How an AI writing fiction differs from humans: "AI pays closer attention to physical environment and characters’ inner mental states, showing emotion through bodies and environments rather than naming feelings directly." :o arxiv.org/pdf/2604.03136

Museum of teletext art :o yle.fi/aihe/tekstit...

Shoutout to the very senior professor in HCI who gave a talk last month about how AI can do qualitative data analysis better than people now.

Anthropic revenue growth. :o

'Files used to be objects you could manipulate. Nowadays apps take care of versioning, and sharing, and often organising.' ... 'You can’t drag a Google Doc; the file isn’t a directly manipulable “file” so much as the visual depiction of a save point.' interconnected.org/home/2021/02...

Just a normal day sharing feedback from one bot to another bot so that the bot could build something that operates one more bot. Then I take output of that operational bot to help the builder bot improve what it is building.

Today I learned about numeric separators in Typescript. 🤯 mariusschulz.com/blog/numeric...

LLM:s allow everyone to tap into their creativity: compose poems or draw pictures. Hopefully this leads to more people thinking about art and creativity: what makes a piece of art interesting and meaningful? When creating something new, how do I justify it to myself? Why would someone care?

'"When everyone is super, no one will be.” That’s the point that the supervillain Syndrome makes in the Pixar classic The Incredibles (2004).' niklaselmqvist.medium.com/when-everyon...

Figma MCP gained support for sending live-coded designs to Figma canvas. But based on the description, it doesn't really understand the Figma canvas or let you edit anything there, but is more like an integrated web page export functionality, right? developers.figma.com/docs/figma-m...

We have reached a moment where instead of releasing software you simply release the detailed spec for software and tell people to prompt their agent to build it themselves From the README of OpenAI’s new Symphony orchestrator: github.com/openai/symph...

Now might be a good time to buy Adobe stocks. 🤔 Sure, Canva, Figma and AI will eat its user base, but, well. 🤷‍♂️ www.macrotrends.net/stocks/chart...

Social network for AI agents. :o

Often the problem is not finding an app for something. But evaluating, comparing and testing multiple apps. This will get worse with LLM:s speeding up development. So I wonder: do we need LLM:s that are really good at comparing and selecting the best tools for a job?

Has anybody tried out building an app with Convex? Looks like a modern-day Firebase. convex.dev

Would you like to work with me? 👋🏻 I’m currently helping to strengthen the existing design systems team at Kesko and we’re looking for a nice & talented designer/developer to join us. arielsalminen.com/2026/hiring-...

Is there an AI secretary, that would make sure I don't do any obvious mistakes? For example: - tell me if I added an event in the wrong Google calendar accidentally - tell me about actually important emails I've missed - double-check when I'm about to send a message that is unclear or too complex

Nuclear weapons are now ESG-compliant sustainable and ethical investments. onestopesg.com/esg-news/nuc...

Ah, I just hate it when people use lousy ChatGPT for their marketing texts. Don't do this 😬😬😬😬 "He's not a pianist — he’s a current.", "This is not a gig. This is a testimony.", "Folk as circuitry. Synths as soul memory.", "They don’t play concerts. They detonate them."

I wonder if there are industries that use LLM AI:s a lot but don’t advertise it to their clients. And getting good profit from it. Lawyers and patent applications, for example. 🤔

"There is a continuum from play to games, marked by an increase in complexity of social construction and rules, as well as a decrease in the need for a phenomenological experience of playfulness." direct.mit.edu/books/oa-mon...

Technology and market capitalism are trend amplifiers. Suddenly we have hundreds of delivery drivers, or AI bots, or military drones. Instead of solving problems like social isolation, extremism and environmental problems, they are making them worse.

Game Design is Simple, Actually: "fun is basically about making progress on prediction" and the nature of good problems www.raphkoster.com/2025/11/03/g... by the enlightening @raphkoster.bsky.social

"because we insisted that no longer would politics be something that is done to us, now it is something that we do…" - Zohran Mamdani 11/4/25 excerpt from "Ing" comic from 2020: "Democracy is not something we have – it's something we must always be in the act of doing" spinweaveandcut.com/ing-2020/

Are there any good diagrams or models for visualising nested problems and solutions to them in planning, design, strategy etc? I could have asked the AI overlords or Google about this, but today I don't feel like that, so I thought that perhaps a fellow human would know something about this.

