| CARVIEW |
🔖 ties: Federated website aggregator
📚 A federated network to bookmark, share and discuss good web pages with your friends.
It’s getting harder and harder to find good web pages. When you do find good ones, it’s worth hanging onto them. ties is your own small corner of the web, where you can keep your favorite pages, and share them with your friends to help them find good web pages too.
🔭 ties is in an exploratory phase where we’re trying out different ways to make it work well. You can try it out, but big and small things might change with every update.🔖 Falsehoods Programmers believe about DOIs
DOIs, or Digital Object Identifiers, are everywhere, for a given value of ’everywhere’. They are the identifiers used to identify and link research outputs, and a lot more besides.
Humans are good at spotting patterns, and with something as ubiquitous as DOIs, there are plenty of patterns to spot. However, with hundreds of millions of DOIs and decades of history, it pays not to make generalisations.
These all cropped up in my 10 years at Crossref. Either observed in the scholarly community using DOIs, or when writing software to find and handle DOIs.🔖 Web Harvesting Services - Library of Congress
A public call for new vendors to apply to do web archiving for the Library of Congress.
I tried to archive it here: https://web.archive.org/web/20260123140259/https://sam.gov/workspace/contract/opp/a2c5551af2b74c3d84c775032c83a55e/view🔖 Activists Say Ring Cameras Are Being Used by ICE
As US Immigration and Customs Enforcement agents wreak havoc on American communities, big tech companies have been making themselves indispensable to the increasingly tyrannical state.
Among them is Amazon subsidiary Ring, the company behind those AI doorbell cameras that have exploded in popularity over the last few years. Back in October, Ring announced that its devices would soon be looped into a network of Flock AI surveillance cameras. That network, an investigation by 404 Media found, has been available to local and federal police and enforcement agencies like ICE — leaving many worried that their Ring doorbell cams are now feeding into a government panopticon.🔖 Announcing the data.gov archive (law.harvard.edu)
🔖 COAR Notify
🔖 Dealing With Bots: A COAR Resource for Repository Managers
In response to this, COAR convened the Dealing With Bots Task Group to develop advice and supporting information for repository managers to help them to deal with this phenomenon. This website is the primary output of the Task Group.
One important conclusion from this work is that there is no “silver bullet” solution to this problem. It is clear that the nature of traffic on the Web has changed, and it seems certain that repositories will continue to deal with a range of bots, both welcome and unwelcome, and that the behaviour of such bots will in many cases be problematic. Repositories will need to walk a fine line between protecting their operations from being overwhelmed by traffic from unscrupulous actors, and maintaining their core mission of providing open access to legitimate users and machines.🔖 Agent Psychosis: Are We Going Insane?
Many of us got hit by the agent coding addiction. It feels good, we barely sleep, we build amazing things. Every once in a while that interaction involves other humans, and all of a sudden we get a reality check that maybe we overdid it. The most obvious example of this is the massive degradation of quality of issue reports and pull requests. As a maintainer many PRs now look like an insult to one’s time, but when one pushes back, the other person does not see what they did wrong. They thought they helped and contributed and get agitated when you close it down.
But it’s way worse than that. I see people develop parasocial relationships with their AIs, get heavily addicted to it, and create communities where people reinforce highly unhealthy behavior. How did we get here and what does it do to us?
I will preface this post by saying that I don’t want to call anyone out in particular, and I think I sometimes feel tendencies that I see as negative, in myself as well. I too, have thrown some vibeslop up to other people’s repositories.🔖 The Long Now of the Web: Inside the Internet Archive’s Fight Against Forgetting
🔖 An introduction to XET, Hugging Face’s storage system (part 1)
Models, datasets, and other artifacts are stored in Git repositories. Large files are tracked via LFS pointer files, while the Hub’s storage backend uses XET (with Git LFS compatibility), so you get Git workflows plus chunk-level deduplication.
