Shellsharks Blogroll - BlogFlock

Files are going away. Here's how to manage invisible documents - Johnny.Decimal

2026-11-14T06:10:06.000Z

In this post, we'll investigate how modern apps are moving away from the notion of a 'file', and how this makes our job more difficult.

The solution is simple: use your JDex to tell yourself where things are. This trick works just as well if you want to save links to documents stored on a website, in SharePoint, or anywhere else.

This post has an accompanying video. Sign in with your Pro account to view it.

MindNode modernises

One of my most beloved apps, MindNode, has rebuilt their app from scratch. What was MindNode is now MindNode Classic, and while they say it's still maintained, evidence suggests otherwise.

Figure 0260A. 😐

That's fine. Software marches on, so let's download the new MindNode and move with the times. Okay, done that. Now let's go to File > Open… and open the mindmap I was working on yesterday.

Sad trombone sound 🪊

Open… opens another copy of MindNode's own document browser. You never get to just choose a file to open." autoplay controls height="1080" loop muted playsinline preload="auto" width="1728"> Download the video

Figure 0260B. How Gen Z opens files?

That is … not what I was expecting. I was expecting this.

Figure 0260C. How Gen X opens files.

There is no file

This will be familiar to those of you who use online tools like Google Workspace, Canva, and one of my favourites, tldraw. There, you have no notion of 'a file' that can be saved. Where would that file even be?

This abstraction is at the root of an issue made famous by the 2021 Verge article 'File not found: A generation that grew up with Google is forcing professors to rethink their lesson plans'. How would you know how filesystems worked if you'd never used one?

In this model, your files are managed entirely by the app that owns them. MindNode is using an Apple CloudKit database to store my 'files'. I put that in quotes because it really is incorrect to call them 'files' any more. They're now 'documents' in a database.

To what end?

There is a good reason for this: sharing.

How would Joe Lunchbox go about using this file on his iPad? Or sharing it with a colleague? I don't mean emailing a copy, I mean real-time collaboration. Did I mention that his colleague uses MindNode on his Apple Vision Pro?¹ That's difficult if you don't know what you're doing. But it's enabled for free by CloudKit.

Not only is it free as in money, it's very simple for a developer to set up. They get the storage, security, and sharing provided by Apple as part of the CloudKit service. I know this because I'm using it myself for the app I'm building. It's a really great feature.

One more place a thing can be

Okay so what? Files don't exist, they're stored in a database. What's my Gen X problem?

My problem is that this is yet another place that a thing can be.

When I sit down to work on this mindmap – where I'm trying to clarify the extend-the-end pattern for Felicks on Discord, and for myself so I can explain the idea more clearly – I need to reconfigure both my computer and my mind. I need to recall my notes, figure out what I was going to do next, and of course I need to open up the mindmap.

The list of places I need to recall for this trivial task is now:

My JDex note (Obsidian).
The Discord thread (a URL).
The location of this mindmap.

That's not a long list, but it's this silent friction that makes our computers hard to use: every time you context switch to a new task you need to figure out where this stuff is. And of course the list of possible places is different for each thing that you do.

Undiscoverable

'The location of this mindmap' has now become very un-discoverable. I can't go pecking about my filesystem for it. In writing this post I created a test file with a unique name and used the command-line mdfind, which hooks into the system Spotlight database, to find it. No results.

I asked Claude – an absurdly efficient finding machine – to tell me where it was. It spun its wheels with find commands for 3 full minutes until I interrupted it. That was when we realised that the document is buried in a SQL database, invisible even to the all-knowing LLM.

Forgettable

Discoverability is one side of the discoverability/rememberability coin. (I probably won't take that analogy much further.)

Let's say I come back to this thing in a month. I can't even assume that I'll remember that I created this mindmap. In a world where all your stuff is in one folder and to 'do work' you open that folder and rummage around, this isn't a problem.

But in a world where my stuff is in databases stored in CloudKit, I need to remind myself that it even exists before I know to go looking for it.

Figure 0260D. Trying to figure out where a thing is.

The solution

The solution of course is your Johnny.Decimal index, or JDex. You need a place that you control, and this is it.

As a reminder, the typical structure of a JDex entry is this.

# 11.11 Title of entry

- Locations:
  
- Related to:
  

---

This is where the text of the note goes,
so on and so forth, blah blah.

The 'frontmatter' above the --- line varies depending on the information you need to store. If you use Obsidian you might use YAML properties. I just use a bulleted list.

And so in this list of locations, just tell yourself that there's something in MindNode. I find the gentle pointer to be enough: you don't need to be explicit. Just guide yourself to the place.

- Locations:
  - MindNode

Figure 0260E. Your JDex tells you where everything is.

That's it. Keep it simple and you'll be more likely to do it – because you still need to remember, in the moment, to create this breadcrumb.

Use the 'deep link' if you have one

A 'deep link' is one that opens a document directly, rather than just the app or site. They're typically very long, with a bunch of random characters at the end. Here's one to a Google Drive document.

https://docs.google.com/document/d/1N5ZnznL3LkHmt-vm4z7jF1YKGqzbbhmkPs3J0aiiC5k

MindNode will generate a deep link at File > Copy Link to Document.

So now you can use this link in your JDex entry, and clicking it will open the document directly.

- Locations:
  - [MindNode](https://mindnode.com/document/4CZ02g…)

Formalising 'Locations'

I'm going to make this simpler and more useful by recommending you assign standard zero category 01 Locations to your system.

You'd then define an ID for each location, and use links. For example if you define 01.11 MindNode and link to it from any note like this…

# 11.11 Your ID

- Locations:
  - [[01.11 MindNode]]

…then you'll get a list of backlinks from the MindNode note that tells you what you've stored there.

Not having given this deep thought yet, I'd probably suggest creating broad IDs for classes of locations, then extending for specific locations. Because you only get 90 locations.

01 Locations
   01.11 Apps & websites
         + Canva
         + MindNode

– then your link would look like

- Locations:
  - [[01.11 Apps & websites+ MindNode]]

'My online life' is now 'My digital life' and other content edits - Johnny.Decimal

2026-11-14T04:02:42.000Z

1 category in the Life Admin System has been renamed.

