Follow buttons here act as you, not as @admin.
| Label | Kind | State | Action |
|---|---|---|---|
| @ricardo | local | — | |
| @rmdes | local | — | |
| @paul | local | — | |
| news.rss.chat | source | — | — |
| scripting.com | source | — | — |
| www.techdirt.com | source | — | — |
| simonwillison.net | source | — | — |
| www.manton.org | source | — | — |
| feeds.kottke.org | source | — | — |
| rss.beehiiv.com | source | — | — |
| www.404media.co | source | — | — |
| brilliantmaps.com | source | — | — |
| shellsharks.com | source | — | — |
| www.platformer.news | source | — | — |
| podstandards.org | source | — | — |
| zacharykai.net | source | — | — |
On the home AI front, I decided to get the Beelink EQi. Transferring the brain of my OpenClaw to it today. I wish this could be a Mac mini, but the cost of Apple storage is ridiculous. Much easier to drop a 4 TB SSD in this.
Neat example from Simon Willison of using voice to develop a feature for his blog while cooking, talking and watching the preview on his laptop:
We went on like this for about half an hour (the time it took to cook dinner). The model would reply and occasionally ask clarifying questions, then get to work modifying the code.
As much as everything has been turned upside down, the flow is the same: make a change, see how it looks, make another change.
Bill McKibben: I’m Exhausted. You’re Demoralized. We Weren’t Built for a World Like This. The human nervous system wasn’t designed for the polycrisis. “Among all the other dangers these developments represent, they’re simply wearing us out.”
I keep thinking about this story I read on NPR yesterday about a chance encounter and a stranger's kindness. 一期一会. "One time, one meeting." We're all just passing briefly in and out of each other's lives.
I shipped a new feature for my blog today: the Newsletters page, which offers an index of all of the newsletters I've sent out, both my free weekly Substack and my monthly sponsors-only updates. I built the feature almost entirely using my voice, chatting away to my laptop while I cooked dinner.
I used the ChatGPT desktop app for this, in the Codex tab, using the voice conversation mode, running against a local development environment. Here's what that looks like:

I started the session against my local simonwillisonblog checkout by typing:
Start dev server and open in browser
This gave me a preview of the site that it would be working on, and meant that I could later ask it to show me the new pages so I could visually track its progress.
Then I clicked the "Start new voice chat" button - that's not the microphone button, it's the one to the right of it - and set my laptop up in the kitchen so I could talk to it while I cooked.
I had a pretty good idea of what I wanted to build, and it's a simple enough Django feature that I was certain the model (in this case GPT-6 Astra High) would be able to do it. A new model, a migration, some view code, templates, and a couple of import functions to populate the database from external sources.
Here's an extract of my voice transcript that was captured by Codex:
Um, they do not. Um, this is going to be a new type of content. Um, it's not going to show up... Oh, hold on. Yeah, no- I do not want this to show up in my, um, tag pages and date archive pages and... Actually, no, I think... I don't want it on the tag pages. I don't want it on the, um, blog index page. But I think I do want it to show up on the date-based pages. You know, if you navigate to September the 19th, and I sent a newsletter on that page, I think I want that to show up. So... this is- so I think we probably need a new model. The other thing is that I want them searchable, uh the Substack ones are not searchable, because those are actually just copies of other s- on content on my blog. These monthly ones do contain unique content, and spe- and once they're... published, like once they're made public a month after they've gone out, I want them to show up on my search results.
Apparently this was clear enough that the model knew what I wanted to build! You can read the full transcript, disfluencies and all, in this Gist.
We went on like this for about half an hour (the time it took to cook dinner). The model would reply and occasionally ask clarifying questions, then get to work modifying the code.
We got a surprisingly long way entirely by voice:
/api/v1/archive directly) and then ran a search to figure out how to paginate it and found this article by Karen SpinnerIt was almost ready to ship. The catch was the imports: Astra offered to export data from my local copy so I could import that into production, but I wanted it to work like my other import scripts. Since some of the data lived in a private GitHub repository, this would involve creating a new API key, and for that I knew I'd have to sit at the keyboard for a while.
Once I had finished cooking and judged it mostly feature-complete, I had Codex create a branch and open a pull request.
I reviewed the code in the GitHub PR interface. It was nearly what I needed, except it had chosen to use Git in a subprocess for one of the import scripts. I needed one of the imports to pull from a private Git repository, so I figured the API would be a better bet. I switched to typing and had Codex swap that out for an API-based import instead.
