog to digital, mechanical to electronic , I’ve thought about this trip a lot. And I’m thinking about it again in this new era of building software with LLMs. I think about that mechanic. This guy who grew up around mechanical cars that could be physically inspected, diagnosed, and repaired. So much of his experience and knowledge unusable in the face of a computerized car. You can tell when a mechanical switch has failed with your eyes, but not a digital one. You need a computer to help you understand the computer. Will this be my future? If a codebase was made with the assistance of an LLM, will its complexity and bugs only be inspectable, understandable, diagnosable, and fixable with an LLM? “Hey, can you help me, there’s a problem with my codebase?” “Ok, I can confirm the issue, but I can’t fix it without hooking your codebase up to an LLM.” Reply via: Email · Mastodon · Bluesky

Electronics Automotive hardware Hardware
0 views
Jim Nielsen 1 months ago

Speed is Not Conducive to Wisdom

Speed has become the primary virtue of the modern world. Everything is sacrificed to it. Move fast (and break things, not as a goal but as a consequence). Wisdom requires allowing yourself to be undone by experience: Experiencing these can be slow and uncomfortable, but if you keep up your speed you can outrun them — never reflecting on what happened in your wake. Speed is how you avoid reckoning. It guarantees you miss things, and you can’t learn from what you don’t notice. Wisdom’s feedback loop is slow. Wise people I’ve met seem unhurried. I don’t think it’s because they’re slow thinkers or actors. I think it’s because they’ve learned that important things take the time they take, no amount of urgency changes that. Wisdom is chasing all of us, but we’re going too fast to notice what it’s trying to teach us. Reply via: Email · Mastodon · Bluesky An opinion dismantled by reality. An artifact torn apart by the real world. An idea destroyed by its own shortsightedness.

career Career culture Culture
0 views
Jim Nielsen 1 months ago

Fewer Computers, Fewer Problems: Going Local With Builds & Deployments

Me, in 2025, on Mastodon : I love tools like Netlify and deploying my small personal sites with But i'm not gonna lie, 2025 might be the year I go back to just doing builds locally and pushing the deploys from my computer. I'm sick of devops'ing stupid stuff because builds work on my machine and I have to spend that extra bit of time to ensure they also work on remote linux computers. Not sure I need the infrastructure of giant teams working together for making a small personal website. It’s 2026 now, but I finally took my first steps towards this. One of the ideas I really love around the “local-first” movement is this notion that everything canonical is done locally, then remote “sync” is an enhancement. For my personal website, I want builds and deployments to work that way. All data, build tooling, deployment, etc., happens first and foremost on my machine. From there, having another server somewhere else do it is purely a “progressive enhancement”. If it were to fail, fine. I can resort back to doing it locally very easily because all the tooling is optimized for local build and deployment first (rather than being dependent on fixing some remote server to get builds and deployments working). It’s amazing how many of my problems come from the struggle to get one thing to work identically across multiple computers . I want to explore a solution that removes the cause of my problem, rather than trying to stabilize it with more time and code. “The first rule of distributed computing is don’t distribute your computing unless you absolutely have to” — especially if you’re just building personal websites. So I un-did stuff I previously did (that’r right, my current predicament is self-inflicted — imagine that). My notes site used to work like this : It worked, but sporadically. Sometimes it would fail, then start working again, all without me changing anything. And when it did work, it often would take a long time — like five, six minutes to run a build/deployment. I never could figure out the issue. Some combination of Netlify’s servers (which I don’t control and don’t have full visibility into) talking to Dropbox’s servers (which I also don’t control and don’t have full visibility into). I got sick of trying to make a simple (but distributed) build process work across multiple computers when 99% of the time, I really only need it to work on one computer. So I turned off builds in Netlify, and made it so my primary, local computer does all the work. Here are the trade-offs: The change was pretty simple. First, I turned off builds in Netlify. Now when I Netlify does nothing. Next, I changed my build process to stop pulling markdown notes from the Dropbox API and instead pull them from a local folder on my computer. Simple, fast. And lastly, as a measure to protect myself from myself, I cloned the codebase for my notes to a second location on my computer. This way I have a “working copy” version of my site where I do local development, and I have a clean “production copy” of my site which is where I build/deploy from. This helps ensure I don’t accidentally build and deploy my “working copy” which I often leave in a weird, half-finished state. In my I have a command that looks like this: That’s what I run from my “clean” copy. It pulls down any new changes, makes sure I have the latest deps, builds the site, then lets Netlify’s CLI deploy it. As extra credit, I created a macOS shortcut So I can do , type “Deploy notes.jim-nielsen.com” to trigger a build, then watch the little shortcut run to completion in my Mac’s menubar. I’ve been living with this setup for a few weeks now and it has worked beautifully. Best part is: I’ve never had to open up Netlify’s website to check the status of a build or troubleshoot a deployment. That’s an enhancement I can have later — if I want to. Reply via: Email · Mastodon · Bluesky Content lives in Dropbox Code is on GitHub Netlify’s servers pull both, then run a build and deploy the site What I lose : I can no longer make edits to notes, then build/deploy the site from my phone or tablet. What I gain : I don’t have to troubleshoot build issues on machines I don’t own or control. Now, if it “works on my machine”, it works period.