14 My online life is now 14 My digital life

5 IDs in the Life Admin System had their descriptions and examples edited.

14.11 My computers & servers
14.12 My mobile devices
14.13 My wi-fi & network devices
14.14 My data storage & backups
14.15 My accessories

1 ID and 1 category in the Small Business System had links added.

11.21 Business plan
13 Money earned, spent, saved, & owed

A few other minor content edits and a typo correction.

fluttering - James' Coffee Blog

2026-11-14T00:00:00.000Z

Between amber leaves
The fluttering bird dances —
Making a new home?

Notable links: November 13, 2026 - Werd I/O

2026-11-13T22:50:23.000Z

Most Fridays, I share a handful of pieces that caught my eye at the intersection of technology, media, and society.

Did someone forward this to you? Subscribe for free.

The Two Rails of Leadership

I’ve been thinking a lot about my entrepreneurship journey lately — and how I can draw lessons from it to help journalism — and this piece from Corey Ford landed at the right time.

“Culture and strategy are the two things a leader is responsible for designing; everything else is delegable. […] Your job as a leader isn't to come up with the next big thing. It's to build the culture, the people, and the process that can continually do this: a team whose velocity of learning outruns the pace of change. That capability lives on the culture rail. The strategy is what it produces.”

I basically punched the air.

The thing is, so many of us have got into the trap of either managing the incremental changes that we feel are needed to keep the lights on, or defining ourselves by the product we create rather than who we create it for. In doing so, we build inward-looking cultures that put us in danger of losing touch with the world we’re trying to serve.

Corey’s example of the newspaper is apt: do you produce words on a page, or do you serve a specific need for context and understanding? Only one of those answers will survive change. Someone asked me recently what I thought the business model for news should be, and I responded that it’s the wrong question to start with. The question is: how can we be most valuable to the people we’re trying to serve?

I’m constantly referencing Gina Chua, who points out that we should care more about saving what journalism does for us than the specific institutions that produce it today. A culture centered on providing the effects of journalism will stand the test of time; one centered on institutionalism will not.

Can existing media institutions recenter themselves around solving needs rather than finding ways to optimize the same products, workflows, and ideas? I don’t know the answer; it’s certainly not a given. But I do know that it starts with building the right culture. And if those institutions can’t adapt their cultures — well, we’ll just need to build new institutions.

Introducing the Atmosphere Community Cooperative

AT Protocol is the underlying protocol that powers Bluesky, Chicago.com, Sill, Surf, Leaflet, and many other social apps and services. It’s designed to be open, decentralized, and interoperable: the idea is that no one entity owns it, and any tool built on the protocol is theoretically interoperable with all the other tools. It’s one piece of an emerging, diverse open social web that’s free of monopolistic control.

At the News Product Alliance summit last month, Joe Germuska and I presented on AT Protocol with Frank Sharpe and Andjela Padejsky from Chicago Public Media. The protocol has allowed them to build Chicago.com more efficiently: they know two codebases they write on the protocol will work together. It’s also proven to be useful scaffolding for LLM-assisted development, which has allowed their development team to punch above its weight. There hasn’t been much attention paid to the protocol in spaces like journalism: our presentation and events like Protocols for Publishers are the exception. Nonetheless, I’m certain it will be an important building block.

But a building block that can be manipulated by a single owner introduces risk that newsrooms and other mission-driven organizations shouldn’t tolerate. (Consider what happened to Twitter.) Independence is vital; without it, a building block becomes a single point of failure.

Now the folks behind AT Protocol ecosystem pillars like the annual Atmosphere conference have filed to form a co-operative organization.

“As we look toward the third AtmosphereConf in 2027, we’re happy to announce that we’re transforming into a cooperative dedicated to supporting the health, growth, and stability of the Atmosphere.

This is the first post from the Atmosphere Community Cooperative, a new global cooperative based in Canada that is governed by and for the community, for the good of the ecosystem and the people inside it.”

This is a big deal for lots of reasons. The first is that it further formalizes the idea that a bunch of AT Protocol support and governance is independent of Bluesky. The protocol itself is already supported by a working group at the IETF, and the directory of identities is forming as an independent Swiss Association. This, then, is a third independent ecosystem org, ensuring that it’s not just the protocol that’s decentralized, but governance and support of it, too.

I do also think it’s really interesting that the Atmosphere Community Cooperative will be a real co-operative organization, adhering to international co-op principles, and established in Canada. Between this, the Swiss Association, Bluesky’s Delaware Public Benefit Corporation status, the IETF working group, and other participants like Dutch public interest organization the Modal Foundation, there is a plurality of geographies, organizational approaches, and communities involved in the protocol’s development and success. That feels like much-needed organizational innovation in addition to the technical gains. All of which makes AT Protocol safer to build on.

Medium is launching a new note-taking app called TK

Medium is launching its new app, TK:

“The company is launching a new app, called TK, that combines note taking and journaling with something resembling a private social network. The app is meant to get you writing more, and help you share that writing with friends and fans. And it’s part of a much bigger idea about blogging, and a much broader thesis about the state of the internet.”

I’ve been using it for a while and it reminds me of some of my best experiences on the internet. It also contains a bit more AI than you’d expect to see in a writing app, but it doesn’t hinge on those features: for some people they’re useful ways to add more information, fill in blanks, or do some smart retrieval from notes you’ve already written, but you can get a great experience from TK without using it at all.

The Verge article notes that it’s a bit like group chat, and I think that’s right. The way I’d put it is this: if Medium (or any blog, including this one) is a magazine where complete pieces are published, TK represents the messy conversation, note-taking and idea-sharing that exists before that point. That’s an imperfect description, though, because a lot of this really is conversation: something that will never be a published thought. Some of the interactions remind me of LiveJournal at its height, which is probably the biggest compliment I can give to any new service.

That’s also why I’m not worried about it not operating on an open social web protocol. It’s intentionally closed for the right reasons, in the same way that my group chats are closed. These are ideas that I want to share with select people, not the world. And because of that, I’m more likely to be open, be wider with my thoughts, and share less-finished ideas. All of those are good things. And then, should you publish: yes, put it on your own domain, or somewhere like Medium, depending on your values and preferences.