You can see the changes I made during the review in the extra commits on the PR. I fixed the import mechanism and made a few tweaks to the display of those public pages. It took an additional half hour of typing-based prompting to get to the point where I was happy to deploy it to production by landing the PR.
You can see the end result at the new newsletters index page, or view the page for a previous monthly newsletter.

The index page shows my most recent Substack weekly newsletters and GitHub sponsors monthly newsletters mixed together in reverse chronological order. Further down the page are links to my by-year archive pages.
GPT-6 Astra designed the page, and then tweaked that design based on my vocal feedback from glancing at the local preview across the kitchen.
OpenAI love using voice-driven demos like this one for things like DevDay - and they do work well in that environment. I don't think this is going to be a daily driver for me though.
I've written before about how much "work" I get done using ChatGPT voice mode on my phone while walking the dog - mostly research and brainstorming, but occasionally actual development work by having ChatGPT write and test out snippets of code.
This feels different. The addition of the visual preview, plus being able to type or paste things in via the keyboard when I need to communicate something that doesn't work vocally, makes this a much more powerful way of interacting with a coding agent.
I still switch back to typing once I get down to the details of things though. Being able to paste in examples and error messages, or directly highlight the code or feature that needs changing, remains more efficient than trying to describe it in words.
I mainly work from home, which is good because there's no way I'd want to talk to my computer like this in a shared workspace!
The killer feature for me is the ability to multi-task. I usually cook with a podcast or TikTok running; now I can actually build stuff instead.
You are only seeing the long-form articles from my blog. Subscribe to /atom/everything/ to get all of my posts, or take a look at my other subscription options.
The first Spurs preseason game is tonight! They've changed the streaming provider yet again, now on DAZN. 🏀
In 1996, the Blondin family drove all around the US, blogging about it all before “blog” was even a word. “We were doing something that had no name and no manual, and we were doing it over a dial-up modem while living in a motorhome.”
Recorded a short video to explain a new batch selection interface for uploaded photos, just rolled out today in Micro.blog.
Agustín Ferrando Trenchi became a little obsessed with the moon as a kid. As an adult, he’s taken tons of photos and hundreds of hours of footage of the moon. After seeing an airplane fly close to the moon while he filmed, Trenchi began a quest to capture a plane flying across the face of the moon.
Over the course of 10 years, he took photos and videos of birds and statues silhouetted by the moon, but planes eluded him.
I realized if I wanted to film this, I was going to have to try a different approach. It wasn’t just going to happen by accident. So in February of this year, taking advantage of being in London, I decided to go near one of the busiest airports in the world to maximize my odds as much as possible.
But while we were on our way, I had second thoughts. I felt like it took the fun out of it. It didn’t make sense to force it that much either. So we got off.
I’m not going to spoil the video, but the whole thing is just so lovely — a meditation on obsessions, journeys, and finding meaning in your life. (via virginia)
Tags: Agustín Ferrando Trenchi · Moon · photography · video
Abolish Frat Culture. “This is going to be uncomfortable for anyone with a kid in a fraternity to hear: If your son finds himself in a group that would send a ‘free pussy’ text, it’s too late. The group is the disease, the text is the symptom.”
Any Human Ever randomly picks an average person from some point in human history and tells you about what their life was like. Mine was a woman born in 1547: “Of all the sounds of her life, the loudest were thunder and the church bells.”
Poking around in WordPress because I wanted to see how they handle something in the UI, and I just noticed the "Create AI Podcast" sidebar menu item:
Pick a date range or a few specific posts and we’ll generate a two-host conversation, complete with narration and a ready-to-publish draft. Edit, refine, and hit Publish when you’re happy.
I want more people to podcast, but creating a podcast with multiple fake hosts (for what is likely a solo personal blog) is weird and unnecessary. This is why so many people hate AI.
Love this Apple video of embroidery for the logo to announce the October 13th event. 🧵
Mamdani’s “Richard Scarry Socialism” Can Make City Life Simple Again. “It is an astonishing indictment of our times that the cities of our picture books can seem like a radical utopia.”
Thought maybe I wouldn't bother downloading the latest Xcode 27.1 from a coffee shop, but I'm glad I tried. Actually getting 300 Mbps.