web-development Web Development cloud Cloud devops DevOps
0 views
Jim Nielsen 1 months ago

Prototyping with LLMs

Did you know that Jesus gave advice about prototyping with an LLM? Here’s Luke 14:28-30: Suppose one of you wants to build a tower. Won’t you first sit down and estimate the cost to see if you have enough money to complete it? For if you lay the foundation and are not able to finish it, everyone who sees it will ridicule you, saying, ‘This person began to build and wasn’t able to finish.’ That pretty much sums me up when I try to vibe a prototype . Don’t get me wrong, I’m a big advocate of prototyping . And LLMs make prototyping really easy and interesting. And because it’s so easy, there’s a huge temptation to jump straight to prototyping. But what I’ve been finding in my own behavior is that I’ll be mid-prototyping with the LLM and asking myself, “What am I even trying to do here?” And the thought I have is: “I’d be in a much more productive place right now if I’d put a tiny bit more thought upfront into what I am actually trying to build.” Instead, I just jumped right in, chasing a fuzzy feeling or idea only to end up in a place where I’m more confused about what I set out to do than when I started. Don’t get me wrong, that’s fine. That’s part of prototyping. It’s inherent to the design process to get more confused before you find clarity. But there’s an alternative to LLM prototyping that’s often faster and cheaper: sketching. I’ve found many times that if I start an idea by sketching it out, do you know where I end up? At a place where I say, “Actually, I don’t want to build this.” And in that case, all I have to do is take my sketch and throw it away. It didn’t cost me any tokens or compute to figure that out. Talk about efficiency! I suppose what I’m saying here is: it’s good to think further ahead than the tracks you’re laying out immediately in front of you. Sketching is a great way to do that. (Thanks to Facundo for prompting these thoughts out of me.) Reply via: Email · Mastodon · Bluesky