I like it very much, is what I’m saying.

Spacefast and beyond

Spacefast is instructive as one of the first truly AI-native infrastructure services I’ve seen, and as a genuinely useful service that solves a bunch of problems for people building with AI, particularly at the enterprise level.

As the announcement post says:

“In a world where writing code is as easy as talking to a magic box, it is still not as easy as it should be to build or share what you made. Services we use were built for a different time.”

The idea is that if you’re using agents to build prototypes, you can tell them to save their work to Spacefast and it’ll take care of all the hosting, databases, etc transparently in the background. It also includes version control, so prototypes can be rolled back to previous iterations really easily. The whole thing is accessible both via an AI interface and over the web.

To set it up, you just drop a prompt into your AI service of choice, and it does the rest for you. I plugged it into Claude and it went away and set everything up; meanwhile, I could follow progress through the web interface.

I asked Claude to build a one-shot web page that makes my Bluesky posts fall like rain. Publishing was indeed seamless. The web app Claude made defaulted to private, but I was able to make it public really easily. A second app, a homepage for US news based on my Speaking Truth to Power feed, was not quite as smooth, but I was able to adjust and fix the problem quickly.

And then you have an easy way to build prototypes using AI and share them, without really having to worry about any of the infrastructure. At the enterprise level, that includes SSO and top-level access policies, so you can let people play and share prototypes internally without running the risk of any of that work being shared them externally. That saves enterprise organizations (including newsrooms) from having to set any of that up themselves.

I’m talking about prototypes because I think that’s what this is particularly suitable for. Longer-lived applications, even if they run internally, should be code reviewed and held to a higher level of scrutiny. (It must be said that hosted apps run on Automattic’s infrastructure, so they could be shared much more widely — but there are safety and maintenance implications to consider.) But if you’re building something to show a few colleagues, or external people as part of your user research, I think Spacefast genuinely makes the process far easier and safer for businesses than it was before.

It also solves a problem for Automattic. As more people vibe code sites, fewer people will use WordPress. Putting a service like Spacefast in place, and establishing it early, helps ensure that people will stay in its ecosystem.

And more:

Here are some of the articles I didn't get a chance to go into in depth this week.

A strategic shift for Online News Association

ONA has long been a gathering post for digital folks in news, but there won't be a conference in 2027 and its staff is being laid off. One stated reason is that both professional development and sponsorship budgets have sunk, which is probably the most worrying thing to me about this announcement. In a time of crisis we need more opportunities for new skills and collaborations, not fewer.

Top 50 US news websites: Four in five see traffic drop in October

Only nine of the fifty biggest US news sites avoided a year-on-year decline, in the month before a midterm election. The steepest drops are brutal: 60% at Newsweek, 47% at Politico. Substack is listed as one of the outliers, but why is it in this list to begin with?

How to fight back against AI-generated local news and pink slime: double down on civic engagement

AI-generated news sites are taking on local news: first in news deserts, and now in places with established newsrooms. Some of it is barely-disguised partisan content. The solution for human journalists: "be at the farmers’ markets, organize events, serve as a public forum, give people a chance to meet and greet your journalists, and be really a part of the fabric of the community."

Postal Service Surveilled Protest Activities Online With Little Rules, Oversight

"The U.S. Postal Service used social media surveillance to monitor and report on protests across the political spectrum, government records show." It turns out it has an "internet Covert Operations Program", or iCOP for short. This seems to run counter to its usual description as an independent agency.

Apple Notes importer for LAS/SBS - Johnny.Decimal

2026-11-13T22:07:21.000Z

The Apple Notes import feature is broken. It's always been garbage – I just don't think Apple cares about it at all.

In the current version, you can't import a nested folder of notes using the 'Preserve folder structure on import' feature. It just doesn't work. This breaks the LAS/SBS import process – all your notes end up in one flat folder.

Download now includes an import utility

The system downloads now include an import utility. Find it in the JDex files > Apple Notes folder.

Double-click it, approve a bunch of warnings and prompts, wait a minute while each of the imported notes is moved into its proper folder, and you're done. More details are in the README.

Until Apple breaks something else… 🫡

2026 Self-Host User Survey Results - Kev Quirk

2026-11-13T20:59:00.000Z

2026 Self-Host User Survey Results

by Ethan Sholly

Visualized results across 6,200+ responses to 2026's annual user survey.

Read post ➡

Fascinating look at how people (like me) self-host their tech stack across the Internet.

There were some surprises for me, like Debian being above Ubuntu, and Android being far more popular that iOS.

Interesting reading if you like data.

Thanks for reading this post via RSS. RSS is ace, and so are you. ❤️

You can reply to this post by email, or leave a comment.

Gadget Review: Thermal Master P3 Lite - and Linux Instructions ★★★★⯪ - Terence Eden’s Blog

2026-11-13T12:34:48.000Z

The good folks at Thermal Master have sent me (yet another) thermal camera. The P3 Lite has a detachable macro lens. As well as being good for scanning your house for leaks and looking for wildlife, it can focus down on tiny electronic components.

I've previously reviewed their P3 with adjustable macro focus, the chunky T2 Max, and the wildlife spotting DV2 - how does this compare?

Like most thermal cameras, it has a low-ish resolution - 256×192, but their X³ software upscales it to 512×384 running at around 25fps. That's pretty decent for a small plug-in camera.

It also has a magnetically attachable lens for macro-focus, which allows you to see close up details.

You can read the full specs but, basically, it is suitable for all domestic tasks.

Sample Photos and Video

Note, all photos were originally 6MB JPGs - I've losslessly converted them to JPEG-XL and kept the (minimal) EXIF.

Point the camera at something and you'll see which bits are hot. Useful for checking your radiators are working.

Images are stored in portrait mode and have been rotated. As you can see, some of the UI elements remain in portrait mode. Here's a just-finished dishwasher:

A 30 second video is about 12MB - it does include audio from your phone's microphone. I've recompressed and rotated this. It is a tap running cold, then hot, then cold again.

https://shkspr.mobi/blog/wp-content/uploads/2026/11/Hot-Water.mp4

Video quality is, I think, excellent.

