Search Results: "anarcat"

24 September 2026

Russell Coker: Links September 2026

Bjorn Stahl of the Arcan project wrote an insightful blog post The Day of a new Command-Line Interface: Shell [1]. He does a really good job of identifying problems and strategies for dealing with them, the UI of the results isn t suitable to what I do though. Using sqlite as an executable format (replacement for ELF) is one of the wildest ideas I ve seen in a while, it might actually have some good use cases too [2]. Andrew Dana Hudson wrote an interesting LongNow article about SciFi s focus on space travel when it seems unlikely [3]. TheDriven has an informative article about the Hands Off Our Fuel astroturf campaign that uses farmers as the face of an attempt to give tax exemptions to mining companies [4]. The Conversation has an interesting article about the new Chinese regulations restricting access to AI systems to prevent emotional dependence and prevent minors from having virtual partners or relatives [5]. Elana Hashman wrote an informative blog post about Python virtualenvs, has useful information for people who aren t Python experts and just need to get stuff done [6]. We need more systems in the NTP pool, I would consider doing this if I had a server with an IP address that won t change and no risk of bandwidth throttling [7]. Cory Doctorow has some interesting thoughts on the CEO fetish for AI even though it is almost never doing any good [8]. Steve Rose wrote an insightful article for The Guardian about the tech fascists of Silicon Valley [9]. Jacqui Lambie calls Pauline Hanson an angry little woman and a bloody coward , One Nation hate the rule of law [10]. The Conversation has an interesting article about Regime Change, the book about Trump by Maggie Haberman and Jonathan Swan [11]. Elvira Barry wrote an insightful article about the many failures of the Russian legal system, it s interesting to note that ALL those failures match things that the Trump regime is doing [12]. Pyotr Kurzin Geopolitics YouTube channel has an interesting interview about continental vs maritime powers in the context of what the US and Russia are doing [13]. Maxinomica has an interesting video about how China obtained a monopoly on rare earth metals and what happens next [14]. Anarcat wrote an insightful blog post The people vs the AI overlords about the problems that AI software development is bringing [15]. John Goerzen wrote an interesting blog post discussing the issues related to LLM use in FOSS projects [16]. Cybernews has an article about an exploit for Samsung, Xiaomi, Oppo, and OnePlus phones due to bad OEM kernel drivers and bad OEM API proxies, no company should release devices with proprietry drivers [17]. Scott Santens wrote an insightful article about how retraining fails laid-off workers and why UBI is needed when workers get replaced by LLMs [18]. libroot has some information on the Snowden archive and why so little of it has been used by journalists, we need more answers about this [19]. Drew Devault s list of Weird little guys of FOSS is interesting history about some people who are best avoided [20].
Related posts:
  1. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  2. Links February 2026 Charles Stross has a good theory of why AI is...
  3. Links September 2024 CNA Insider has an insightful documentary series about Chinese illegal...

25 August 2026

Antoine Beaupr : A more nuanced view of LLMs

Also in this series:
After ranting and railing about LLMs or "AI" as the optimists (or accelerationists?) call it, I figured it might be important to be a little more honest about my use of LLMs and how I think about it more practically in the world.

The Debian vote context This is not a coming out. I am not using LLMs on a daily basis, and this blog is, again, written out of my cold dead hands in a dying world, with over-engineered hardware and (to a certain extent, hi Emacs!) software, powered by 100% green energy built on stolen land. There is a vote going on in Debian. If you're unfamiliar with it, you can catch up at LWN. So far I've essentially said "LLM is bad" which is not a very balanced or useful opinion. Obviously, people are using LLMs, sometimes unknowing or unwillingly, and we need to take that into account. Furthermore, there has been many different blog posts on Debian planet about this. Some that I found balanced, good summaries, even if I didn't fully agree with them, at least some did the basic civil service of being short. But others were just not only Wrong but also so long that I couldn't finish that I just had to write something.1 This is not an explanation of the ballots, nor how I will vote. This vote is Debian's failure of framing that debate in a reasonable way: we have 8 options on the ballot with many duplicates. We have failed to do the hard work of summarizing and aggregating options into a meaningful set. I doubt the final vote will represent a readable position we can rally around. I have not read the two months of debates on the topic either. Normally, before voting, I take a cursory look at the debate to see points of view I might have missed. But in this case, it will just make me sad, add noise, and I'm already pretty sure on where I stand on this. So let me describe how I use LLMs and how I think they fit in our work, as computer engineers and hobbyists.

My LLM use

Debian Packaging An astute reader has pointed out that I maintain a package in Debian made to use Anthropic. It's actually multiple packages: As I previously explained in response, I am not entirely comfortable with this work: it's a compromise. In fact, I first uploaded llm to the contrib section of Debian, where we keep software that depends on other non-free software, but I was told that, since yt-dlp was in main, llm belonged there as well. So I moved it to main, alongside similarly controversial tools like llama.cpp or the python-openai library.