design Design ai AI
0 views
Jim Nielsen 1 months ago

I Tried Vibing an RSS Reader and My Dreams Did Not Come True

Simon Willison wrote about how he vibe coded his dream presentation app for macOS . I also took a stab at vibe coding my dream app: an RSS reader. To clarify: Reeder is my dream RSS app and it already exists, so I guess you could say my dreams have already come true? But I’ve kind of always wanted to try an app where my RSS feed is just a list of unread articles and clicking any one opens it in the format in which it was published (e.g. the original website). So I took a stab at it. (Note: the backend portion of this was already solved, as I simply connected to my Feedbin account via the API .) First I tried a macOS app because I never would’ve tried a macOS app before. Xcode, Swift, a Developer Account? All completely outside my wheelhouse. But AI helped be get past that hurdle of going from nothing to something . It was fun to browse articles and see them in situ . A lot of folks have really great personal websites so it’s fun to see their published articles in that format. This was pretty much pure vibes. I didn’t really look at the code at all because I knew I wouldn’t understand any of it. I got it working the first night I sat down and tried it. It was pretty crappy but it worked. From there I iterated. I’d use it for a day, fix things that were off, keep using it, etc. Eventually I got to the point where I thought: I’m picky about software, so the bar for my dreams is high. But I’m also lazy, so my patience is quite low. The intersection of: the LLM failing over and over + my inability to troubleshoot any of it + not wanting to learn = a bad combination for persevering through debugging. Which made me say: “Screw it, I’ll build it as a website!” But websites don’t really work for this kind of app because of CORS. I can’t just stick an article’s URL in an and preview it because certain sites have cross site headers that don’t allow it to display under another domain. But that didn’t stop me. I tried building the idea anyway as just a list view. I could install this as a web app on my Mac and I'd get a simple list view: Anytime I clicked on a link, it would open in my default browser. Actually not a bad experience. It worked pretty decent on my phone too. Once I visited my preview deploy, I could "isntall" it to my home screen and then when I opened it, I'd have my latest unread articles. Clicking on any of them would open a webview that I could easily dismiss and get back to my list. Not too bad. But not what I wanted, especially on desktop. It seemed like the only option to 1) get exactly what I wanted, and 2) distribute it — all in a way that I could understand in case something went wrong or I had to overcome an obstacle — was to make a native app. At this point, I was thinking: “I’m too tired to learn Apple development right now, and I’ve worked for a long time on the web, so I may as well leverage the skills that I got.” So I vibed an Electron app because Electron will let me get around the cross site request issues of a website. This was my very first Electron app and, again, the LLM helped me go from nothing to something quite quickly (but this time I could understand my something way better). The idea was the same: unread articles on the left, a preview of any selected articles on the right. Here’s a screenshot: It’s fine. Not really what I want. But it’s a starting point. Is it better than Reeder? Hell no. Is it my wildest dreams realized? Also no. But it’s a prototype of an idea I’ve wanted to explore. I”m not sure I’ll go any further on it. It’s hacky enough that I can grasp a vision for what it could be. The question is: do I actually want this? Is this experience something I want in the long run? I think it could be. But I have to figure out exactly how I want to build it as a complementary experience to my preferred way of going through my RSS feed. Which won't be your preference. Which is why I'm not sharing it. So what’s my takeaway from all this? I don’t know. That’s why I’m typing this all out in a blog post. Vibe coding is kinda cool. It lets you go from “blank slate” to “something” way faster and easier than before. But you have to be mindful of what you make easy . You know what else is easy? Fast food. But I don’t want that all the time. In fact, vibe coding kinda left me with that feeling I get after indulging in social media, like “What just happened? Two hours have passed and what did I even spend my time doing? Just mindlessly chasing novelty?” It’s fun and easy to mindlessly chasing your whims. But part of me thinks the next best step for this is to sit and think about what I actually want, rather than just yeeting the next prompt out. I’ve quipped before that our new timelines are something like: The making from nothing isn't as hard anymore. But everything after that still is. Understanding it. Making it good. Distributing it. Supporting it. Maintaining it. All that stuff. When you know absolutely nothing about those — like I did with macOS development — things are still hard. After all this time vibing, instead of feeling closer to my dream, I actually kinda feel further from it. Like the LLM helped close the gap in understanding what it would actually take for me to realize my dreams. Which made me really appreciate the folks who have poured a lot of time and thought and effort into building RSS readers I use on a day-to-day basis. Thank you makers of Feedbin & Reeder & others through the years. I’ll gladly pay you $$$ for your thought and care. In the meantime, I may or may not be over here slowly iterating on my own supplemental RSS experience. In fact, I might’ve just found the name: RxSSuplement. Reply via: Email · Mastodon · Bluesky Ok, I could use this on my personal computer. I don’t know that I’ll be able to iterate on this much more because its getting more complicated and failing more and more with each ask ( I was just trying to move some stupid buttons around in the UI and the AI was like, “Nah bro, I can’t.”) I have no idea how I’d share this with someone. I don’t think I’d be comfortable sharing this with someone (even though I think I did things like security right by putting credentials in the built-in keychain, etc.) I guess this is where the road stops. Nothing -> Something? 1hr. Something -> Something Good ? 1 year.