There are a variety of different thermal palettes to choose from - and the camera excels at spotting where a circuit's hot spots are. Here's a Raspberry Pi:

So what about the macro lens? Here's a close-up shot of a circuit board where I haven't used the included lens:

Not bad, but what if we click the magnetic lens on?

Whoa! Much better. Here's a video demonstrating it.

https://shkspr.mobi/blog/wp-content/uploads/2026/11/lens-on.mp4

Hardware

The camera has a reasonably long USB-C protrusion - so it will fit in your phone even if you use a chunky case.

Because it is USB-C, you can have it facing you or facing away from you. Handy if you want some thermal selfies!

The lens is a simple affair. Note that the protective cover it ships with might not come away cleanly:

I had to prise off the sticker to prevent that residue dirtying the main lens.

It ships with a USB-C extension cable (with USB-A converter) - that's handy if you want to keep camera and phone somewhat separate. It also makes it easy to plug it in to a computer. More on that later.

There's a C to Lightning adapter if you're still stuck on an old iPhone.

Finally, it comes with a little carry case emblazoned with the logo.

The App

The Thermal Master Android App isn't brilliant. It's slow to start and has a bafflingly complicated interface. It takes a couple of clicks to take a photo, rather than the single click you'd expect.

As you can see from the photos, despite the timestamp being rotated, other on-screen elements aren't. You can possibly change this manually in the settings, but it is rather annoying that it doesn't do it automatically. Speaking of which, there are lots of settings to fiddle with - but you'll probably just end up on one that you like and leave it there.

The app does work, but I wish it were snappier.

Linux Usage

There is a Windows App. I couldn't get it working in either VirtualBox or Wine, so here's how to display the camera's feed in Linux.

Plug in the camera via the included USB-C extension cable. It will show up in lsusb as 3474:42b2 Thermal Master Technology Co., Ltd. P2 MAX

Note: the device was previously known as the P2Max and this is reflected in the firmware.

Add a new udev rule with sudo nano /etc/udev/rules.d/99-thermal-p2max.rules

Copy and paste this into the file, then save and exit.

# Thermal Master P2 Max USB Rules
# Allow read/write access for all users
SUBSYSTEM=="usb", ATTRS{idVendor}=="3474", ATTRS{idProduct}=="42b2", MODE="0666"

Reload the rules with sudo udevadm control --reload-rules && sudo udevadm trigger

Hurrah! Your Linux machine can now control the camera. There's two ways to view it - and easy way and a hard way.

The easy way is to use https://ferrolho.github.io/thermal-master-viewer/

Open the website in Chrome, click the "Connect camera" button, after a moment you should have a view like this:

The harder way is to use dedicated Linux tools like OpenP3 or Thermal Camera Viewer - you may need to hack around with them so they see the device. Usually that's as simple as adding the PID to the code.

Cost

You're looking at around £220 for the camera and can expect to pay about £40 for the macro lens. You can sometimes find them bundled for slightly cheaper.

Readers of this blog get an extra 6% off at Thermal Master or Amazon UK/US with code THERMLITE6.

Unfortunately, that's pretty cheap for a thermal camera. At that price, I'd recommend buying one to share with your local community. If you are a member of a hack-space, these are invaluable for finding hot-spots on your electronics.

The image quality is excellent. There's no bogus "AI" features to hallucinate what it thinks it saw - you just get the raw image. It would be lovely if it presented as a proper webcam on Linux, but there are open source tools to bridge that gap. The detachable macro lens is handy, and the magnet grip is reasonably strong - but I wish it had a strap or something to prevent it getting lost.

There's not a lot of difference between Thermal Master's various cameras. If you want to swap between wide shots and macro shots, this is probably as good a device as any.

2026-11-13 09:52: Started writing letters to people a month ago now, but haven't had a single one... - Kev Quirk

2026-11-13T09:52:00.000Z

Started writing letters to people a month ago now, but haven't had a single one back.

My handwriting must be really bad! 🙃

Thanks for reading this post via RSS. RSS is ace, and so are you. ❤️

You can reply to this post by email, or leave a comment.

Winter air - James' Coffee Blog

2026-11-13T00:00:00.000Z

Preparation is everything. I started my walk by pulling my hands over the sleeves of my jumper that I was wearing under my coat. If my hands were under my sleeves, I would have two layers of protection from the cold. Gloves seem to make my hands feel colder, whereas my sleeves hold the vestiges of radiated warmth that keep me more comfortable. With my coat sleeves extending over my hands, I was ready for my walk.

The shade foreshadowed winter: the air was cold, but not quite cold enough to see my breath. I walked at pace, paying less attention to the immediate surroundings and more on finding my stride – finding warmth – that would propel me through the walk. Acclimatising (to the extent one can in the cold), the sparse trees started to catch my attention, as well as the remaining golden leaves from the autumn season. When does winter begin, I wonder. Has it begun already, with fierce cold air during day and delicate frost lay over night?

I notice my shoulders are tense on this walk. I am tensing my muscles as if doing so will help me hold onto heat. Relief comes when I make my way out of the shade and into the sunlight, where golden warmth shines on my face and I start to gain heat. I have found my stride. I look up and around to appreciate the trees – the twisting red colourful ends on the top of one tree’s branches, the silvering bark that shines where leaves once fluttered in the air – which warms my heart. It’s a beautiful day.

Back home, I have the heating on for the first time this season. I am already thinking about gloves and layers and days spent at home reading and drinking tea, the latter dreaming about reading (books aside from textbooks) being more of a hope because I read in cycles and it can take me quite a while to get started with reading in my free time. Nevertheless, I have a few books and mangas I would like to read this winter – words that carry warmth and knowledge that will carry me through the cooling season outdoors.

Spacefast is a new kind of infrastructure for the vibe coding age - Werd I/O

2026-11-12T18:59:18.000Z

Link: Spacefast and beyond, by Batuhan İçöz

As the announcement post says:

“In a world where writing code is as easy as talking to a magic box, it is still not as easy as it should be to build or share what you made. Services we use were built for a different time.”

Investigating The President - Molly White's activity feed

2026-11-12T18:04:15.000Z

Investigating The President

My favorite thing about building research tools like the map of the Trump family’s crypto empire is when someone else does something great with them. Glad I could help Coffeezilla with this one; it’s a really clear explanation of how Trump is getting rich off the presidency.