OpenAI and Anthropic usage An important part of my work is technology watch. I keep tabs on thousands of (new and old) software projects, follow news, and generally try to keep my skills up to date. It's a pretty impossible race, especially as I grow older, but I still think I'm doing the right choices in my job. Testing large language models is part of that work. At first, I was using ChatGPT's web interface, but it was annoying to copy-paste things into a browser, so I looked for different interfaces. For a while I tried gptel, a "simple, extensible LLM client for Emacs" but I found it kind of terrifying. Giving a LLM control over an Emacs buffer seems like a security nightmare, so I stopped doing that. So I use the llm command-line tool to talk to Anthropic's API. I started that in the summer of 2025, when I bought 20$USD of API credits. Before that, I paid for a ChatGPT subscription and then OpenAI credits, which expired and sent me over to Anthropic, which seemed then to have better ethics. As it turns out, Anthropic is also happy to work for the US military (which is a big red line for me). Anthropic also won't let you talk about the genocide in Gaza, it is destroying physical books, and is blackmailing us to use their product for security coverage. Needless to say, Anthropic and "Claude" are not my friends, but they seem like the lesser evil in current "frontier models". So I have renewed, a couple of weeks ago, another 20$USD of API credits with Anthropic.

Actual prompts and responses So what does 20$ give you at Anthropic anyways? What am I using LLMs for and how? The neat thing with llm is that everything is logged in a sqlite database, so there are some answers that are easy to get: 
> llm logs status
Logging is ON for all prompts
Found log database at /home/anarcat/.config/io.datasette.llm/logs.db
Number of threads logged:   7
Number of turns logged:     12
Number of legacy conversations: 543
Number of legacy responses: 970
Database file size:         9.61MB
That is 10MB of logs, with about a thousand prompts. My logs go back to 2024-03-07, a little over two years ago, and include a mix of Anthropic and OpenAI responses. I used it more in 2024 than 2025, and if the trend continues, I will have used it less in 2026 again: 
> llm logs list -n 0  --json   jq -r .[].datetime_utc   sed 's/-.*//'   sort   uniq -c 
    527 2024
    357 2025
     98 2026
It looks like about 10 prompts per month right now, down from a peak of about 60 per month in 2024. It's pretty difficult to analyze those actual logs to get more patterns and I won't run the prompts through a model again to process them.

How I'm using models now At first, I was using it partly for benchmarking model's capabilities, like Simon Willison does with his pelicans, clearly not trusting its output. But I was impressed by the capacities of the Claude Opus 4.5 model when it wrote this script in January. Impressed, but also scared: it's the first time I felt I could delegate the entirety of my programming to a model. Just run the code, if it works, it works, right? So what do I use it now? As an example, here are the 10 last prompts in my history:
  1. there is now Claude 5, and a fable model, maybe you know about it?
  2. impress me
  3. not impressive, i already know all of this
  4. chat
  5. in postfix, i have a 300k mailing that happens regularly here. normally, it delivers within about...
  6. is there a way i could have drained the maildrop queue faster without removing the milter?
  7. the problem was that rspamd was timing out on the FUZZY_CALLBACK check. how do i disable that?
  8. how do i disable all spam checks? i just want rspamd to add dkim signatures
  9. how do the default_destination_concurrency_limit and initial_destination_concurrency settings int...
  10. mic check
The first one was me trying to confirm which model I am using, which is not always obvious when going through the whole llm stack I've been using. The following two are an attempt at seeing what the model is capable of and I was "not impressed", to which Claude answered that I have a "high bar", which, fair enough. The chat is me failing to use a command line, which shows that perhaps I need to readjust that "high bar", again. The next five are a rather embarrassing debacle in a large Postfix mailing that went sideways, and where I couldn't find an actual Postfix expert of my level to help. The fabled Claude Fable 5 answered rather correctly, but dangerously, that I could empty the queue by disabling the non_smtpd_milters. What Fable (and myself) did not realize is that the milter was also adding DKIM signatures, so while the mailing was expedited, it was done without those precious signatures, which got us promptly blocked at Gmail. We have recovered since, and, thanks to the model and reading the Postfix manual for the hundredth time, that pickup(8) is single-threaded and that we needed to review the architecture of that mailing (and our spam filters) a bit. Many tickets ensued. The last one is a test I did to make sure my last uploads of llm-anthropic and its dependency worked correctly.
Note that the above excludes 5 questions I asked Anthropic while writing this article, where I asked for synonyms and "what nanometer scale are arduino processors built from? how is an arduino CPU printed?", a question which Wikipedia furiously evades providing a good answer.
Those prompts are pretty typical of my LLM use: I'm testing the models to see if they work at all, but also, out of desperation, I fire off a prompt after I fire off questions to colleagues or search engines (in that order). It's often weird edge cases like the Prometheus query language, Python's matplotlib, LaTeX, Elisp, optimizations, and so on. I use models for translation a lot. Being fully bilingual, it is common for me to think of a word in French or English and fail to find exactly the right word for that in the other language. Models help with that, and are also useful to find synonyms. Those are low-token uses that seem pretty innocuous to me, but I realize the irony of this after writing about the tower of Babel.

What I am not using models for I am not using models to write prose. I am not using models to read prose. If it's generated with LLMs, I stop reading. I am not using models to write code, with the exception of that single Python script above. I am generally not using models to review code, with exceptions. If I get stuck on a hard problem, I might feed a piece of code to the model. I repeatedly fed asncounter into Claude to try to fix a performance regression I had introduced. It found micro-optimizations that taught me a thing or two about Python's internal implementations, but overall, it was mostly a waste of time. This was in June 2025, so perhaps now models would fare better. I have not tried again. I am not using LLMs to do Debian packaging. When I can, I manually review the diffs of packages I upload into Debian, still, by hand. I do this for the reasons outlined in The Four Horsemen of the LLM Apocalypse, because I refuse to be complicit in the:
  1. aggressive and illegal scraping of the servers I steward
  2. world-wide computer hardware shortage (making it, by the way, nearly impossible to run presumably clean local models) and the attack on our job conditions (also discussed in The people vs the AI overlords)
  3. death of copyright and free software
  4. complication and enshifitication of everything, and the destruction of our communities
  5. the imperialist Nerd Reich that wants to take over the world
Like I reluctantly use Intel computers, I do fire off a prompt. But I still hold on to the dream that we can build communities of practice that hold human knowledge collectively and not offload that as a utility to some megalomaniac billionaire.