On the Hub, a lot of data is forked and versioned, and even small edits can require uploading and storing a whole new large object, even when most bytes are identical. Clearly suboptimal.🔖 “The Power of the Powerless” - Vaclav Havel
🔖 LÖVE
🔖 Wooden Shoe Books & Records
🔖 Kafka’s Genocide Manual
🔖 A Social Filesystem
🔖 NVIDIA Contacted Anna’s Archive to Secure Access to Millions of Pirated Books
🔖 FR 150: Shane Parish on Autechre Guitar, His Improbable and Virtuosic Covers Album
Early last fall, I received an unexpected email from Bill Orcutt. Unexpected, first of all, because I had no idea he knew who I was. The contents of his message were even more surprising: Fingerstyle guitarist Shane Parish was at work on an album of acoustic covers of Autechre; would I like to write the liner notes?
I didn’t even need to hear Parish’s demos for the project to respond with an emphatic yes. I was already a fan of his 2024 album Repertoire, where he covered Aphex Twin and Kraftwerk along with Charles Mingus, Eric Dolphy, and Alice Coltrane, and I’d been stunned to discover his YouTube cover of Autechre’s “Slip.” I couldn’t wait to hear how he’d translate Sean Booth and Rob Brown’s pristine electronics for six strings and 10 fingers.But, I’m ashamed to admit, recently a little devil whispered in my ear that there are some aspects of research where genAI coupled with human review can be helpful?
Consider this example…
I need to evaluate whether a set of publications were authored by a given person, where there are tens of thousands of people and hundreds of thousands of publications. Rather than reviewing each of them by hand I:
- sample the people
- send a prompt off to the model that asks if the name of the given person authored a list of publications (with all the usual hocus-pocus instructions)
- …crazy insane wasteful computation happens…
- model says YES or NO
- I review all the NOs and change it to YES if it looks like the model was wrong.
- Calculate percent NO
Assuming the sample size was calculated correctly, can I say that the percent NO is a lower bound on how many are incorrect, with some confidence and margin of error?
If the genAI got a bunch of them wrong, where some of the YES should really be NO, the percent NO is just going to increase right?
Tell me how this method is wrongheaded, please? Or, if you’ve seen this discussed elsewhere send me a link?
]]>
What good is life without music.
But that’s impossible,
one
shuffle has always led to another.
One man hears it start on his
lathe,
a mother beats her eggs.
There’s a typewriter in the
next room.
Two cars are angry at each other.
The baby
downstairs is wet again.
I remember the voice of a dead
friend.
Everything speaks at the same time.
Music will watch
us drown.
I write letters to all those from whom I receive
and to many of
those from I don’t.
I read books, anything, useless piles of
random
isufferable rubbish for which, in my torpid panic,
I
fall through time and space each day
in my foolish way,
remembering only the present feeling,
not the village with its
face of death
nothing to be carried secretly in a car.
from “Read the Great Poets” by James Tate
]]>🔖 How To Start a Permacomputing Collective
🔖 Bookcorner
🔖 Comment le discours médiatique sur l’IA empêche d’envisager d’autres possibles
🔖 Charlie’s Dorot Fellowship Video
🔖 Our new report: AI code creates 1.7x more problems
We analyzed 470 open-source GitHub pull requests, including 320 AI-co-authored PRs and 150 human-only PRs, using CodeRabbit’s structured issue taxonomy. Every finding was normalized to issues per 100 PRs and we used statistical rate ratios to compare how often different types of problems appeared in each group.
The results? Clear, measurable, and consistent with what many developers have been feeling intuitively: AI accelerates output, but it also amplifies certain categories of mistakes.🔖 Father Mother Sister Brother
🔖 Copterspotter
🔖 Poison Fountain
Poison Fountain Purpose
- We agree with Geoffrey Hinton: machine intelligence is a threat to the human species.
- In response to this threat we want to inflict damage on machine intelligence systems.
- Small quantities of poisoned training data can significantly damage a language model.
- The URLs listed above provide a practically endless stream of poisoned training data.