Data: Mapping Trump’s crypto empire

Book Review: This Is for Everyone - Tim Berners-Lee ★★★☆☆ - Terence Eden’s Blog

2026-11-12T12:34:10.000Z

Is this a biography of a man, or of his mission? Is this the story of how the web came to be, or a lament for what happened to it? Do you crave the inside gossip, or the PR sanitised gloss?

This Is For Everyone, but I think it ends up being slightly unsatisfying for anyone.

It starts with the normal charming family anecdotes. No grand revelations or hidden secrets, just a lovely ramble through growing up. It's hard to write a memoir without delving in to some of the frailty that makes us human. I wasn't expecting a "tell all" or histrionics, but divorces are waved away in half a sentence and the emotional aspects of life as the inventor of the web are almost totally absent.

Where his emotions shine through is the constant undercurrent of rage and disappointment at how the web has been corrupted.

those operators turned the supposedly non-profit DNS address system into a nauseating free-for-all. Domain names would become an asset class for investors, rather than a tool for innovators

[…]

DNS fell into the hands of some rather unscrupulous operators. What had been an unremarkable piece of back-end internet infrastructure was temporarily transformed into a revolting for-profit bonanza.

He is unashamedly catty about some of the villains of the piece. I won't regurgitate their names or give them the oxygen of publicity, but it is nice to see a senior tech figure call out some of the madness and harm caused by others. Similarly, he is harsh in his criticism of how some browser vendors have behaved. We can argue about whether XHTML was a mistake, but the "moat" of working with non-standards compliant pages makes for a successful monopoly.

It isn't exactly a whinge, but you can hear the exasperation in his tone. He gave the world a magnificent gift and some people are trying to poison the well for all of us.

A few chapters are dedicated to the rise of AI - and its perils. Similarly the tension between addictive algorithms and staying sane. These are certainly interesting thoughts, but perhaps belong in a different book. Memoirs are, typically, not about the future.

A large part of the book is taken up with Tim's "difficult second album" - Solid. I desperately love the idea of Solid - keeping your data personal and only allowing access for explicit and limited purposes. But I also think it is becoming clear that most people just don't care that much and, even if they could be bothered with all the admin, there's no signs that businesses want to get involved with it.

At times, it feels like the book is mostly a pitch document for why Solid is the future. Again, not exactly what I was expecting from a memoir.

I think every single part of this book is good - the personal memoir, the history of the web, the future of tech - but as a whole they feel a little disjointed. I think I'd rather have read three different books rather than have them jammed in together. I suspect the hard-core geeks will skip the domestic stuff and the average reader will be somewhat baffled by the dive into deepfakes and Mastodon.

It does, however, contain a message of hope. The web is still here and it will last in one form or another for eternity. We have to keep working for it though.

[M]y work with Danny Weitzner had shown me that it was not enough simply to release new technology and hope for the world to improve. You had to develop technology and society together. You really had to fight, in a principled and continuous way, for human rights.

We need that tattoo'd inside the eyelids of every tech executive who has tried to subvert the web for their own ends.

News needs entrepreneurs. Where will they come from? - Werd I/O

2026-11-12T10:00:00.000Z

“What is the New York Times doing?”

I’ve lost count of the number of times I’ve heard that in various journalistic settings: inside newsrooms, in the hallways at conferences, in cross-newsroom chats. It makes some sense: the Times is the industry’s 800-pound gorilla, and its size and wealth give it the power to innovate and maneuver.

On the other hand, I often found myself thinking: where is the imagination?

Copying a larger newsroom with a different focus and set of resources available to it is not a strategy. Neither is incremental optimization. And they’re certainly not innovation.

Unlike the Times, which has cemented its position through a smart portfolio of services like games, recipes, and product recommendations, news as an industry is dying. In the US we lose at least two newsrooms a week. Newsroom leaders can’t afford to be followers.

Newsrooms need to change. The question is: how?

Most newsrooms are a fraction of the size of the New York Times and can’t follow its strategy. (And anyway, the Times has taken those positions.) But they can follow its openness to radical change.

Relative safety vs existential threat

Fifteen years ago, a16z co-founder Ben Horowitz published his definition of a peacetime CEO vs a wartime CEO:

Peacetime in business means those times when a company has a large advantage vs. the competition in its core market, and its market is growing. In times of peace, the company can focus on expanding the market and reinforcing the company’s strengths.

In wartime, a company is fending off an imminent existential threat.

As much as I have differences with a16z’s business priorities and its impact on democracy in the years since, I think this is a useful high-level frame. Whereas Horowitz’s companies needed to win against their competitors, newsrooms need to adapt to a rapidly changing world — but it’s an existential threat nonetheless.

A peacetime CEO can focus on improving what the company already does: it’s already winning, and the way to improve its standing is to optimize. On the other hand, a wartime CEO needs to take more risks and be unafraid to fundamentally change the business or to kill sacred cows. A peacetime CEO may feel like they don’t need to understand their customers more deeply, because it’s already working; a wartime CEO needs to understand them incredibly deeply so they can beat their competition.

Let’s ditch Horowitz’s unpleasant war language and call these scenarios relative safety and existential threat. And then let’s turn this lens towards news.

Journalism is operating in an existential threat environment. We are in the fastest period of technical change in my lifetime, and the fastest period of democratic change. The institutions people trust, the places they get their information, and the pressures on anyone who tries to hold power to account are all shifting at once. This should also imply the fastest period of journalistic change. Instead, most newsrooms are in decline; the industry certainly is.

Newsrooms are organizationally really interesting. Very often, the CEO (or effective CEO) is a journalist rather than an experienced business operator, and is hyper-focused on the journalistic process and reporting great stories. I don’t think that’s unreasonable on its face: journalism is, after all, at the core of what a newsroom does, and putting a journalist in charge enshrines journalistic values at the center of the organization. But it does mean the hard task of building strategy often ends up sitting elsewhere. The hard task of building a culture that supports innovation, meanwhile, is often missing in action.