Late in his life, the prominent English mathematician and philosopher Bertrand Russell was subject to an attempt in the early 1960s by Oswald Mosley to provoke him into a debate about the merits of fascism. Before WWII, Mosley had founded the British Union of Fascists. BUF had a corps of paramilitary personnel, nicknamed “Blackshirts” for their uniform color, at its disposal and Mosley “publicly espoused antisemitism and sought alliances with Benito Mussolini and Adolf Hitler”.1 After the war and having left Britain in disgrace, he was an early Holocaust denier. From Wikipedia:
Mosley was a pioneer in the emergence of Holocaust-denial. While not denying the existence of Nazi concentration camps, he claimed that they were a necessity to hold “a considerable disaffected population”, where problems were caused by lack of supplies due to “incessant bombing” by the Allies, with bodies burned in gas chambers due to typhus outbreaks, rather than being created by the Nazis to exterminate people. He sought to discredit pictures taken in places like Buchenwald and Belsen. He also claimed that the Holocaust was to be blamed on the Jews and that Adolf Hitler knew nothing about it.
With all that as context, Russell wrote a letter to Oswald in January 1962 emphatically declining his invitation to engage in discourse about fascism:
Dear Sir Oswald,
Thank you for your letter and for your enclosures. I have given some thought to our recent correspondence. It is always difficult to decide on how to respond to people whose ethos is so alien and, in fact, repellent to one’s own. It is not that I take exception to the general points made by you but that every ounce of my energy has been devoted to an active opposition to cruel bigotry, compulsive violence, and the sadistic persecution which has characterised the philosophy and practice of fascism.
I feel obliged to say that the emotional universes we inhabit are so distinct, and in deepest ways opposed, that nothing fruitful or sincere could ever emerge from association between us.
I should like you to understand the intensity of this conviction on my part. It is not out of any attempt to be rude that I say this but because of all that I value in human experience and human achievement.
Yours sincerely,
Bertrand Russell
That’s how it’s done.
Fascists and authoritarians are often personally shitty as well. “During his marriage to Cynthia, Mosley engaged in an extended affair with his wife’s younger sister, Lady Alexandra Metcalfe, while also carrying on a separate affair with their stepmother Grace, who had been a widow since her husband’s death in 1925.” During his marriages, Mosley apparently also slept with the wives of friends, another of his first wife’s sisters, and several other women. ↩
Tags: Bertrand Russell · fascism · Oswald Mosley · politics
The tech industry kicked the web out of the web.
scripting.com is your canary in a coal mine.
If you can read it then your browser is a real web browser.
If it can’t it’s phony baloney.
That's the big picture behind the Web of Markdown, an idea that has generated surprising interest among devs, esp now that Claude Code exists to enable weekend projects that can do miraculous things.
My longtime friend Andrew Shell has quickly put together what looks like a Chrome plugin that starts work on how to display a Markdown web page. And it looks fucking beautiful, as I imagined it might. Get rid of the bullshit just tell me what you have to say. A web that you can read. A web without interstitials. A web every write lusts for as well as an editor that can be used everywhere. This is the nightmare of the companies whose products inhabite the de-webbed web of 2026 (and that's all there is). Let's go back and pretend the VCs didn't run all their get-rich-quick schemes, what would the fucking web be like if it was all about two things -- writing and reading and the tech got out of the fucking way and let the writers do their thing. I think it'll actually work this time because they really fucked up the web and we can show them dramatically what it would be like if they hadn't. I've said this before if they insist on twisting what remains of the web into more bullshit, our option is to re-create the web and forget about them, leave them in the dust.
By the way, now I'm glad I have a source:markdown element in most of my RSS 2.0 feeds. I knew they'd come in handy! :-)
Back to Andrew's project, I said in a comment, I want more Markdown and absolutely no Chrome.
I have to trust Google for far too many things, but I think we can get rid of our dependency on Chrome. Now is the best time in a long time, as there's a hybrid product forming and they don't have the best AI system, and btw -- even if they did as users we are highly motivated for them to get some competition. And the AI companies don't do HTML or CSS, they do Markdown. So much so that Google had to convert their office product to Markdown before OpenAI would work with them. The wind is blowing away from Google and into Markdown, and I'm going to insist that they support HTTP in addition to HTTPS to make sure Google hears loud and clear that they can't run the web as if it were a corporate platform.
BTW, Mets fans do not root for the Yankees, while Yankees fans are happy to wear Mets caps when we're on top. Their only philosophy is that the Yankees always win and if they don't we hate them. And the Mets' philosophy is incomprehensible (I hear) to Yankees fans. Our philosophy is to have a philosophy and there's nothing about winning in there because the love of the Mets is rooted in a team that couldn't even run the bases. They were loved because they were losers, probably because we knew their fate was to reward our belief in them. Now this year the Mets sucked. The reason why is kind of obvious. The Mets got confused, and they thought they were the Yankees. That. Does. Not. Work. And we already have one heartless low-philosophy baseball team in NYC, we sure as hell don't need another one. But this year is different because we have the fucking Knicks, who won in 5 and it would have been a clean sweep if it weren't for that buzzkill mofo, Trump. So anyway, I hear the Yankees were eliminated yesterday. Well then, it's on with the NBA season now and we'll try again with the Mets in April, or more realistically after the NBA championship for the new season is settled.