web-development Web Development swift Swift
0 views
Jim Nielsen 1 months ago

The Blandness of Systematic Rules vs. The Delight of Localized Sensitivity

Marcin Wichary brings attention to this lovely dialog in ClarisWorks from 1997: this breaks the rule of button copy being fully comprehensible without having to read the surrounding strings first, perhaps most well-known as the “avoid «click here»” rule. Never Register/​Register Later/​Register Now would solve that problem, but wouldn’t look so neat. This got me thinking about how you judge when an interface should bend to fit systematic rules vs. exert itself and its peculiarities and context? The trade-off Marcin points out is real: "Never Register / Register Later / Register Now" is fully self-describing and avoids the «click here» rule. However, it kills the elegant terseness that makes that dialog so delightful. “Now / Later / Never” is three words with no filler and a perfect parallel structure. It feels like one of those cases where the rule is sound as a guideline but a thoughtful design supersedes the baseline value provided by the rule. Rules, in a way, are useful structures when you don’t want to think more. But more thinking can result in delightful exceptions that prove better than the outcome any rule can provide. I suppose it really is trade-offs everywhere : As software moves towards “scale”, I can’t help but think that systematic rules swallow all decision making because localized exceptions become points of friction — “We can’t require an experienced human give thought and care to the design of every single dialog box.” What scale wants is automated decision making that doesn’t require skill or expertise because those things, by definition, don’t scale. Then again, when you manufacture upon inhuman lines how can you expect humane outcomes? Reply via: Email · Mastodon · Bluesky When you choose to make decisions on a case-by-case basis, the result can be highly-tailored to the specific context of the problem at hand. However, within a larger system, you can start to lose consistency and coherence across similar UX decision points. When you choose to make system rules override the sensitivities of individual cases, you can lose the magic and delight of finding waypoints tailored exclusively to their peculiarities.

design Design
0 views
Jim Nielsen 2 months ago

Continuous, Continuous, Continuous

Jason Gorman writes about the word “continuous” and its place in making software. We think of making software in stages (and we often assign roles to ourselves and other people based on these stages): the design phase, the coding phase, the testing phase, the integration phase, the release phase, and so on. However this approach to building and distributing software isn’t necessarily well-suited to an age where everything moves at breakneck speed and changes constantly. The moment we start writing code, we see how the design needs to change. The moment we start testing, we see how the code needs to change. The moment we integrate our changes, we see how ours or other people’s code needs to change. The moment we release working software into the world, we learn how the software needs to change. Making software is a continuous cycle of these interconnected stages: designing, coding, testing, integrating, releasing. But the lines between these stages are very blurry, and therefore the responsibilities of people on our teams will be too. The question is: are our cycles for these stages — and the collaborative work of the people involved in them — measured in hours or weeks? Do we complete each of these stages multiple times a day, or once every few weeks? if we work backwards from the goal of having working software that can be shipped at any time, we inevitably arrive at the need for continuous integration, and that doesn’t work without continuous testing, and that doesn’t work if we try to design and write all the code before we do any testing. Instead, we work in micro feedback loops, progressing one small step at a time, gathering feedback throughout so we can iterate towards a good result. Feedback on the process through the process must be evolutionary. You can’t save it all up for a post-mortem or a 1-on-1. It has to happen at the moment, evolving our understanding one piece of feedback at a time (see: Gall’s law , a complex system evolves from a simpler one). if code craft could be crystallised in one word, that word would be “continuous”. Your advantage in software will be your ability to evolve and change as your customers expectations evolve and change (because the world evolves and changes), which means you must be prepared to respond to, address, and deliver on changes in expectations at any given moment in time. Reply via: Email · Mastodon · Bluesky