Their LLM use I am forced into So that's me. Clearly, I'm going against the grain here. Everywhere I look, I see LLM-generated code and projects. Slop and botnets have flooded the web. I use Wadamesh, clearly vibe-coded, because it's the best graphical interface for MeshCore that runs on portable devices. I wish it was made by a human, in a community I could participate in, but it isn't, and I don't. I package the above llm toolset, which is more and more vibe-coded, but I still review the diffs. And I have to say: I trust Simon here. The code is verbose as hell, feels overengineered, and llm feels slow, but it generally works, and Simon is still at the gate. The Anthropic SDK is another thing entirely. The 0.91.0 to 0.120 upload, for example, was nuts: 
 806 files changed, 72281 insertions(+), 1478 deletions(-)
I explicitly did not review that entire diff. It feels like there's a lot of garbage there to just have a shim between a proprietary API and Python. But this is the hand I've been dealt.

Larger projects LLM use LLMs are being used in the Linux kernel, Firefox, rsync, Rust, and other places. I don't feel good about this, particularly in Rust, but they at least made a decent policy. I am glad GCC made a policy against LLM contributions and I support the human Emacs project. We need to have a set of foundational tools that are "clean" in the sense that they are built upon a community of people that understand how they are built. Maybe that's naive or even impossible. The Linux kernel and GCC, in particular, are massive projects that have long grown past the scale of a single person's understanding. But the theory was that a community of humans can understand collectively. Now we seem to be throwing up our hands and giving up on that community. That LLMs will just fix the problem, whatever it is. But we're all just one rug pull away from being completely incapable of managing those projects. The argument there is that we'll just switch to local models, but no one is actually doing that. All I see is people use local models as a corner case (for privacy) or as in theory, but in reality, everyone uses the centralized frontier models right now. We just can't fallback. We're in the same situation we were, a decade or two ago, when Microsoft decided it would kill free office alternatives by making Office free for non-profits. It worked: thousands, if not millions of schools, community groups and individuals stopped looking for alternatives (including free software but also "piracy") for Office and embraced what seemed like a generous offer. Now Microsoft pulled the plug and Over 170,000 Nonprofits Lost All Their Data. I'm afraid the rug pull on LLMs will be much worse: never mind that Linus won't be able to use his tireless helper to fix obscure kernel bugs; we're looking at a collapse of the economy so large that we are already talking about bailing out the companies responsible. In a sense, the most striking thing about the Debian vote is it has actually no option to completely refuse upstream LLM contributions. It seems the community has taken it for granted that it's now impossible to build Debian entirely without LLMs. We lost the battle even without a fight, it seems.

A plea for small If it has really become impossible for us to manage the complexity we have built, maybe it's time to stop and think about what we're doing in the first place. We're struggling to even bootstrap our current toolchain! This is one of the things I like the most about working on the mesh: it's low tech, small Arduino devices that is built with decades-old semiconductor processes that is understandable by human beings. Maybe the answer lies more in single-purpose devices like those communicators and simpler multi-purpose computers than what we have now, which is what the permacomputing movement is about. Small is beautiful, let's scale it down.
  1. and yes, I'm sorry this has gotten this long, I hope you will forgive those 3000 words.

31 May 2026

Russell Coker: Links May 2026

Ron Garrett wrote an interesting blog post about the mathematical possibility of abiogenesis [1]. Cory Doctorow wrote an interesting blog post about the way the current antics of right wing extremists are forcing permanent changes in society away from the old systems [2]. William Angel wrote an insightful blog post comparing the costs of a Macbook and the Openrouter hosted service for LLMs [3]. The Register has an informative article about the threat that management systems built in to Intel and AMD CPUs pose to data sovereignty in EU owned cloud providers [4]. But this is just the first stage of building sovereign clouds, all significaant cloud services run at least 2 types of CPU and adding EU manufactured CPUs at a future time will be easy. Benn Jordan made an interesting YouTube video about the infrasound problems caused by data centers, we need FOSS to measure infrasound [5]. amarok on the Purism forum made a great post about how to setup profiles in Firefox for different uses [6]. fralb5 wrote an informative post on the Purism forum about how to use a Librem 5 (or any other FOSS Linux phone) to firewall spyware on an Android phone [7]. Michael Prokop wrote an interesting blog post about debugging input event problems on Linux which turned out to be due to an analogue headphone connection [8]. This gave me some useful pointers to investigating an input device problem which is probably very different. Patrick Boyle made an insightful youtube video about the ridiculous IPO of SpaceX, it seems like a scam from start to finish [9]. Anarcat wrote an insightful blog post about the LLM apocalypse comparing it to the horsemen of the apocalypse [10]. The Wikimedia Foundation (that runs wikipedia.org among other things) is sacking union organisers and trying to corporatise the organisation which means stealing the donations from the community [11]. Tianon Gravi wrote an informative blog post about containers, Debian, and Docker options [12]. We need a lot more work on these sorts of things in Debian. Memory Tagging and how it improves C/C++ memory safety is an interesting paper from Google researchers giving an overview of the benefits of tagged memory hardware for pointer validation on SPARC and ARM64 [13]. In 2013 a faulty beer fridge motor acted as a spark gap transmitter and blocked mobile phone service for several Melbourne suburbs [14].
Related posts:
  1. Links April 2026 Charles Stross wrote an interesting blog post about the apparent...
  2. Links February 2026 Charles Stross has a good theory of why AI is...
  3. Links March 2026 Krebs has an interesting article about the Kimwolf botnet which...