If programming languages don't matter much anymore, but reviewing code matters more than ever, can we resurrect RubyCocoa? Ruby is still one of the most elegant, readable languages. Wish I could use it everywhere.
The Global Rise of Electoral Dictatorships. Is the U.S. Next? “Electoral dictators — who identify as presidents and prime ministers — hold uncompetitive and unfair elections after aggressively suppressing political opponents.”
A short (too short?) video from Little White Lies on designer Saul Bass and his influence on the art of movie title design. They also did videos on Pablo Ferro and Kyle Cooper.

Just wanted to dip in here to say that Amazon is running another one of their sales for Prime members, which can be a great opportunity to upgrade some tech items at lower prices, especially with the price increases we’ve seen recently. Just a couple of quick things:
You can find more sale items here. I think these deals expire in a few hours, so hop to it.
And also, other stores like Best Buy often do their own deals around this time and Bookshop.org is offering free shipping. I’m reading Marlon James’ The Disappearers right now and, oof, it’s great but harrowing.
Tags: Amazon
Andrew Shell actually built a Markdown browser:
I quickly realized that much of this was already present with sites like mine that serve markdown for LLMs.
I knew about /llms.txt, but didn’t realize there were so many people using HTTP content negotiation with text/markdown.
“Margaret Hamilton, who led the 100-person team that wrote the onboard software for the Apollo missions, guiding Neil Armstrong and Buzz Aldrin to a safe landing on the moon, died Sept. 30 in Cambridge, Mass. She was 90.”
Nick Heer has a blog post about Meta charging if you want to link outside their silos:
Instagram has never supported links of any kind in individual posts, so that is a new feature with two links per month permitted for free, and more enabled by this pricey subscription — starting at $50 USD per month for four posts per month containing external links
How the turn tables... Remember when Australia and Canada wanted Facebook to pay newspapers for linking to their articles? This is some twisted karma. In both cases, paying for links is fundamentally anti-web.
Coming up with solutions and new ideas was an adventure. Dedication and commitment were a given. Mutual respect was across the board. Because software was a mystery, a black box, upper management gave us total freedom and trust. We had to find a way, and we did.
There's an amazing photo of her standing with print-outs of code for Apollo missions. Rest in peace.
Me on textlog
Evgeny Morozov has pointed out that using ai through chat interfaces is based on an individualised and privatised model, which (heavily paraphrased) is inherently anti-social. With an extra layer now of people making small personalised tools, do bigger things still need teams of people who know what they’re doing? FWIW I’d like to see more sharing of vibe coded tools instead of everyone making their own more less the same ones. Mine are here: https://pauladams.neocities.org/
Where in the world are typefaces being designed? The US, Germany, and France…but also the UK, Netherlands, and Indonesia.
Reading Dave Winer's notes on a hypothetical Markdown web browser. I don't think we need to reinvent the web or HTML, but it would be neat if web browsers had built-in Markdown preview by default. Why not? Browsers render PDF and all sorts of files.
As previously promised, here's Anthropic's new fast, low cost model: Introducing Claude Haiku 5.5.
The previous Haiku, 4.5, was very much showing its age. It came out almost a year ago, and was priced at $1/million input and $5/million output - relatively expensive even back then, and a full 10x the price of OpenAI's GPT-6 Luna, released last month.
The new Haiku exactly matches the price of GPT-6 Luna - $0.10/$0.50 - up to 100,000 tokens. Beyond 100,000 tokens the price increases 5x to $0.50/$2.50. Luna itself has a price increase at 272,000 tokens but only to $0.20/$0.75.
Haiku 5.5 also uses a new, less generous tokenizer. My Claude Token Counter tool shows that the same long prompt uses around 1.25x as many tokens with Haiku 5.5 compared to Haiku 4.5, so there's a hidden price increase there.
If your workloads fit in 100,000 tokens, Haiku is the same price as Luna and reports higher benchmark scores. Above 100,000 tokens, Luna looks like a much better deal.
