Design: Weekly Summary (November 10-16, 2025)

I would describe this week's design chatter as a sort of tidy mess — lots of careful craft, a few sharp rants, and a steady hum about tools that either save time or make trouble. To me, it feels like people are wrestling with two things at once: making things that look and feel right, and making systems that actually survive day-to-day life. You see it in type and paint, in icons and stations, in tiny apps and ambitious AI agents. It's like watching someone refurbish an old house while also deciding which rooms to knock through.

Material things: type, paint, and the journal that feels like a pocket

Typography and physical objects popped up a surprising number of times. There's a tactile hunger in these pieces. I’d say they read like people wanting things to have texture again, not just pixels that behave like paper.

Take the short, obsessive note on typography so1o.xyz. The piece is frank about the fetish part — not in a weird way, but in that small, nagging devotion you get when you can’t stop fiddling with a logotype at 2 a.m. The author spends time on alignment, weight, and the tiny choices that separate a decent poster from one that sings. To me, it feels like someone describing the last 10 percent of a recipe: all the seasoning and the pan flicks that make a dish memorable. They keep nudging readers to start making, even if it's just a single poster. There’s a little impatience with big ideas that never leave the sketchbook — which, frankly, is refreshing.

Then there’s the Design Museum notice about Harland Miller, flagged by Ian Mansfield. Giant letter paintings going up in London, big canvases that mix medieval manuscript vibes with pop aesthetics. That felt like the other side of the typography coin — scale and spectacle. Where the typography fetish piece is the quiet workshop, Miller’s show is the public square. I’d say it’s the same love of letters, but shouted through speakers in a market rather than whispered over a drafting table.

And the journal project from Suleika Jaouad — The Alchemy Journal — slots neatly into this conversation. The object is part reporter’s pad, part sketchbook, part diary. You can almost feel the warm corners from which the idea came: summers in Tunisia, slow mornings. The post reads like someone handing you a small, sensible tool and saying, try this and see what sticks. It’s practical but earnest. I’d say it’s the kind of thing you buy and then defend in public, like a favourite umbrella.

These three posts point to a recurring idea: design isn't just visual cleverness. It’s the insistence on execution, on making an object that carries use and memory. One is small and meticulous, one is large and theatrical, and one is intimate and useful. They all care about letters and surfaces, but in different scales. If you like type, you’ll want to click through.

Tools and the indie maker instinct

There’s a neat cluster of posts about tooling, apps, and making stuff happen fast. They’re mostly pro-hands-on: build, ship, test, fix.

Creativerly writes about ShareBridgely, a niche macOS app that sits between a browser bookmarklet and the native share sheet. The post is full of small-person energy — someone solving a tiny personal pain and then deciding to shape it into a utility. The arc is familiar: you’ve got a reading habit, you juggle tools like Readwise and Reeder, and then you make a small app because the middle bit is awkward. That story always reads like a good recipe: start with what annoys you, make something, iterate.

On a bigger scale, Mike Coats gives a practical walkthrough of using PCBWay for prototyping. There are details — Inkscape, KiCAD, UV printing, stencils — and a lot of hands-on assembly talk. It’s the sort of post that feels like sitting beside a friend at a bench while they teach you how to solder. He praises the customer service and the quality, but not without the minor gripes you always get with manufacturing. That honesty matters. It’s not marketing; it’s a how-to that helps you imagine your own prototype rolling off a bench.

Then there’s the design-and-build sprint from Not So Common Thoughts: From Sketch to Site in Hours. This one is refreshingly blunt about process. Analysis, experiments, iterative changes, a critique with another designer that actually moved the needle — that sort of back-and-forth you wish more agencies wrote about. The piece also nods to AI as a speed lever but not a magic wand. I’d say it feels like a studio session where someone turned a messy whiteboard into something you can click through before lunch.

What ties these posts together is a bias toward making — towards the kind of design that ends at a thing you can touch or use. The stories range from small utilities to physical boards. They read like people who'd rather be in the garage than at a planning meeting. If you tend to tinker, these are the posts to bookmark.

Interaction and the little things that matter

There was a neat tension between grand systems and the trivial-sounding details that actually shape everyday experience.

jerlendds writes about gestures — the slow rise of gesture-based interaction and its place in creative thinking and human-robot interaction. The post isn’t starry-eyed. It talks about models, research, and the messy real-world stuff you need to get right for gestures to feel intuitive. To me, gestures are like learning to kick-start a scooter — awkward at first, then oddly natural, and then you miss it if it’s gone. The piece shows the promise but also reminds you of the calibration cost.

That human-scale interaction debate pops up again in a grumpy but important critique from Michael J. Tsai about macOS Tahoe’s new icons. He and other contributors feel the icons have lost detail and personality. There’s talk about blandness, accessibility issues for low-vision users, and a general sense that something human was smoothed over into insignificance. I would describe this as a small outrage, but a meaningful one: when icons look the same, the thing loses a bit of its soul. It’s like when all the corner shops in a street get replaced by the same chain; you still have what you need, but the character is gone.

These posts together remind you that interaction design isn’t only about cool new gestures or shiny screens. It’s also about the quiet cues — contrast, shape, micro-detail — that help you find your way without thinking. Losing those is quietly bad.

Maintenance, governance, and the slow death of features

A louder theme this week is governance — not government, but the housekeeping of design.