15 May 2026

Antoine Beaupr : The Four Horsemen of the LLM Apocalypse

I have been battling Large Language Models (LLM1) for the past couple of weeks and have struggled to think about what it means and how to deal with its fallout. Because the fight has come from many fronts, I've come to articulate this in terms of the Four Horsemen of the Apocalypse.
Sound track: Metallica's The Four Horsemen, preferably downloaded from Napster around 2000, but now I guess you get it on YouTube.

War: bot armies Let's start with War. We've been battling bot armies for control of our GitLab server for a while. Bots crawl virtually infinite endpoints on our Git repositories (as opposed to downloading an archive or shallow clone), including our fork of Firefox, Tor Browser, a massive repository. At first, we've tried various methods: robots.txt, blocking user agents, and finally blocking entire networks. I wrote asncounter. It worked for a while. But now, blocking entire networks doesn't work: they come back some other way, typically through shady proxy networks, which is kind of ironic considering we're essentially running the largest proxy network of the world. Out of desperation, we've forced users to use cookies when visiting our site. We haven't deployed Anubis yet, as we worry that bots have broken Anubis anyways and that it does not really defend against a well-funded attacker, something which Pretix warned against in 2025 already. (We have a whole discussion regarding those tools here.) But even that, predictably, has failed. I suspect what we consider bots are now really agents. They run full web browsers, JavaScript included, so a feeble cookie is no match for the massive bot armies.

Side note on LLM "order of battle" We often underestimate the size of that army. The cloud was huge even before LLMs, serving about two thirds of the web. Even larger swaths of clients like government and corporate databases have all moved to the cloud, in shared, but private infrastructure with massive spare capacity that is readily available to anyone who pays. LLMs have made the problem worse by dramatically expanding the capacity of the "cloud". We now have data centers that defy imagination with millions of cores, petabytes of memory, exabytes of storage. I thought that 25 gigabit residential internet in Switzerland could bring balance, but this is nothing compared to the scale of those data centers. Those companies can launch thousands, if not millions of fully functional web browsers at our servers. Computing power or bandwidth are not a limitation for them, our primitive infrastructure is. No one but hyperscalers can deal with this kind of load, and I suspect that they are also struggling, as even Google is deploying extreme mechanisms in reCAPTCHA. This is the largest attack on the internet since the Morris worm but while Robert Tappan Morris went to jail on a felony, LLM companies are celebrated as innovators and will soon be too big to fail.2 Which brings us to the second horsemen, famine.

Famine: shortages All that computing power doesn't come out of thin air: it needs massive amounts of hardware, power, and cooling. Earlier this year, I've heard from a colleague that their Dell supplier refused to even provide a quote before August. Dell! In February, Western Digital's hard drive production for 2026 was already sold out. Hard drives essentially doubled in price within a year, and some have now tripled. A server quote we had in November has now quadrupled, going from 10 thousand to FORTY thousand dollars for a single server. But regular folks are facing real-life shortages as well, as city-size data centers are being built at neck-breaking speed, stealing fresh water and energy from human beings to feed the war machine. We've been scared of losing our jobs, but it seems that Apocalypse has yet to fully materialize. Regardless for engineers, the market feels tighter than it was a couple years ago, and everyone feels on edge that they will just have to learn to operate LLMs to keep their jobs. Update: it turns out I was clearly too optimistic. Cisco is laying off 4,000 or 5% of its staff in a jolly announcement celebrating a record $15.8 billion revenue, and Meta will lay off 8,000 or 10% of its workforce, in horrifying conditions. See also the jobloss.ai tracker which counts 125,000 jobs lost since January 2025, as of May 2026. Which brings us, of course, to Death.

Death: security and copyright Our third horseman is one I did not expect a couple of months ago. Back at FOSDEM, curl's maintainer Daniel Stenberg famously complained about the poor quality of LLM-generated reports but then, a few months later, everyone is scrambling to deal with floods of good reports. In the past two weeks, this culminated in a significant number of critical security issues across multiple projects. Chained together, remote code execution vulnerabilities in Nginx and Apache and two local privilege escalations in the Linux kernel (dirtyfrag and fragnesia) essentially gave anyone root access to any unpatched server to the web. As I write this, another vulnerability dropped, which gives read access to any file to a local user, compromising TLS and SSH private keys. All those vulnerabilities were released without any significant coordination while people scrambled to mitigate. Many people including Linus Torvalds are now considering issues discovered through LLMs to be essentially public. This puts some debates about disclosure processes in perspective, to say the least. But this is not merely the death of the traditional coordinated disclosure process, the C programming language, or the Linux kernel: remember that those bots are trained on a large corpus of copyrighted material. Facebook has trained their models on pirated books and Nvidia has done deals with Anna's Archive to secure access to large swaths of copyrighted material. The US Congress seems to think LLM outputs are not copyrightable, like any other machine outputs. With many people now vibe coding their way out of learning or remembering how computers work, is this the Death of Copyright? And that, of course, brings us to the final horseman: Pestilence.

