August 20, 2026 · 8 min read
Moderna's personalized cancer therapy points toward a stranger form of generative AI: patient data goes in, a one-person medicine comes out. The drug is becoming the output of a computational system.
August 16, 2026 · 6 min read
Why calling something a failure mode turns an individual mistake into evidence about the system, and why that can reveal design or hide responsibility.
August 14, 2026 · 6 min read
TSMC published a tiny English vocabulary for technicians. It looks like a word list until it starts to reveal the factory, the interface, and an older agreement between humans and machines.
August 5, 2026 · 11 min read
AI can give us more knowledge, more options, and more reasons to wait. What happens when better advice raises the threshold for ordinary action?
August 11, 2026 · 8 min read
I asked AI to review a Finnish cultural website. The useful part was not the score, but discovering the invisible idea of good design behind it.
August 3, 2026 · 8 min read
Patrick Collison's Book of Kells experiment shows the difference between recovering a visual style and rebuilding the cultural systems that give beauty meaning.
August 1, 2026 · 11 min read
OpenAI says Astra has produced ten advances in mathematics and theoretical computer science. The mathematicians around its earlier breakthrough are already showing us what the news means for time, attention, authorship, and mathematical culture.
August 1, 2026 · 9 min read
I can tell ChatGPT and Claude how I want them to write hundreds of times. The correction feels like teaching, but the product keeps restoring a stronger voice underneath mine.
August 1, 2026 · 10 min read
AI makes the first version of a website almost effortless. The harder work begins later, when every reasonable edit risks making the product forget why it looks and behaves the way it does.
August 1, 2026 · 6 min read
Every change I make to an AI draft teaches the system something about what I mean by quality. The published article shows what survived, but the edit history may reveal more about my voice than the final text alone.
July 31, 2026 · 9 min read
Retailers are entering AI assistants to preserve their taste, but every successful integration teaches shoppers to begin somewhere other than the store.
July 31, 2026 · 10 min read
AI can generate answers instantly. The harder skill is staying inside uncertainty long enough to discover what is actually true.
July 31, 2026 · 11 min read
Vibe Jam, ETHGlobal, Kaggle, Topcoder, InnoCentive, HeroX and XPRIZE all turn uncertain problems into competitions where the crowd absorbs much of the risk of experimentation.
July 31, 2026 · 9 min read
A therapy session can happen and help a patient, yet remain financially unreal until it has been translated into the codes, evidence and permissions an insurer accepts.
July 30, 2026 · 4 min read
The AI race looks like a competition for more GPUs, but some of the fastest machines in history are still limited by memory, coordination and waiting.
July 30, 2026 · 8 min read
A modern device is a coordinated organization of CPUs, GPUs, neural processors, signal processors and tiny controllers. Software design increasingly means deciding which machine receives each part of the work.
July 29, 2026 · 11 min read
Interaction design has strong tools for screens, flows and journeys. Interaction sequence maps can expose the hidden choreography of users, interfaces, sensors, services, AI and human recovery.
July 29, 2026 · 8 min read
Brainrot is an attention format. AI slop is a production system. Generative AI did not invent our appetite for repetitive nonsense, but it industrialized the supply and gave recommendation algorithms an infinite field of cultural experiments.
July 29, 2026 · 5 min read
In comics, visual novels, memes, explainers, and serialized AI media, the image can be imperfect because its real job is to carry us toward the next panel.
July 29, 2026 · 8 min read
A screen shows the product for a moment. A URL preserves its objects, state, position and history, making the experience shareable, revisitable and legible to people, browsers and agents.
July 29, 2026 · 10 min read
The hidden job market includes work companies have not yet named. Missing would analyse organisations, identify absent capabilities, design the role and find the person who can build it.
July 29, 2026 · 5 min read
A pillow is a small argument between filling, fabric, seams and gravity. Learning to draw one means learning to see the invisible forces beneath a visible form.
July 29, 2026 · 7 min read
AI has made execution so easy that the harder questions arrive first: who needs this, why now, what deserves to be built and which future is worth committing to.
August 14, 2026 · 12 min read
AI may be more interesting as a way to reopen old films, code, research, manuscripts and archives than as another machine for making new things.
August 3, 2026 · 11 min read
AI is not only automating work. It is beginning to automate making itself, one of the oldest ways humans experience agency, growth, belonging, and meaning.
July 29, 2026 · 5 min read
I connected an old hard drive and asked what was on it. The computer could answer almost everything except what had happened.
July 26, 2026 · 12 min read
A YouTube channel reached around 60 million views by repackaging scenes other people made. It made me wonder where online arbitrage ends and a durable company begins.
July 26, 2026 · 14 min read
What if every continuing interest had a small app that could absorb what I learn, survive my attention moving elsewhere, and occasionally reveal that one forgotten bet had become valuable?
July 26, 2026 · 5 min read
We keep trying to increase intelligence, speed, and output while overlooking the narrower resource that decides what becomes real: the human ability to keep returning to something.
July 26, 2026 · 8 min read
AI can now generate one hundred working apps faster than a single product can find one hundred paying users. The challenge reveals that software has become cheap enough to function as a draft.
July 26, 2026 · 9 min read
Claude Design can import a codebase, use a real design system, and sync with Claude Code. Yet Anthropic still treats the moment a design becomes software as a handoff.
July 25, 2026 · 8 min read
Senior consulting roles such as Transformation Lead do more than sell or deliver projects. They turn organisational uncertainty into a problem, a budget and work for everyone else.
July 25, 2026 · 9 min read
A Finnish directory of software companies becomes more interesting when read backwards: not as a list of vendors, but as a map of problems organisations were willing to pay to solve.
July 25, 2026 · 10 min read
I uploaded my DNA once. Genomelink and YourRoots show how a fixed biological file becomes an endless product of traits, ancestors, maps, hints, and unfinished identity.
July 25, 2026 · 7 min read
AI has made implementation cheap enough that code is no longer my main bottleneck. The harder limit is deciding what a product should do and which ideas deserve to exist.
July 24, 2026 · 12 min read
The interesting part of AI graphic design is not its ability to imitate Bauhaus or Brutalism. It is what happens when obscure visual languages collide.
July 21, 2026 · 9 min read
AI has made production cheap enough that the scarce act is no longer creation, but choosing what deserves to enter a product, a company, or a life.
July 21, 2026 · 8 min read
What Venezuela's citizen-built earthquake tools reveal about AI, emergency infrastructure, and software that acquires public responsibility before it has governance.
July 14, 2026 · 8 min read
Luxury museums, open AI models, and private wealth services all point to the same shift: value is moving from the thing being sold into the system surrounding it.
July 13, 2026 · 6 min read
I wanted one small drawing lesson at a time. The first response showed me the whole path instead.
July 12, 2026 · 8 min read
What Japan's one-month AI startups reveal about the difference between building a model, deploying a product, and changing an organization.
July 12, 2026 · 8 min read
What fourteen University of Washington speakers reveal about where influence, compensation, and professional value migrate as UX careers advance.
August 1, 2026 · 13 min read
Planet began in my inbox as a small satellite startup. Eleven years later, it has become part of how forests, borders, wars, disasters, and physical change are remembered.
December 3, 2021 · 4 min read
An opinionated archive note on why the old data-visualization habit of making complexity visible should be treated as essential AI design, not a decorative afterthought.
November 3, 2017 · 3 min read
A first-person field note from AI Frontiers 2017 in Santa Clara, before applied deep learning had fully settled into everyday product interfaces.