A small number of newsrooms are lucky enough to employ a Chief Strategy Officer. The 19th (where I worked as CTO) has Alexandra Smith; The Marshall Project has Ebony Reed. The largest newsrooms have multiple full-blown strategy teams. But most have no-one. That leaves strategy in the hands of a Product lead — if one exists.

Sometimes, particularly in smaller newsrooms, no-one’s watching the ball at all. Other times, the Product lead acts as if we’re in a time of relative safety, tweaking dials and optimizing workflows without meaningfully engaging with the existential threat. But building culture and strategy to solve for the industry’s existential problems is where the real work needs to be done.

And that work is mostly not being done. The Reuters Institute for the Study of Journalism’s 2026 Trends and Predictions report finds that 59% of digital leaders are focused on improving their core products rather than reinventing them. ASU’s recent News + AI Field Report, written by my JSK advisor Djordje Padejski, rounds up AI-assisted innovation projects across newsrooms between 2022 and 2025. As I noted at the time, only 20% of them involved business experimentation, and most of those were conducted in programs that paid newsrooms to do business experimentation.

My conclusion is that it’s simply not within the comfort zone of most newsroom leadership, which is editorially-focused, not business-focused. They concentrate on building an editorial culture to produce great journalism, but don’t create a broader culture of innovation.

Given that strategy often sits with the Product lead by default, most Product leads are optimizing when they should be reinventing, and culture often isn’t addressed at all, how can we redefine Product leadership to include the skills, mandate, resources, and training to bring about real change?

Newsroom entrepreneurs

After conversations with product and foundation leaders at this year’s News Product Alliance Summit, I had an epiphany.

In most newsrooms, the Product lead needs to lead strategy, product, platform, technology, brand, design, and audience; they need to figure out how to build revenue into the offering; they need to be hyper-focused on the underlying needs of the people and communities they’re addressing. In tech, where the Product role originated, product managers are far narrower: they tend to a feature, or even part of a feature. Here, we’re not talking about product managers at all.

Almost a decade ago, I led West Coast investments at Matter, the media startup accelerator. Prior to that, I went through the program myself with my own startup. We backed cohorts of early-stage startups with the potential to be transformative for the media industry. They got a cash infusion, but more importantly, they got a five-month program that was rooted in d.school design thinking but touched on everything from culture to revenue models. That training genuinely changed everything about how I build products and teams.

Not all of the startups succeeded: media is a tough market. But most of the founders who had both received that training and deployed it in service of bringing their businesses to life had gone on to be change-makers elsewhere. The program found people with the right kind of prototype-driven, experimental mindset and a bias towards action and it gave them directed training, a network, and resources. It turns out that people with these traits are valuable — and, if they’re allowed to, can make transformational strides.

Matter alumni include high-level product leaders at companies like Hearst and Snapchat, and even the CTO at the Biden-era HHS. Change-making alumni like Jenn Brandel are well-known across journalism. Some have gone on to take big swings like building hospitals in India. And I know I couldn’t have led engineering, IT, and security at ProPublica without everything I learned there.

The Product leaders we need in today’s newsrooms are more like this. They’re not product managers: they’re entrepreneurs. People who can build their wings on the way down, aren’t afraid to take big risks, have the cross-functional skills to execute, and can forge radically collaborative cultures that are centered on the needs of their communities.

We need them to operate this way, and we should redefine (and potentially rename) our Product lead roles to take this into account. News needs entrepreneurial, wildly experimental, human-centered thinking with a mandate to execute at speed, reporting to and working alongside the CEO. And, for a few rare operators who both have a journalistic background and these entrepreneurial traits, they could be the CEO.

The trick is, there aren’t that many people who can operate this way in news. So now that we’ve re-scoped the role, the question becomes: how can we make more of them? And how can we make sure they have that mandate and leeway to operate?

Matter as an accelerator no longer exists, although Columbia University’s Sulzberger Executive Leadership Program shares its leader and much of its DNA. Media executives can learn from the program, and I highly recommend its leader Corey Ford’s coaching (I’m a happy mentee). But what’s missing from the ecosystem is the experience of building something completely new from the ground up, and the on-the-job experience it provides.

Building a mission-driven startup in media is training-by-fire. With Matter gone, there are very few other opportunities to do it — and very little pipeline for change-making entrepreneurs.

We need that pipeline. So let's build it.

Building an ecosystem of innovation

This is the first part of a series. In future posts I’ll discuss:

A proposal for how we might, concretely, build an entrepreneurial pipeline in media
How we might ensure that newsrooms and media institutions give innovators leeway to work

In the meantime, if this speaks to you, please reach out; I would love to have a conversation. I’m particularly interested in speaking with people and organizations that have been working on the same problem.

Leaders are responsible for culture and strategy. Everything else is secondary. - Werd I/O

2026-11-12T06:16:57.000Z

Link: The Two Rails of Leadership, by Corey Ford in Point C

I’ve been thinking a lot about my entrepreneurship journey lately — and how I can draw lessons from it to help journalism — and this piece from Corey Ford landed at the right time.

“Culture and strategy are the two things a leader is responsible for designing; everything else is delegable. […] Your job as a leader isn't to come up with the next big thing. It's to build the culture, the people, and the process that can continually do this: a team whose velocity of learning outruns the pace of change. That capability lives on the culture rail. The strategy is what it produces.”

I basically punched the air.

CASA, conscience and the absence of constraint - destructured

2026-11-12T05:00:00.000Z

When I posted the CASA document a few weeks ago, I expected — indeed, hoped for — some debate. I just didn't expect the first major critique to come from an institutional source. Open Source Science (OS-Sci), an international outfit that partners with organizations like GitHub and the Rust Foundation, has published a blogpost outlining their objections.

If you haven't read "Computing As Social Activity," start there, but I also encourage you to read "Does Free/Libre Open Source need a conscience?" The issues demand serious consideration, and if CASA is worth putting into practice, it must be able to stand up to the pressures of disagreement.