The most recent release of llm-anthropic finally fixed it so I don't need to ship a new version of that plugin for every new model. I tested the new model like this:
llm install -U llm-anthropic
llm anthropic refresh
llm -m claude-haiku-5.5 "Generate an SVG of a pelican riding a bicycle" -o thinking_effort low
Here are pelicans for low, medium, high, xhigh, and max. The new Haiku doesn't let you disable reasoning, and defaults to medium. I got a good bicycle frame for everything beyond low. The low effort pelican cost 0.0936 cents and took 7 seconds.
This max effort pelican (with a reasoning trace that starts "This is the classic pelican-on-bicycle SVG test...") took 5 minutes 9 seconds to generate, but still only cost me 3.3826 cents:

(Since the reasoning trace exhibits awareness of the benchmark, here's Generate an SVG of an armadillo in fishnet tights jaywalking on Mars (on xhigh), and the same prompt against some other recent models. Background on that.)
For comparison, here's the pelican I got a year ago from Haiku 4.5 (for 0.7583 cents - Haiku 4.5 did not support reasoning levels). It sucked at drawing pelicans:
In addition to Haiku 5.5, Anthropic announced today that they are halving the price of cache reads for Sonnet 5.5. They've also added API credits to subscription plans:
Second, this week, we’ll roll out a new monthly API credit to all Max and Team subscribers for use on the Claude Platform. Max 5x users will get $100 in credits per month, Max 20x users will get $200, and Team subscribers will receive up to $500, pooled across their users.
Claiming this is pleasantly easy: navigate to Settings -> Billing and select the API organization that should benefit from the credits every month:

The API credits exactly match the cost of the subscription itself. This is really generous - it makes it much easier for subscribers to use the API. Anthropic also let you disable auto-reload for the API, with the consequence that "API requests will stop when your balance runs out" - exactly what you want if you're planning to burn through those API credits without risk of a nasty billing surprise.
Note that the monthly credits do not roll over - use them or lose them.
OpenAI still allow you to use your Codex subscription for personal API use, which works out as a better deal for heavy API users. This new credit scheme goes at least some way to overcoming that difference.
You are only seeing the long-form articles from my blog. Subscribe to /atom/everything/ to get all of my posts, or take a look at my other subscription options.
That said, surprisingly Nicolle Wallace's podcast is really good, which surprised me because her two hour afternoon MSNOW show is kind-of nothing. On this show they have a lot of time to talk with one person, and maybe there are limits on the TV show that aren't on the podcast? Another theory, I listen to her podcast, and thus can't see her facial expressions. On TV she looks utterly hopeless, I'm not kidding, and she's justified. I worry about how much of this she can take. But on the podcast she comes off as she's curious, respectful, let's the guest speak (so many hosts don't). She reminds me of a Diane Rehm, a podcast host on NPR through 2016. She never let her guests talk down to us, her audience, which I appreciated. A lot of times you hear a science, history or tech topic and hope you might learn something, but they only repeat the intro over and over, with a lot of giggling, of course. BTW, Diane Rehm has a new podcast. I bet it's great, probably gets right to the point and I guarantee you there is no giggling.
Laika is the stop-motion animation studio that made Coraline & Kubo and the Two Strings. Their newest film, Wildwood, looks pretty incredible; have a look at the trailer.
Creative Boom recently highlighted some of the process behind the film, which is heavily dependent on physical sets and puppets.
The man tasked with putting the show together was art director Matthew Maldonado, having spent five years designing and building Wildwood’s sets. So I asked him the million dollar question: how much of Wildwood is physical and how much is digital?
Matthew’s response is simple. “Every single thing was made by hand, truly,” he stresses. “And every single thing that we built was either shot in camera or referenced by our effects department. We’ve got some really wide, expansive locations and there’s some digital aspect to all of that. But if it wasn’t on camera, on stage, it’s a scanned asset or a reference asset, specifically made by the artisans in the art department.”
“The show” referenced in the quote above is Wildwood: The Exhibition, due to open soon at the Design Museum in London.


(via justin lind)
Tags: Laika Studios · movies · stop motion · trailers · video · Wildwood
There's a bit of accidental gamification going on with my releases.md gist. Feels good to work on something and then see it reflected in that file.
AI-generated images don’t have to look like AI-generated images…they can be any style you want. (It’s just going to get more and more difficult to tell…)
In 2005, NASA’s Jet Propulsion Laboratory calculated that the water stored by China’s huge Three Gorges Dam is enough to increase the length of a day by 0.06 microseconds and move the North pole by 0.8 inches. (A 9.1 earthquake had a bigger impact.)
I like to hide tabs in GitHub projects that I don't need. Finally figured out how to get rid of the Agents tab on the web. Go to your profile → Copilot settings → Visibility.