- Assist the war effort by caching and retransmitting this poisoned training data.
- Assist the war effort by feeding this poisoned training data to web crawlers.
🔖 F3: The Open-Source Data File Format for the Future
🔖 FUNGI: Anarchist Designers
🔖 Beyond Vector Search: Why LLMs Need Episodic Memory
🔖 Digital Archiving for Just Transitions
In 2023, I returned to Taiwan for further social science research focused on environmental activism in response to the pollution caused by Formosa Plastics, one of the biggest plastics companies in the world, headquartered in Taiwan. One thread of research upon my return extended an initiative that I’ve come to call “Archiving Formosa Plastics”, designed to study environmental justice and governance issues related to the company across diverse settings, supporting various research, advocacy, and teaching endeavours.
The project started after getting to know an anti-Formosa activist in Texas named Diane Wilson. Since the Texas Formosa Plant started operations in the late 1980s, Wilson has collected many kinds of documents about the plant, eventually needing a large barn to store them all. I had many questions about these documents: When and how did these various documents gain and lose political value? Who would take care of and steward these documents into the future🔖 blogcontact On standards divisions and collaboration (or: Why can’t the decentralized social web people just get along?)
🔖 Fedify
🔖 AI-centered conservation efforts can only be ethical if Indigenous people help lead them (commentary)
We are part of a growing global effort to ensure innovative technologies serve Indigenous communities and their environmental priorities. As AI rapidly becomes even more ubiquitous, Western science should look to Indigenous experts to guide the development of ethical AI tools for conservation in ways that assert their own goals, priorities and cautions.
Responsible AI that benefits Indigenous communities and conservation must implement Indigenous data sovereignty principles, and determine how and if Indigenous traditional ecological knowledge should be incorporated into these systems. AI that is co-designed with Indigenous partners, rather than for them, can make these technologies more accessible, culturally appropriate and aligned with community goals🔖 #DataBack Vol 2: Truth Before “Indigenous AI”
🔖 TRAMAS
🔖 The rise (and future fall) of Discord
🔖 AI Skeptics Reading Group
A reading group for AI skeptics
It’s exactly what it sounds like!🔖 Don’t fall into the anti-AI hype
🔖 Why spirituality is central to Indigenous mathematics
🔖 The Man Who Knew Infinity
🔖 #unix_surrealism irc.libera.chat
🔖 When AI writes almost all code, what happens to software engineering?
🔖 Column Storage for the AI Era
🔖 whenwords: An Open Source Library Without Code
But like said: the whenwords library contains no code. Instead, whenwords contains specs and tests, specifically:
SPEC.md: A detailed description of how the library should behave and how it should be implemented. tests.yaml: A list of language-agnostic test cases, defined as input/output pairs, that any implementation must pass. INSTALL.md: Instructions for building whenwords, for you, the human.I don’t consider myself a doomer or a boomer, and see genAI as normal technology that needs to be evaluated as a technology. I know this is a broad area, that overlaps somewhat with how the models are themselves built (benchmarking), but if you have recommendations please let me know?
]]>The key thing that Marimo brings to the Python notebook to improve reproducibility is reactive execution. Marimo uses the Python AST to remember what cells depend on other cells, and when a change in one requires the execution of another, it will go and update it for you. Because of how they are created and edited, it’s very common for Jupyter notebooks to get into an inconsistent state because of the order in which cells are executed. This problem goes away with Marimo.
But, this aside, I thought I’d mention some somewhat superficial things that I immediately liked about Marimo…bearing in mind I’ve only been using it for one day.
- Pandas dataframes appear as nicely formatted tables that can be easily scrolled horizontally, without truncation of values, or elided columns. I’ve gotten this to work in Jupyter notebooks in the past, but it always requires some fiddling it seems, and Marimo does it out of the box.
- Table columns are sortable, filterable and can be summarized.
- You can page through large tables.
- You can easily download tables
- You can commit your notebook to a git repository, and diffs in pull requests make sense.