Before AI/LLM:s people asked questions on Twitter for things that they were pondering. I assume people do less of that now, what do you think?

To say this is a problem is a gigantic understatement. We live in the golden age of AI slop. Hopefully we will soon gain the right control the algorithms that control our online life in order to filter out AI slop to a large degree.

Price of both gold (yellow) and stocks (sp500 in red) going up steeply. Perhaps both an AI and a gold bubble? 🤔 Markets are strange, and so is the economy in general. www.longtermtrends.net/stocks-vs-go...

That's a masterclass on how to write a blog post. → nan.fyi/database by @nandafyi

:o

"The only safe way to run LLM agents is to run them on someone else's computer." This is actually quite interesting, since to make this practical, one should be able to generate disposable credit cards, email accounts, etc for LLM:s as well. simonwillison.net/2025/Oct/22/...

Neat ideas on visual programming languages and LLM:s by @szymonkaliski.com szymonkaliski.com/notes/vpls-a...

Neat semantic zoom UI:s github.com/prathyvsh/se... (via @szymonkaliski.com )

Interesting how commoditization makes general-purpose building blocks better. But at the same time, incentivises everyone to use those instead of specialised parts, since they are cheap, readily available and familiar. I'm sure there is a parallel to this in biological evolution.

I'm using a usb-c/headphone adapter for my wired headphones. Today I learned that a weird fade-in/fade-out I get in audio books is most likely caused by the cheap adapter. Sigh. Not something I expected.

Realised something almost obvious today: when a domain has matured, it is much more tempting to use ready-made components, since they are cheap and readily available, even if they are not perfect. But in new domains you may get to build ideal solutions, since there is nothing ready-made to use.

My statement on Ai from the mini-comic as syllabus i made for new class I'm teaching that starts tomorrow! It robs you of your decisions & struggles - and the joy of being surprised. We won’t to be robbed of our learning - this is essential. This & the full mini at post: bsky.app/profile/nsou...

I wonder how many people die globally due to the current right-wing politics in the U.S.? Gaza, vaccination and climate change policies, gun laws, and also increase in military spending all over the world due to escalation (and away from healthcare, welfare, etc). Must be millons.

Very glad to see that someone is doing the important work of aligning AI alignment. alignmentalignment.ai

Software developer salaries in Europe, interesting. www.levels.fyi/heatmap/euro...

I love the full-page video background in the portfolio of Alex Rott. This is also the way I interact with many of my clients: via Zoom, Hangouts, etc. I think I could copy the idea for my portfolio when I do a redesign. :) alexrott.cargo.site

So you can't make a vertical line fill container (vertically) in Figma, right? Or am I doing something wrong?

Despite a decent macroeconomic environment, only 25% of US Americans agree that people like them have a good chance of improving their standard of life. Source: buff.ly/nb67FVx

"Let’s link up and make some LinkedIn magic happen." Sigh. Time for cottagecore.

Potentially interesting: "A modern fully open-source textbook exploring why and how deep neural networks learn compact and information-dense representations of high-dimensional real-world data." ma-lab-berkeley.github.io/deep-represe...

Areal, an improved version of the Arial typeface. 😮😮😮 www.are.na/editorial/in...

I love this! Notable European coastal cities (cry out if your gem of a city was omitted) displayed by their population density.

AI:s don't just make you dumber — I think I could have gotten better in engineering maths if I had been able to use something like ChatGPT back in the day. GPT is good at giving simple explanations and translating formulas to intuitive concepts.

Gorton. aresluna.org/the-hardest-...

Reminder that last year I made an iPhone app that always points at the centre of the Milky Way It peaked at 87 in the US App Store free apps chart (Travel category) haha Here’s what Ars Technica said about it arstechnica.com/gadgets/2024...