To that end, I have a few points in response:

CASA is not a legal paradigm. It doesn't "shift rights" or "claim jurisdiction." It isn't written as a license, doesn't seek to "bind" with contracts.¹ It doesn't even assert "rights," which might imply a legal framework. The absence of an enforcement mechanism (consistent with the emphasis on personal agency) is not an oversight. The principles may conflict with FOSS in particular scenarios, but there is no coercion that could "break" the freedoms of the Free Software movement. Just as a personal commitment to pacifism may dissuade you from using software for weapons targeting, adopting CASA principles ought to change how you use computers, but not because it has impinged on your freedom. You remain free; all that has changed is the range of things you might choose to do with that freedom. That domain, the scope of things you would wish to do with computers, is where CASA operates.
Agency defines community in CASA. OS-Sci's conclusion points to a common tension between those two concepts. It's a tension the core principles are structured to address. It's true (in line with my previous point) that "CASA offers no procedure, no judge, no due process that decides whose community counts." What it offers instead is a domain and a standard. Principle 1 establishes the domain: Computing is inherently social. As such, using a computer brings you into community with others. As indicated in the discussion, that opens onto other types of community, too, but CASA's focal point is the social domain of computing. If accounting for community entails concern, and concern puts agency before utility, then agency becomes the standard for assessing how well that social relation functions as a community. In short: You are, however briefly and indirectly, in community with anyone brought into the scope of your computer use, and ought, therefore, to respect their agency.
There are times we should feel conflicted. Insisting on agency conflicts with Free Software's freedom 0 only when your aim is to violate someone's agency. If you're not out to violate agency, then there is no conflict. To take one of OS-Sci's examples, designing a weapons targeting system brings the people targeted into the scope of an engineer's computer use. Abstractly, the engineer knows this — what else, after all, is the point of a targeting system? CASA provides a framework for thinking through the implications. Will recognizing that community, and therefore the other person's agency, leave the engineer feeling conflicted about their task? Good; that's CASA is working as intended.
Neutrality isn't granted. OS-Sci illustrates "the neutrality of the tool" with the kitchen knife analogy. Not all software is as neutral as a kitchen-knife, though.² Some is more akin to a landmine, purpose-built to cause indiscriminate harm. The computing landscape is rife with software that functions more like shackles or a snare. Abstracting it all as "code" allows us to ignore the fact that some software is malicious by design and difficult to turn to ethical ends. From a CASA perspective, the best way to prevent unethical use is not with a license, but by making software that lends itself less easily to malicious use.
Ethical reasoning clarifies freedom. It's difficult to think of a more timely or straightforward illustration than a recent post from the GNU Emacs mailing list. There, co-maintainer Eli Zaretskii answered a hypothetical by arguing that consistent application of FOSS principles would require the project to accept contributions they knew to have been produced by slave labor.³ This goes beyond the usual defense that FOSS can't be expected to solve every wrong in the world: By making it a matter of fidelity to "software freedom", what Zaretskii has described is a policy of benefitting from slavery. Institutional slavery has historically solicited exactly that sort of complicity as a defense against calls for abolition. My guess is that, in just about any other context, Zaretskii would disavow slavery outright, but for many FOSS advocates, a schematically narrow focus on software freedom has become a permission structure for suspending ethical reasoning — even when ethics might serve the interests of freedom. As OS-Sci puts it: "the four freedoms matter more, in principle, than any potential misuse they enable." The problem is not that FOSS advocates have no conscience, but rather the perception that conscience cannot coexist with freedom, even in cases where ethical reasoning helps to avoid fundamental contradictions in how we understand freedom.⁴

So yes, FOSS needs conscience. More importantly, though, it needs a framework in which people can discuss and reason about the ethical concerns that arise from the consciences they already possess. If not CASA, then something else.

There is, in fact, a long history of attempting to redress FOSS problems with the introduction of new licenses. If I had thought that those efforts were sufficient, I wouldn't have seen a need for something like CASA. At this point, the prospects for licensing our way to more ethical computing seem rather dim. ↩
"Do Artifacts Have Politics?" suggests that even the neutrality of a kitchen knife is open to debate, but for the case I'm making here, it's not necessary to follow the argument that far. ↩
For what it's worth, I'm not at all convinced that Zaretskii's conclusion really is consistent with the Four Freedoms, which are often said to entail open source distribution as a necessary condition for the freedoms it describes. Why shouldn't they also entail freedom from slavery? After all, slaves generally aren't at liberty to use or modify software "for any purpose," and the slaves in the hypothetical were explicitly coerced into producing software to their captor's specifications. No code contributed under those circumstances really deserves to be called "free as in speech." For the purposes of my point above, though, the internal consistency of Zaretskii's argument matters less than his suspension of any consideration ruled "out of scope" for FOSS. ↩
For example: "A movement that […] wants to bind use to moral conditions, does not stand alongside the FSF and OSI — it stands against them." Note the terminological shift from "ethical" to "moral." There's a philosophical debate to be had on the precise distinction, but in popular use, "moral" tends to connote traditional precepts that are incumbent upon every person in a society, while "ethics" leaves more room for personal variance. CASA is meant as a domain-specific ethos. I don't see any reason why OS-Sci should shift to talking about "moral conditions" unless it's to raise the specter of repressive moralizing campaigns. ↩

Editorial policy: no paid endorsements - Johnny.Decimal

2026-11-12T02:42:57.000Z

I write about and recommend technology. Our upcoming video discusses MindNode. Soon I'll be talking about email, where I recommend Fastmail. I've discussed 1Password, and of course there's my beloved Bear.

In the past, I have published affiliate links for some of these services. This makes me feel awkward – I don't want to tell you how good Fastmail is and follow it with a little footnote: ¹ Clicking this link earns me money.

When I see others do this, I'm suspicious. I don't want you to be suspicious of what I write. So.

1. We do not earn money from endorsements

When I recommend something, I want you to know it's because I think it's the best thing. Not because it's earning me money.

This means we will not include, say, an affiliate link to a product that we recommend.

You might still get a deal

This weekend's post will include a MindNode offer. This is a public offer that has nothing to do with me. I'll still include offers like this where they might benefit you but not me.

This doesn't preclude advertising

I'm not saying that I'll never advertise anything. (Though, let's be realistic.) But if I did, it would be obvious: it would be an advert, plain and clear. I'll say "I'm telling you this because they paid me".

This policy lives at Support > Policies

I've copied this policy there. If we add to it in the future, I'll blog an update.

morning window - James' Coffee Blog

2026-11-12T00:00:00.000Z

Drops of condensed air
Backlit by a golden glow --
Contrast, visible.