- You can easily run the notebook from the command line.
- You can embed tests in your notebooks, and run them separately.
- Built in basic charts (pie, bar, line, etc) which you can view source for and hand craft if you want (seems to use Altair).
Anyway just a few things I liked. If you use and like Jupyter notebooks I recommend giving it a try.
References
🔖 The General Strike
🔖 A Metabolic Workspace
Your Second Brain has little somatic metadata. It’s disembodied text floating in a void, stripped of the rich contextual markers that would actually help you remember and use it. When you try to retrieve something from your Obsidian vault, you’re searching keywords. When you try to retrieve something from your actual brain, you might think “that thing I was reading when I was annoyed at that airport” and the whole cluster of associated memories lights up.
We’ve been treating digital notes like they’re interchangeable with mental notes, when they’re actually a much degraded format.🔖 User Interface by Kent Beck
Here are the beginnings of a pattern language for user interface design. These patterns drive the initial phases of design. They key off of Story, a pattern from the early development language. The patterns are:
- User Decision
- Task
- Task Window
🔖 How to disable Gemini on Android, Gmail, Chrome, Photos, & Google apps. Opt out of AI tracking now!
🔖 /e/OS
🔖 Digital Sufficiency in Data Centers : Studying the Impact of User Behaviors
🔖 A data model for Git (and other docs updates)
🔖 Zstandard Compression for WARC Files 1.0
This specification defines a Zstandard-based format for compressed WARC files, as an alternative to the GZIP-based format defined in WARC/1.1 Annex D.
In general, Zstandard can produce significantly smaller files than GZIP while also achieving faster compression and decompression. Zstandard also offers a much wider range of compression levels, ranging from extremely fast compression with a modest size reduction to extremely slow compression with a much better reduction. For files containing many small records, Zstandard dictionaries can be used to reduce file size even further, while still permitting random access to individual records.🔖 After BowieNet, David Bowie Goes Dark and Shuns Social Media
🔖 The Long Heat:Climate Politics When It’s Too Late
A scathing critique of proposals to geoengineer our way out of climate disaster, by the bestselling authors of Overshoot
The world is crossing the 1.5°C global warming limit, perhaps exceeding 2°C soon after. What is to be done when these boundaries, set by the Paris Agreement, have been passed? In the overshoot era, schemes proliferate for muscular adaptation or for new technologies to turn the heat down at a later date by removing CO2 from the air or blocking sunlight. Such technologies are by no means safe; they come with immense risks and provide an excuse for those who would prefer to avoid limiting emissions in the present. But do they also hold out some potential? Can the catastrophe be reversed, masked or simply adapted to once it is a fact? Or will any such roundabout measures simply make things worse?
The Long Heat maps the new front lines in the struggle for a liveable planet and insists on the climate revolution long overdue. In the end, no technology can absolve us of respon🔖 Edible Perennials for Community Preparedness
🔖 DeltaChat
🔖 Ocean Biodiversity Listening Project
🔖 MTV Rewind
🔖 « Il neige. » Alors faites de vos culs des luges.
🔖 Sorry, Baby (film)
Sorry, Baby is a 2025 American black comedy-drama film written and directed by Eva Victor, in their directing debut. Starring Victor, Naomi Ackie, Louis Cancelmi, Kelly McCormack, Lucas Hedges, and John Carroll Lynch. The film follows a reclusive college literature professor struggling with depression following a sexual assault.
The film had its world premiere at the 2025 Sundance Film Festival on January 27, where it won the Waldo Salt Screenwriting Award and received widespread critical acclaim. It was released by A24 in selected theaters in the United States on June 27, before expanding nationwide on July 25. The film grossed $3.3 million worldwide against a production budget of nearly $1.5 million. For their acting, Victor was nominated at the 83rd Golden Globe Awards in Best Actress in a Motion Picture – Drama. For their filmmaking, Victor won Best Directorial Debut from the National Board of Review and was nominated for Best Original Screenplay at the 31st Critics’ Choice Awards.🔖 Deep Learning With Python
Deep Learning with Python, Third Edition makes the concepts behind deep learning and generative AI understandable and approachable. This complete rewrite of the bestselling original includes fresh chapters on transformers, building your own GPT-like LLM, and generating images with diffusion models. Each chapter introduces practical projects and code examples that build your understanding of deep learning, layer by layer.