programming Programming devops DevOps
0 views
Jim Nielsen 2 months ago

Code as a Tool of Process

Steve Krouse wrote a piece that has me nodding along: Programming, like writing, is an activity, where one iteratively sharpens what they're doing as they do it. (You wouldn't believe how many drafts I've written of this essay.) There’s an incredible amount of learning and improvement, i.e. sharpening , to be had through the process of iteratively building something. As you bring each aspect of a feature into reality, it consistently confronts you with questions like, “But how will this here work?” And “Did you think of that there ?” If you jump over the process of iteratively building each part and just ask AI to generate a solution, you miss the opportunity of understanding the intricacies of each part which amounts to the summation of the whole. I think there are a lot of details that never bubble to the surface when you generate code from English as it’s simply not precise enough for computers . Writing code is a process that confronts you with questions about the details. If you gloss over the details, things are going to work unexpectedly and users will discover the ambiguity in your thinking rather than you (see also: “bugs”). Writing code is a tool of process. As you go, it sharpens your thinking and helps you discover and then formulate the correctness of your program. If you stop writing code and start generating it, you lose a process which helped sharpen and refine your thinking. That’s why code generation can seem so fast: it allows you to skip over the slow, painful process of sharpening without making it obvious what you’re losing along the way. You can’t understand the trade-offs you’re making, if you’re not explicitly confronted with making them. To help me try to explain my thinking (and understand it myself), allow me a metaphor. Imagine mining for gold. There are gold nuggets in the hills. And we used to discover them by using pick axes and shovels. Then dynamite came along. Now we just blow the hillside away. Nuggets are fragmented into smaller pieces. Quite frankly, we didn’t even know if there were big nuggets or small flecks in the hillside because we just blasted everything before we found anything. After blasting, we take the dirt and process it until all we have left is a bunch of gold — most likely in the form of dust. So we turn to people, our users, and say “Here’s your gold dust!” But what if they don’t want dust? What if they want nuggets? Our tools and their processes don’t allow us to find and discover that anymore. Dynamite is the wrong tool for that kind of work. It’s great in other contexts. If you just want a bunch of dust and you’re gonna melt it all down, maybe that works fine. But for finding intact, golden nuggets? Probably not. It’s not just the tool that helps you, it’s the process the tool requires. Picks and shovels facilitate a certain kind of process. Dynamite another. Code generation is an incredible tool, but it comes with a process too. Does that process help or hurt you achieve your goals? It’s important to be cognizant of the trade-offs we make as we choose tools and their corresponding processes for working because it’s trade-offs all the way down. Reply via: Email · Mastodon · Bluesky

programming Programming writing Writing
0 views
Jim Nielsen 2 months ago

Re: People Are Not Friction

Dave Rupert puts words to the feeling in the air: the unspoken promise of AI is that you can automate away all the tasks and people who stand in your way. Sometimes I feel like there’s a palpable tension in the air as if we’re waiting to see whether AI will replace designers or engineers first. Designers empowered by AI might feel those pesky nay-saying, opinionated engineers aren’t needed anymore. Engineers empowered with AI might feel like AI creates designs that are good enough for most situations. Backend engineers feel like frontend engineering is a solved problem. Frontend engineers know scaffolding a CRUD app or an entire backend API is simple fodder for the agent. Meanwhile, management cackles in their leather chairs saying “Let them fight…” It reminds me of something Paul Ford said : The most brutal fact of life is that the discipline you love and care for is utterly irrelevant without the other disciplines that you tend to despise. Ah yes, that age-old mindset where you believe your discipline is the only one that really matters. Paradoxically, the promise of AI to every discipline is that it will help bypass the tedious-but-barely-necessary tasks (and people) of the other pesky disciplines. AI whispers in our ears: “everyone else’s job is easy except yours” . But people matter. They always have. Interacting with each other is the whole point! I look forward to a future where, hopefully, decision makers realize: “Shit! The best products come from teams of people across various disciplines who know how to work with each other, instead of trying to obviate each other.” Reply via: Email · Mastodon · Bluesky