The Atmosphere Community Cooperative is another step towards building a truly open social web ecosystem - Werd I/O

2026-11-11T18:49:04.000Z

Link: Introducing the Atmosphere Community Cooperative

Now the folks behind AT Protocol ecosystem pillars like the annual Atmosphere conference have filed to form a co-operative organization.

“As we look toward the third AtmosphereConf in 2027, we’re happy to announce that we’re transforming into a cooperative dedicated to supporting the health, growth, and stability of the Atmosphere.

This is the first post from the Atmosphere Community Cooperative, a new global cooperative based in Canada that is governed by and for the community, for the good of the ecosystem and the people inside it.”

Frankenstein - Joel's Log Files

2026-11-11T18:40:00.000Z

Back in my younger days, my high school would hold yearly book reading events. The choice one day was Frankenstein, by Mary Shelley.

Quite a bit of students would line up, waiting for their turn to read a few paragraphs of the novel.

To be honest, while I found the event very interesting, in practice it was a bit boring. I never stayed to listen for more than a few minutes. I didn’t enjoy the voices much. I believe many of the kids there only did this for bonus points, not putting much effort into performing and adding emotion to the written words.

Besides, I ignored the legacy of the novel itself, it just seemed like an old story that we were being forced to listen to.

I always wanted to participate and reading aloud. I actually really love doing so with family or at church, but I was too shy to do it at school.

I didn’t realize how important this novel was until much later. Very weird of me since I already loved classic science fiction stories. Little did I know, that before H.G. Wells or Jules Verne, there was Mary Shelley.

Learn from me, if not by my precepts, at least by my example, how dangerous is the acquirement of knowledge, and how much happier that man is who believes his native town to be his world, than he who aspires to become greater than his nature will allow.

Other adaptations

I may as well mention that I watched some adaptations of the novel after I read it.

First, I read Junji Ito’s Manga version, which follows a lot of the same plot up to a point in the story where it changes completely and goes for a more spooky take on the creature and its actions. However, there are some things that happen way too fast, which I would have liked to see better developed.

Nonetheless, it was an interesting take on the characters and the message of the novel.

The art of this is also amazing, to be expected from Junji Ito’s style. The design of the creature was also pretty terrifying!

The finale ends up very similar to the novel. Though the story felt less nuanced, it still hit during the important moments.

Afterwards, I watched the movie! Not Guillermo del Toro’s latest adaptation, but the original film from 1931—which is going to become part of the public domain next year.

Its story pales in comparison to the original novel. However, I really appreciated it by what it is on its own, and I am willing to call it a masterpiece for the time. It has some dated moments and the fact that Victor and Henry swapped names is hilarious to me, but it still contains the original essence of the novel, as the creature—as dumbed down as it is compared to the book—is still misunderstood and feared.

The film features many iconic elements to the mythos, such as the “It’s alive!” scream, the character design, and the conclusion of the film, which is completely different to the novel, but has been parodied multiple times. There were some genuinely shocking and original scenes for the time, which caught me off-guard even today.

While not as thought-provoking, it is definitely worth a watch, just don’t expect it to be like the original at all.

Speaking of which, let’s get to it!

I have love in me the likes of which you can scarcely imagine and rage the likes of which you would not believe. If I cannot satisfy the one, I will indulge the other.

The novel!

There have been many influential books that I’ve read before, be it Dune, Roadside Picnic or The House on the Borderland. However, if we are talking about the cultural impact a book has had, it turns out that this one is on a whole different level.

There are plenty of classic works that have been bested by successors—sometimes the only thing of note is the fact that they did it first. But after two centuries, Frankenstein remains at the forefront of the genre.

At eighteen years old, Mary Wollstonecraft Shelley was part of a contest to make a spooky story. The participants were her husband Percy Shelley, Lord Byron and John Polidori. Only the latter would end up completing the challenge as well, with The Vampyre, a work that kickstarted the vampire genre, but like many others, became a stepping stone, a prototype for more known works like Dracula.

The story should be known by everyone by cultural osmosis at this point. Victor Frankenstein has quite the life ahead of him, a good family and promising studies. However, he ends up obsessed with the secret to creating life, and once he discovers it, he decides to play god. He builds a living, intelligent being, ignoring his friends, family and future during the whole endeavor. However, as soon as his creation comes to life, Victor is horrified by it, considering it wretched and vile.

The creature is alone on this world, and rejected by his maker, wanders off. He slowly learns about the world, about humanity, about language and culture. But he realizes that he’ll never be accepted, but hated and feared by any person he meets, so he declares himself an enemy of humanity, and swears revenge on his creator.

The story is a tragedy, a series of events that unfolds and ends in misery for everyone involved. A cautionary tale that resonates today.

I have to say that the writing of this novel is genuinely beautiful. There are so many interesting dialogues and narrations that were truly pleasurable to read, it was rather simple to follow, but it felt profound nonetheless.

The novel does have a weird structure, it’s a bunch of letter that contain a narrative, and another narrative is told within that. There is tragedy after tragedy, and each one feels fated while also completely avoidable.

The dialogues and the events are easy to follow, but but there is that lingering feeling about how hopeless it is. Victor Frankenstein can’t help but doom himself further, assuming the worst and falling victim to his own dumb choices.

I genuinely could’t stop myself from reading this. It took me less than a week, and I read more than half the book in a single day. It is delightful, but I also have to say that there were a few words I had to look up to understand, and some of those long sentences that I had to read with an old times radio voice to get through. I didn’t need to read a Spanish translation at all though, so that’s cool.

Victor Frankenstein considered his creation as perfect up until the moment it was brought to life. Mary Shelley’s creation may not be perfect, but it is very close, a cultural icon with a message that rings true in today’s day and age.

The fallen angel becomes a malignant devil. Yet even that enemy of God and man had friends and associates in his desolation; I am alone.

This is day 69 of #100DaysToOffload

Reply to this post via email | Reply on Fediverse

Note published on November 11, 2026 at 2:44 PM UTC - Molly White's activity feed

2026-11-11T14:44:10.000Z

i mean, just impeachment would've been okay with me... seems a little extreme