Pestilence: slop There is a growing meme that programming is essentially over as we know it. That you can simply vibe-code applications from scratch and it's pretty good. Maybe that's true. So far, most of my attempts at resolving any complex problem with a LLM have often failed with bizarre failures. Some worked surprisingly well. Maybe, of course, I am holding it wrong. I personally don't believe LLMs will ever be good enough to produce and maintain software at scale. They're surprisingly good at finding security flaws right now. But what I see is also a lot of Bullshit, with a capital B. It's not lying: it does not "know" anything, so it can't lie. It's misleadingly cohesive and deliberate, but it lacks meaning, intent, will. I have not been confronted with much slop, apart from the lobster Jesus or the yellow man atrocities, and particularly not in my work. But I see what it is doing to my profession: beyond vibe-coding, people are now token-maxxing, and land-grabbing their colleagues. I don't like what LLMs do to our communities, or the fabric of software we live with. Software does not evolve in a void. It is a team effort, be it free software or a corporate product. Generations of humans have carefully built the scaffolding of technology required for modern networks and software to operate, in a convoluted contraption that no single human fully understands anymore. The idea of simply giving up on that understanding entirely and delegating it to an unproven model is not only chilling, it feels just plain stupid. Not stupid as in Skynet, stupid as in "I can't get inside the data center because the authentication system is down". Except we're in a "the power plant doesn't reboot" or "their LLM found an 0day in our slop" kind of stupid.

The fifth horsemen Researching for this article, I looked up the four horsemen and found out they original seems to have been:
  • Famine
  • War
  • Death
  • Conquest (??)
I was surprised. I grew up thinking about the horsemen being Famine, War, Pestilence, and Death. So I went back to my original source which actually claims the horsemen are: 
Time has taken its toll on you, the lines that crack your face.
Famine, your body, it has torn through, withered in every place.
Pestilence for what you've had to endure, and what you have put others through
Death, deliverance for you, for sure, now there's nothing you can do
So I guess that makes no sense either, which, fair enough, I shouldn't rely on Metallica for theological references. Especially since that song was originally called Mechanix and was "about having sex at a gas station". Anyways. The point is, there are actually five horsemen, and the fifth one is, in my opinion, Conquest. Those companies (and not "AI", mind you) are taking over the world. I sense a strong connection with the "post-truth" world imposed on us by fascists like Trump and Putin. It's not an accident, it's a power grab part of the Californian Ideology3. Just like Airbnb broke housing, Uber destroyed the transportation and Amazon is taking over retail and server hosting, LLM companies are essentially trying to take over if not everything, at least Cognition as a whole. But the capitalization of those companies (OpenAI and Nvidia in particular) are so far beyond reason that their inevitable collapse will likely lead to a global financial collapse of biblical proportions. Because they will inevitably fail like previous bubbles they are built on. And when they fail, I hope it zips all the way back through the blockchain scam, the ad surveillance system, and the dot com then git me back my internet.

The Tower of Babel While I'm off in the woods hallucinating (ha!) on biblical allegories, I feel there's another sign that the apocalypse is coming. The Tower of Babel myth says that humans tried to create a big tower up to heaven and become god. God confounds their speech and scatters the human race. End of utopia. This is what is happening to our human translators now. LLMs being, after all, Language Models, they are excellent at translation work. So much that the only translators not replaced by LLMs right now are interpreters, who translate vocally in real time. But interpreters are worried about their jobs as well. This concretely means we will lose the human capacity, as a civilization, to translate between each other. It is still an open question whether the remaining revision work will be enough for translators to avoid deskilling, but other research has shown that LLM use leads to cognitive decline, impacts critical thinking, and generally, that deskilling is a common outcome. Ultimately, I think this is where LLMs bring us. Towards collapse. So this is a call to arms. Fight back! Poison bots. Build local real-world communities. Go low tech. Moore's law is dead, make use of it. Patch your shit. Go weird. Refuse slop. Train your brain. Refuse distillation. The horsemen will collapse, but let's not go down with them. Butlerian Jihad!
This article was written without the use of a large language model and should not be used to train one.

Updates
  • A paragraph was added about the job apocalypse, which was of course under-estimate.
  • Why Timnit Gebru was fired is extremely important and interesting. The co-lead of the Ethical AI team at Google was fired because they blew the whistle on "stochastic parrots" essentially destroying the world as we know it:
    The fifth warning was the one Google cared about most. [...] The internet would become a place where the dominant voice was a statistical average of dominant voices, presented as a neutral assistant.
    The warnings from the paper are eerily similar to my horsemen:
    1. predicted the hallucination (pestilence)
    2. bias amplification (war?)
    3. environmental cost (famine)
    4. un-auditable training corpus (death?)
    5. "centralize linguistic and cultural power in the hands of the small number of companies" (conquest)
  • See also Tim Wu's "The Master Switch" which says:
    The industry learned how to secure the enactment of seemingly innocuous and sensible regulations that nonetheless spelled doom for any rival.
    People claim the same about Anthropic.
  1. I prefer "LLM" to Artificial Intelligence, as I don't consider models to have "Intelligence" which goes far beyond the analytical traits we train models for. Intelligence requires embodiment and social interaction; machines lack the innate human skills of empathy, feeling and care, which explains a lot of the evils behind the current trends.
  2. It should be noted that Morris also happened to be one of the founder of Y Combinator where he is in good company with other techno-fascists like Peter Thiel, Sam Altman, and so on. Crime, after all, pays.
  3. Probably a good time to watch All Watched Over by Machines of Loving Grace.