ai AI design Design
0 views
Jim Nielsen 2 months ago

You Might Debate It — If You Could See It

Imagine I’m the design leader at your org and I present the following guidelines I want us to adopt as a team for doing design work: How do you think that conversation would go? I can easily imagine a spirited debate where some folks disagree with any or all of my points, arguing that they should be struck as guidelines from our collective ethos of craft. Perhaps some are boring, or too opinionated, or too reliant on trends. There are lots of valid, defensible reasons. I can easily see this discussion being an exercise in frustration, where we debate for hours and get nowhere — “I suppose we can all agree to disagree”. And yet — thanks to a link to Codex’s front-end tool guidelines in Simon Willison’s article about how coding agents work — I see that these are exactly the kind of guidelines that are tucked away inside an LLM that’s generating output for many teams. It’s like a Trojan Horse of craft: guidelines you might never agree to explicitly are guiding LLM outputs, which means you are agreeing to them implicitly. It’s a good reminder about the opacity of the instructions baked in to generative tools. We would debate an open set of guidelines for hours, but if there’re opaquely baked in to a tool without our knowledge does anybody even care? When you offload your thinking, you might be on-loading someone else’s you’d never agree to — personally or collectively. Reply via: Email · Mastodon · Bluesky Typography: Use expressive, purposeful fonts and avoid default stacks (Inter, Roboto, Arial, system). Motion: Use a few meaningful animations (page-load, staggered reveals) instead of generic micro-motions. Background: Don't rely on flat, single-color backgrounds; use gradients, shapes, or subtle patterns to build atmosphere. Overall: Avoid boilerplate layouts and interchangeable UI patterns. Vary themes, type families, and visual languages.

ai AI design Design
1 views
Jim Nielsen 2 months ago

Food, Software, and Trade-offs

Greg Knauss has my attention with a food analogy in his article “Lose Myself” : A Ding Dong from a factory is not the same thing as a gâteau au chocolat et crème chantilly from a baker which is not the same thing as cramming chunks of chocolate and scoops of whipped cream directly into your mouth [...] The level of care, of personalization, of intimacy — both given and taken — changes its nature. I love food and analogies, so let’s continue down that path. Take these three items for example: Which one of these is the best? I’m sure an immediate reaction comes to mind. But wait, what do we mean by “best”? Best in terms of convenience? Best in terms of flavor? Best in terms of healthiness? Best in terms of how ingredients were sourced and processed? Best in terms of price? Best in terms of… It’s all trade-offs. I don’t think we talk about trade-offs enough, but they’re there. Always there. We might not know what they are yet if we’re on the frontier, but we’re always trading one thing for another. “McDonald’s cherry pie is the best cherry pie ever.” That’s a hot take for social media. We wouldn’t accept that as a rational statement applicable to everyone everywhere all the time. People have preferences, products have strengths and weaknesses, that’s the name of the game. “All software in a year will be written by robots.” Also a hot take, not a serious statement. It’s impossible to apply such a generic prediction to everything everywhere all of the time. But also: “software” hand-written by humans is not the same as “software” generated by a machine. To presume the two are equivalent is a mistake. There are trade-offs. Everything has trade-offs, a set of attributes optimized and balanced towards a particular outcome. You get X, but you lose Y. Life is full of trade-offs. Anyone who says otherwise is trying to sell you something. Reply via: Email · Mastodon · Bluesky A McDonald’s cherry pie A Marie Calendar’s cherry pie A homemade Jim Nielsen cherry pie

programming Programming business Business
0 views
Jim Nielsen 2 months ago

w0rdz aRe 1mpoRtAnt

The other day I was looking at the team billing section of an AI product. They had a widget labeled “Usage leaderboard”. For whatever reason, that phrase at that moment made me pause and reflect — and led me here to this post. It’s an interesting label. You could argue the widget doesn’t even need a label. You can look at it and understood at a glance: “This is a list of people sorted by their AI usage, greatest to least.” But it has that label. It could have a different label. Imagine, for a moment, different names for this widget — each one conjuring different meanings for its purpose and use: Usage leaderboard implies more usage is better. Who doesn’t want to be at or near the top of a leaderboard at work? If you’re not on the leaderboard, what’s that mean for your standing in the company? You better get to work! Calling it a leaderboard imbues the idea of usage with meaning — more is better! All of that accomplished solely via a name. Usage dashboard seems more neutral. It’s not implying that usage is good or bad. It just is , and this is where you can track it. Usage wall of shame sounds terrible! Who wants to be on the wall of shame? That would incentivize people to not have lots of usage. Again, all through the name of the thing! It’s worth noting that individuals and companies are incentivized to choose words designed to shape our thinking and behavior in their interest. The company who makes the widget from my example is incentivized to call this a “Usage leaderboard” because more usage by us means more $$$ for them. I’m not saying that is why they chose that name. There may not be any malicious or greedy intent behind the naming. Jim’s law is a variation on Hanlon’s razor : Don’t attribute to intent that which can be explained by thoughtlessness. I do find it fascinating how little thought we often give to the words we use when they can have a such a profound impact on shaping our own psychology, perception, and behavior. I mean, how many “word experts” are on your internal teams? Personally, I know I could do better at choosing my words more thoughtfully. Reply via: Email · Mastodon · Bluesky “Usage leaderboard” “Usage dashboard” “Usage wall of shame”