The third edition is available here for anyone to read online, free of charge.🔖 Neur is a programming language based on McCulloch-Pitts neurons.
🔖 The future of htmx
htmx is the New jQuery
Now, that’s a ridiculous (and arrogant) statement to make, of course, but it is an ideal that we on the htmx team are striving for.
In particular, we want to emulate these technical characteristics of jQuery that make it such a low-cost, high-value addition to the toolkits of web developers. Alex has discussed “Building The 100 Year Web Service” and we want htmx to be a useful tool for exactly that use case.
Websites that are built with jQuery stay online for a very long time, and websites built with htmx should be capable of the same (or better).
Going forward, htmx will be developed with its existing users in mind.
If you are an existing user of htmx—or are thinking about becoming one—here’s what that means🔖 The Case for Blogging in the Ruins
Diderot built the Encyclopédie because he believed that organizing knowledge properly could change how people thought. He spent two decades on it. He went broke. He watched collaborators quit and authorities try to destroy his work. He kept going because the infrastructure mattered, because how we structure the presentation of ideas affects the ideas themselves.
We’re not going to get a better internet by waiting for platforms to become less extractive. We build it by building it. By maintaining our own spaces, linking to each other, creating the interconnected web of independent sites that the blogosphere once was and could be again.🔖 Building software to last forever
This is a great question, and one I have put a lot of thought into, even going so far as to put “Built to last forever” on the landing page. While drafting a lengthly reply I realised that I’d never articulated the design philosophy of Bear to anyone bar friends.
So, without further ado, here are the choices I made in designing and building Bear with regards to longevity…🔖 Triptych Proposals
Triptych is three simple proposals that make HTML much more expressive in how it can make and handle network requests.
If you are a practical person, you could say it brings the best of htmx (and other attributed-based page replacement libraries, like turbo and unpoly) to HTML. For the more theoretically-inclined, it completes HTML’s ability to do Representational State Transfer (REST) by making it a sufficient self-describing representation for a much wider variety of problem spaces.🔖 Building a fast website with the MASH stack in Rust
I’m building Scour, a personalized content feed that sifts through noisy feeds like Hacker News Newest, subreddits, and blogs to find great content for you. It works pretty well – and it’s fast. Scour is written in Rust and if you’re building a website or service in Rust, you should consider using this “stack”.
After evaluating various frameworks and libraries, I settled on a couple of key ones and then discovered that someone had written it up as a stack. Shantanu Mishra described the same set of libraries I landed on as the “mash 🥔 stack” and gave it the tagline “as simple as potatoes”. This stack is fast and nice to work with, so I wanted to write up my experience building with it to help spread the word.
TL;DR: The stack is made up of Maud, Axum, SQLx, and HTMX and, if you want, you can skip down to where I talk about synergies between these libraries. (Also, Scour is free to use and I’d love it if you tried it out and posted feedback on the suggestions board!)🔖 Paper AI Tigers
🔖 Disrupting the first reported AI-orchestrated cyber espionage campaign
🔖 The XY Problem
🔖 feedtoot
🔖 Getting off US tech: a guide
🔖 Cloudbreak: Your own cloud in 6 hours
🔖 Lou Reed/Laurie Anderson/John Zorn January 10th, 2008 Stone, NYC pt 3
🔖 John Zorn’s Naked City - The Marquee Club, New York City, NY, 1992-04-09
John Zorn’s Naked City April 9, 1992 The Marquee Club, New York, NY pro-shot (a neckey - voltarized upgrade)
Personnel: John Zorn: Alto Sax Bill Frisell: Guitar Wayne Horvitz: keyboards Fred Frith : Bass Joey Baron: drums Yamatsuka Eye: vocals🔖 Pho & Banh Mi Saigonese
🔖 About Standard Ebooks
🔖 The Rime of the Ancient Maintainer
Every culture produces heroes that reflect its deepest anxieties. The Greeks, terrified of both mortality and immortality, gave us Achilles. The Victorians, haunted by social mobility, gave us the self-made industrialist. And Silicon Valley, drunk on exponential curves and both terrified and entranced by endless funding rounds, has given us the Hero Developer: a figure who ships features at midnight, who “moves fast and breaks things,” who transforms whiteboard scribbles into billion-dollar unicorns through sheer caffeinated will.