6 March 2026

Antoine Beaupr : Wallabako retirement and Readeck adoption

Today I have made the tough decision of retiring the Wallabako project. I have rolled out a final (and trivial) 1.8.0 release which fixes the uninstall procedure and rolls out a bunch of dependency updates.

Why? The main reason why I'm retiring Wallabako is that I have completely stopped using it. It's not the first time: for a while, I wasn't reading Wallabag articles on my Kobo anymore. But I had started working on it again about four years ago. Wallabako itself is about to turn 10 years old. This time, I stopped using Wallabako because there's simply something better out there. I have switched away from Wallabag to Readeck! And I'm also tired of maintaining "modern" software. Most of the recent commits on Wallabako are from renovate-bot. This feels futile and pointless. I guess it must be done at some point, but it also feels we went wrong somewhere there. Maybe Filippo Valsorda is right and one should turn dependabot off. I did consider porting Wallabako to Readeck for a while, but there's a perfectly fine Koreader plugin that I've been pretty happy to use. I was worried it would be slow (because the Wallabag plugin is slow), but it turns out that Readeck is fast enough that this doesn't matter.

Moving from Wallabag to Readeck Readeck is pretty fantastic: it's fast, it's lightweight, everything Just Works. All sorts of concerns I had with Wallabag are just gone: questionable authentication, questionable API, weird bugs, mostly gone. I am still looking for multiple tags filtering but I have a much better feeling about Readeck than Wallabag: it's written in Golang and under active development. In any case, I don't want to throw shade at the Wallabag folks either. They did solve most of the issues I raised with them and even accepted my pull request. They have helped me collect thousands of articles for a long time! It's just time to move on. The migration from Wallabag was impressively simple. The importer is well-tuned, fast, and just works. I wrote about the import in this issue, but it took about 20 minutes to import essentially all articles, and another 5 hours to refresh all the contents. There are minor issues with Readeck which I have filed (after asking!): But overall I'm happy and impressed with the result. I'm also both happy and sad at letting go of my first (and only, so far) Golang project. I loved writing in Go: it's a clean language, fast to learn, and a beauty to write parallel code in (at the cost of a rather obscure runtime). It would have been much harder to write this in Python, but my experience in Golang helped me think about how to write more parallel code in Python, which is kind of cool. The GitLab project will remain publicly accessible, but archived, for the foreseeable future. If you're interested in taking over stewardship for this project, contact me. Thanks Wallabag folks, it was a great ride!

18 February 2026

Antoine Beaupr : net-tools to iproute cheat sheet

This is also known as: "ifconfig is not installed by default anymore, how do I do this only with the ip command?" I have been slowly training my brain to use the new commands but I sometimes forget some. So, here's a couple of equivalence from the old package to net-tools the new iproute2, about 10 years late:
net-tools iproute2 shorter form what it does
arp -an ip neighbor ip n
ifconfig ip address ip a show current IP address
ifconfig ip link ip l show link stats (up/down/packet counts)
route ip route ip r show or modify the routing table
route add default GATEWAY ip route add default via GATEWAY ip r a default via GATEWAY add default route to GATEWAY
route del ROUTE ip route del ROUTE ip r d ROUTE remove ROUTE (e.g. default)
netstat -anpe ss --all --numeric --processes --extended ss -anpe list listening processes, less pretty
Note that I wrote a whole article about the latter.

Another trick Also note that I often alias ip to ip -br -c as it provides a much prettier output. Compare, before: 
anarcat@angela:~> ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host noprefixroute 
       valid_lft forever preferred_lft forever
2: wlan0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
    link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff permaddr xx:xx:xx:xx:xx:xx
    altname wlp166s0
    altname wlx8cf8c57333c7
4: virbr0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
    link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff
    inet 192.168.122.1/24 brd 192.168.122.255 scope global virbr0
       valid_lft forever preferred_lft forever
20: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff
    inet 192.168.0.108/24 brd 192.168.0.255 scope global dynamic noprefixroute eth0
       valid_lft 40699sec preferred_lft 40699sec
After: 
anarcat@angela:~> ip -br -c a
lo               UNKNOWN        127.0.0.1/8 ::1/128 
wlan0            DOWN           
virbr0           DOWN           192.168.122.1/24 
eth0             UP             192.168.0.108/24
I don't even need to redact MAC addresses! It also affects the display of the other commands, which look similarly neat. Also imagine pretty colors above. Finally, I don't have a cheat sheet for iw vs iwconfig (from wireless-tools) yet. I just use NetworkManager now and rarely have to mess with wireless interfaces directly.

Background and history For context, there are traditionally two ways of configuring the network in Linux:
  • the old way, with commands like ifconfig, arp, route and netstat, those are part of the net-tools package
  • the new way, mostly (but not entirely!) wrapped in a single ip command, that is the iproute2 package
It seems like the latter was made "important" in Debian in 2008, which means every release since Debian 5 "lenny" (!) has featured the ip command. The former net-tools package was demoted in December 2016 which means every release since Debian 9 "stretch" ships without an ifconfig command unless explicitly requested. Note that this was mentioned in the release notes in a similar (but, IMHO, less useful) table. (Technically, the net-tools Debian package source still indicates it is Priority: important but that's a bug I have just filed.) Finally, and perhaps more importantly, the name iproute is hilarious if you are a bilingual french speaker: it can be read as "I proute" which can be interpreted as "I fart" as "prout!" is the sound a fart makes. The fact that it's called iproute2 makes it only more hilarious.