business Business design Design
0 views
Jim Nielsen 3 months ago

Book Notes: “Blood In The Machine” by Brian Merchant

For my future self, these are a few of my notes from this book . A take from one historian on the Luddite movement: If workmen disliked certain machines, it was because of the use that they were being put, not because they were machines or because they were new Can’t help but think of AI. I don’t worry about AI becoming AGI and subjugating humanity. I worry that it’s put to use consolidating power and wealth into the hands of a few at the expense of many. The Luddites smashed things: to destroy, specifically, ‘machinery hurtful to commonality’ — machinery that tore at the social fabric, unduly benefitting a singly party at the expense of the rest of the community. Those who deploy automation can use it to erode the leverage and earning power of others, to capture for themselves the former earnings of a worker. It’s no wonder CEOs are all about their employees using AI: it gives them the leverage. Respect for the natural rights of humans has been displaced in favor of the unnatural rights of property. Richard Arkwright was an entrepreneur in England. His “innovation” wasn’t the technology for spinning yarn he invented (“pieced together from the inventions of others” would be a better wording), but rather the system of modern factory work he created for putting his machines to work. Arkwright’s “main difficulty”, according to early business theorist Andrew Ure, did not “lie so much in the invention of a proper mechanism for drawing out and twisting cotton into a continuous thread, as in […] training human beings to renounce their desultory habits of work and to identify themselves with the unvarying regularity of the complex automaton.” This was his legacy […] for all his innovation, the secret sauce in his groundbreaking success was labor exploitation. Not much has changed (which is kind of the point of the book). The model for success is: As the author says: [Impose discipline and rigidity on workers, and adapt] them to the rhythms of the machine and the dictates of capital — not the other way around. Reply via: Email · Mastodon · Bluesky Look at the technologies of the day Recognize what works and could turn a profit Steal the ideas and put them into action with unmatched aggression and shamelessness

ai AI books Books business Business
0 views
Jim Nielsen 3 months ago

Computers and the Internet: A Two-Edged Sword

Dave Rupert articulated something in “Priority of idle hands” that’s been growing in my subconscious for years: I had a small, intrusive realization the other day that computers and the internet are probably bad for me […] This is hard to accept because a lot of my work, hobbies, education, entertainment, news, communities, and curiosities are all on the internet. I love the internet, it’s a big part of who I am today Hard same. I love computers and the internet. Always have. I feel lucky to have grown up in the late 90’s / early 00’s where I was exposed to the fascination, excitement, and imagination of PCs, the internet, and then “mobile”. What a time to make websites! Simultaneously, I’ve seen how computers and the internet are a two-edged sword for me: I’ve cut out many great opportunities with them, but I’ve also cut myself a lot (and continue to). Per Dave’s comments, I have this feeling somewhere inside of me that the internet and computers don’t necessarily align in support my own, personal perspective of what a life well lived is for me . My excitement and draw to them also often leave me with a feeling of “I took that too far.” I still haven’t figured out a completely healthy balance (but I’m also doing ok). Dave comes up with a priority of constituencies to deal with his own realization. I like his. Might steal it. But I also think I need to adapt it, make it my own — but I don’t know what that looks like yet. To be honest, I don't think I was ready to confront any of this but reading Dave’s blog forced it out of my subconscious and into the open, so now I gotta deal. Thanks Dave. Reply via: Email · Mastodon · Bluesky

web-development Web Development career Career programming Programming
0 views