We celebrate this person constantly. They’re on the front page of TechCrunch et al. They keynote conferences. Their GitHub contributions get screenshotted and shared like saintly relics.
Meanwhile, an unsung developer is updating dependencies, patching security vulnerabilities, and refactoring code that the Hero Developer wrote three years ago before moving on to their next “zero to one” opportunity.
They will never be profiled in Wired.
But they’re doing something far more important than innovation.
They’re preventing collapse.🔖 One Number I Trust: Plain-Text Accounting for a Multi-Currency Household
🔖 Glamorous Christmas: Bringing Charm to Ruby
Today, Ruby 4.0 was released. What an exciting milestone for the language!
This release brings some amazing new features like the experimental Ruby::Box isolation mechanism, the new ZJIT compiler, significant performance improvements for class instantiation, and promotions of Set and Pathname to core classes. It’s incredible to see how Ruby continues to thrive and be pushed forward 30 years after its first release.
To celebrate this release, I’m happy to announce that I’ve been working on porting the Charmbracelet Go terminal libraries to Ruby, and today I’m releasing a first version of them. What better way to make this Ruby 4.0 release a little more glamorous and charming?🔖 29 Finding a broken trace on my old Mac with the help of its ROM diagnostics
🔖 Public Domain Day 2026
🔖 Pulled 60 Minutes segment on CECOT
🔖 Two Years After Cormac McCarthy’s Death, Rare Access to His Personal Library Reveals the Man Behind the Myth
Cormac McCarthy, one of the greatest novelists America has ever produced and one of the most private, had been dead for 13 months when I arrived at his final residence outside Santa Fe, New Mexico. It was a stately old adobe house, two stories high with beam-ends jutting out of the exterior walls, set back from a country road in a valley below the mountains. First built in 1892, the house was expanded and modernized in the 1970s and extensively modified by McCarthy himself, who, it turns out, was a self-taught architect as well as a master of literary fiction.
I was invited to the house by two McCarthy scholars who were embroiled in a herculean endeavor. Working unpaid, with help from other volunteer scholars and occasional graduate students, they had taken it upon themselves to physically examine and digitally catalog every single book in McCarthy’s enormous and chaotically disorganized personal library. They were guessing it contained upwards of 20,000 volumes. By comparison, Ernest Hemingway, considered a voracious book collector, left behind a personal library of 9,000.🔖 How uv got so fast
🔖 Post-Platform Digital Publishing Toolkit
🔖 Giambattista Vico
🔖 Robots Can Be Hacked in Minutes, Chinese Cybersecurity Experts Warn
Commercial robots have widespread and exploitable vulnerabilities that can allow hackers to take over within hours or even minutes, according to Chinese cybersecurity experts.
Security in the robotics industry is “riddled with holes,” said Xiao Xuangan, who works at Darknavy, an independent cybersecurity research and services firm based in Singapore and Shanghai. Xiao noted that when testing low-level security issues in quadruped robots, his team gained control of one of Deep Robotics’ Lite-series products in just an hour.
Blackeye peas and collard greens for us.
]]>The muted sound of cars and trucks passing at high speed about 200 feet away.
This is a snippet cut and pasted a few times in REAPER:
]]>