13 January 2026

Louis-Philippe V ronneau: Reducing the size of initramfs kernel images

In the past few years, the size of the kernel images in Debian have been steadily growing. I don't see this as a problem per se, but it has been causing me trouble, as my /boot partition has become too small to accommodate two kernel images at the same time. Since I'm running Debian Unstable on my personal systems and keep them updated with unattended-upgrade, this meant each (frequent) kernel upgrade triggered an error like this one: 
update-initramfs: failed for /boot/initrd.img-6.17.11+deb14-amd64 with 1.
dpkg: error processing package initramfs-tools (--configure):
 installed initramfs-tools package post-installation script subprocess returned
 error exit status 1
Errors were encountered while processing:
 initramfs-tools
E: Sub-process /usr/bin/dpkg returned an error code (1)
This would in turn break the automated upgrade process and require me to manually delete the currently running kernel (which works, but isn't great) to complete the upgrade. The "obvious" solution would have been to increase the size of my /boot partition to something larger than the default 456M. Since my systems use full-disk encryption and LVM, this isn't trivial and would have required me to play Tetris and swap files back and forth using another drive. Another solution proposed by anarcat was to migrate to systemd-boot (I'm still using grub), use Unified Kernel Images (UKI) and merge the /boot and /boot/efi partitions. Since I already have a bunch of configurations using grub and I am not too keen on systemd taking over all the things on my computer, I was somewhat reluctant. As my computers are all configured by Puppet, I could of course have done a complete system reinstallation, but again, this was somewhat more involved than what I wanted it to be. After looking online for a while, I finally stumbled on this blog post by Neil Brown detailing how to shrink the size of the initramfs images. With MODULES=dep my images shrunk from 188M to 41M, fixing my issue. Thanks Neil! I was somewhat worried removing kernel modules would break something on my systems, but so far, I only had to manually load the i2c_dev module, as I need it to manage my home monitor's brightness using ddcutil.

21 December 2025

Russell Coker: Links December 2025

Russ Allbery wrote an interesting review of Politics on the Edge, by Rory Stewart who sems like one of the few conservative politicians I could respect and possibly even like [1]. It has some good insights about the problems with our current political environment. The NY Times has an amusing article about the attempt to sell the solution to the CIA s encrypted artwork [2]. Wired has an interesting article about computer face recognition systems failing on people with facial disabilities or scars [3]. This is a major accessibility issue potentially violating disability legislation and a demonstration of the problems of fully automating systems when there should be a human in the loop. The October 2025 report from the Debian Reproducible Builds team is particularly interesting [4]. kpcyrd forwarded a fascinating tidbit regarding so-called ninja and samurai build ordering, that uses data structures in which the pointer values returned from malloc are used to determine some order of execution LOL Louis Rossmann made an insightful youtube video about the moral case for piracy of software and media [5]. Louis Rossman made an insightful video about the way that Hyundai is circumventing Right to Repair laws to make repairs needlessly expensive [6]. Korean cars aren t much good nowadays. Their prices keep increasing and the quality doesn t. Brian Krebs wrote an interesting article about how Google is taking legal action against SMS phishing crime groups [7]. We need more of this! Josh Griffiths wrote an informative blog post about how YouTube is awful [8]. I really should investigate Peertube. Louis Rossman made an informative YouTube video about Right to Repair and the US military, if even the US military is getting ripped off by this it s a bigger problem than most people realise [9]. He also asks the rhetorical question of whether politicians are bought or whether it s a subscription model . Brian Krebs wrote an informative article about the US plans to ban TP Link devices, OpenWRT seems like a good option [10]. Brian Krebs wrote an informative article about free streaming Android TV boxes that act as hidden residential VPN proxies [11]. Also the free streaming violates copyright law. Bruce Schneier and Nathan E. Sanders wrote an interesting article about ways that AI is being used to strengthen democracy [12]. Cory Doctorow wrote an insightful article about the incentives for making shitty goods and services and why we need legislation to protect consumers [13]. Linus Tech Tips has an interesting interview with Linus Torvalds [14]. Interesting video about the Kowloon Walled City [15]. It would be nice if a government deliberately created a hive city like that, the only example I know of is the Alaskan town in a single building. David Brin wrote an insightful set of 3 blog posts about a Democratic American deal that could improve the situation there [16].
Related posts:
  1. Links May 2025 Christopher Biggs gave an informative Evrything Open lecture about voice...
  2. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  3. Links April 2025 Asianometry has an interesting YouTube video about elecrolytic capacitors degrading...

30 September 2025

Russell Coker: Links September 2025

Werdahias wrote an informative blog post about Dark Mode for QT programs on non-QT environments (mostly GNOME based), we need more blog posts about this sort of thing [1]. Astral Codex Ten has an interesting blog post about the rise of Christianity, trying to work out why it took over so quickly [2]. Frances Haugen s whistleblower statement about Facebook is worth reading, Facebook seems to be one of the most evil companies in the world [3]. Interesting blog post by Philip Bennett about trying to repair a 28 player Galaxian game from 1990 [4]. Bruce Schneier and Nathan E. Sanders wrote an insightful article about AI in Government [5]. Krebs has an interesting analysis of Conservatives whinging about alleged discrimination due to their use of spam lists [6]. Nick Cheesman wrote an insightful article on the failures of Meritocracy with ANU as a case study [7]. I am mystified as to why ABC categorised it under Religion. David Brin wrote an interesting short SciFi story about dealing with blackmail [8]. Charles Stross has an interesting take on AI economics etc [9]. Corey Doctorow wrote an interesting blog post about the impending economic crash becuase of all the money tied up in AI investments [10].
Related posts:
  1. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  2. Links February 2025 Oliver Lindburg wrote an interesting article about Designing for Crisis...
  3. Links July 2025 Louis Rossman made an informative YouTube video about right to...