The Product Picnic makes a solid stab at this with a post arguing that without UX governance, apps turn into sludge. The term 'sludge' lands. They point to the pandemic-era sign that never gets removed as a clear example: temporary features outliving their usefulness because no one is accountable for pruning. The piece wants designers to reclaim strategy and the power to remove things, not just ship new features.

This complaint feels very practical. I’d say design without maintenance looks like a garden left to its own devices: the flower beds become jungle, and you end up tripping over last year’s trends. There’s a call here for processes that actively retire features, and for designers to hold responsibility beyond the sprint.

That connects neatly with the essay from Not So Common Thoughts about the One Thing Constraint. It uses science fiction world-building as a metaphor: a single, strong change can make a world coherent. Translate that to product design and it’s about not piling on changes until the product loses internal logic. One major shift, done well, beats a hundred small shifts that don't talk to each other. The two pieces together argue for coherence and upkeep, not just growth.

If you work on products, both posts are worth a read. They don’t offer a silver bullet, but they do remind you that curation is a design skill.

Design meets code: the uneasy, productive middle ground

A couple of posts sit squarely at the intersection of design and code. They felt like an ongoing negotiation — designers wanting agency, and engineers wanting predictability.

In the episode with Cursor’s head of design, Ryo Lu, covered by Peter Yang, the message is that designers should code. The discussion shows a concrete example: Ryo built a retro-style operating system, ryOS, using AI and a hands-on approach. What’s neat is the practical stuff — how to manage multiple AI agents, how to avoid terrible AI design patterns, and how a team can operate without a full-time project manager. There’s a lot of humility here. I’d say the conversation feels like an experienced chef explaining how to both run the pass and cook the sauce: messy, exacting, rewarding.

And then there’s the more theoretical, heavyweight paper from Grigory Sapunov: Huxley-Gödel Machine. It’s a mouthful, but the core idea is important. The paper argues that we measure AI agents with the wrong yardstick. Short-term performance misses long-term improvement potential. So they introduce a lineage-based metric, Clade-Metaproductivity, which tries to capture which agents will improve over time. The result is a different kind of agent design that uses fewer CPU cycles and reaches human-level coding performance on benchmarks.

Put these two together and you’ve got a neat paradox. On one hand, designers should learn to code and use AI as a practical tool. On the other, the AI systems themselves need a different kind of evaluation to become reliable collaborators. It’s like training an apprentice: do you reward how fast they finish the first task, or how they improve across ten projects? The paper says the latter matters.

This fold — where craftsmanship meets machine learning — felt like the most future-facing thread. It’s both exciting and a little unnerving.

Heritage, public design, and why stations matter

A quieter note this week was about preserving design that is already in the world. Ian Mansfield reported that Southwark tube station got Grade II listed status. The station’s late-20th-century circular entrance hall and underground layout were singled out. There’s something about this that hits a civic chord for me. Transport design isn’t just signage and tiles; it shapes how a city feels.

Listing a station is a nod to the idea that not all heritage is old stone. Some of it is stainless steel and concrete from the 1990s. It’s like putting a little plaque on your favourite pub to say, yes, this place carries memory too. The piece points out that it’s still a working hub, not a museum. That dual use — public value plus practical function — is worth thinking about when you design public stuff.

The small experiments that tell bigger stories

A few posts were very specific but still quietly instructive. The PCBWay piece I mentioned earlier sits here. So does the Reeder/Readwise thread around ShareBridgely. Small interventions, tiny experiments, tools for personal workflows — these all show how design plays out at personal scale.

The curiosity in these posts is practical: what tool helps me read better, or ship a board, or make a poster? The answers are often imperfect, messy, and subject to iteration. Which is fine. They read like real life.

Where people agreed, and where they grumbled

Agreements were easy to spot. Craft matters. Execution matters more than idea alone. Designers should have a say in how products evolve. Small details — icons, gestures, typography — change how people feel and use things. Tools that shave time off real tasks get applause. Folks also agreed, quietly, that AI is a tool worth learning but needs sensible guardrails.

The grumbles felt a bit louder. There’s fatigue over watered-down aesthetics (hello Tahoe icons). There’s frustration with products that accumulate junk because no one prunes. And there’s impatience with hand-wavy AI design that promises a lot but delivers unevenly.

Those tensions are what kept the week interesting. You could almost map the debate as craft vs scale: make things lovely vs make systems that scale without turning to sludge. It’s not a simple trade-off, and most writers here wanted both, awkwardly and stubbornly.

If one thread runs through all the posts, it’s this: design is not only about newness. It’s about stewardship — of objects, interfaces, experiences, and expectations. You can push for the shiny new gesture or the clever AI, but you also have to sweep the path ahead of you and clear the old signs.

There’s a lot more detail in each piece, and if a particular topic nudged you — the Huxley-Gödel Machine logic, the pcb prototyping workflow, or the Harland Miller paintings — give the authors a click. They wrote the bits that matter much better and with more nitty-gritty than I could squeeze into a round-up. The posts are small pockets of usefulness, and I’d say they’re worth a Saturday afternoon with a mug and maybe some headphones.

Anyway, that’s the week. Design felt both stubborn and curious. People are still making stuff with real hands, and they’re still trying to figure out how to keep it from getting messy. Sometimes you want a tidy shelf, sometimes a rough workbench — and a few of these posts will help you decide what to keep and what to toss.