31 August 2025

Russell Coker: Links August 2025

Dimitri John Ledkov wrote an informative blog post about self encrypting disks and UEFI with Linux [1]. This Coffeezilla video highlights an interesting scam, run a broker for day traders and don t execute trades, just be the counterparty for every trade and rely on day traders losing [2]. First Sight is a Dust SciFi short film that s worthy of Black Mirror [3]. Apple published an interesting article about the operation of the Secure Enclave in iPhone, Mac, and all other significant Apple hardware [4]. Reese Waters made an amusing video about a conservative catfight that s going on, nice to see horrible people attacking each other [5]. Chengyuan Ma wrote an informative summary of the history of the Great Firewall of China [6]. The V2Ray proxy has a nice feature set! Interesting article about the JP Morgan Workplace Activity Data Utility (WADU) AI spyware system [7]. Corporate work is going to become even more horrible. Veritasium has a great video about the history of vulcanised rubber and the potential for significant problems if rubber leaf blight spreads to other countries [8]. Jalopnik has an interesting article about how Reagan killed the safest car ever built [9]. David Brin wrote an interesting article Tolkien Enemy of Progress about fiction that justifies autocratic rule [10].
Related posts:
  1. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  2. Links July 2025 Louis Rossman made an informative YouTube video about right to...
  3. Links April 2025 Asianometry has an interesting YouTube video about elecrolytic capacitors degrading...

31 May 2025

Russell Coker: Links May 2025

Christopher Biggs gave an informative Evrything Open lecture about voice recognition [1]. We need this on Debian phones. Guido wrote an informative blog post about booting a custom Android kernel on a Pixel 3a [2]. Good work in writing this up, but a pity that Google made the process so difficult. Interesting to read about an expert being a victim of a phishing attack [3]. It can happen to anyone, everyone has moments when they aren t concentrating. Interesting advice on how to leak to a journalist [4]. Brian Krebs wrote an informative article about the ways that Trump is deliberately reducing the cyber security of the US government [5]. Brian Krebs wrote an interesting article about the main smishng groups from China [6]. Louis Rossmann (who is known for high quality YouTube videos about computer repair) made an informative video about a scammy Australian company run by a child sex offender [7]. The Helmover was one of the wildest engineering projects of WW2, an over the horizon guided torpedo that could one-shot a battleship [8]. Interesting blog post about DDoSecrets and the utter failure of the company Telemessages which was used by the US government [9]. Jonathan McDowell wrote an interesting blog post about developing a free software competitor to Alexa etc, the listening hardware costs $13US per node [10]. Noema Magazine published an insightful article about Rewilding the Internet, it has some great ideas [11].
Related posts:
  1. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  2. Links April 2025 Asianometry has an interesting YouTube video about elecrolytic capacitors degrading...
  3. Links February 2025 Oliver Lindburg wrote an interesting article about Designing for Crisis...

Antoine Beaupr : Traffic meter per ASN without logs

Have you ever found yourself in the situation where you had no or anonymized logs and still wanted to figure out where your traffic was coming from? Or you have multiple upstreams and are looking to see if you can save fees by getting into peering agreements with some other party? Or your site is getting heavy load but you can't pinpoint it on a single IP and you suspect some amoral corporation is training their degenerate AI on your content with a bot army? (You might be getting onto something there.) If that rings a bell, read on.

TL;DR: ... or just skip the cruft and install asncounter: 
pip install asncounter
Also available in Debian 14 or later, or possibly in Debian 13 backports (soon to be released) if people are interested: 
apt install asncounter
Then count whoever is hitting your network with: 
awk ' print $2 ' /var/log/apache2/*access*.log   asncounter
or: 
tail -F /var/log/apache2/*access*.log   awk ' print $2 '   asncounter
or: 
tcpdump -q -n   asncounter --input-format=tcpdump --repl
or: 
tcpdump -q -i eth0 -n -Q in "tcp and tcp[tcpflags] & tcp-syn != 0 and (port 80 or port 443)"   asncounter --input-format=tcpdump --repl
Read on for why this matters, and why I wrote yet another weird tool (almost) from scratch.

Background and manual work This is a tool I've been dreaming of for a long, long time. Back in 2006, at Koumbit a colleague had setup TAS ("Traffic Accounting System", " " in Russian, apparently), a collection of Perl script that would do per-IP accounting. It was pretty cool: it would count bytes per IP addresses and, from that, you could do analysis. But the project died, and it was kind of bespoke. Fast forward twenty years, and I find myself fighting off bots at the Tor Project (the irony...), with our GitLab suffering pretty bad slowdowns (see issue tpo/tpa/team#41677 for the latest public issue, the juicier one is confidential, unfortunately). (We did have some issues caused by overloads in CI, as we host, after all, a fork of Firefox, which is a massive repository, but the applications team did sustained, awesome work to fix issues on that side, again and again (see tpo/applications/tor-browser#43121 for the latest, and tpo/applications/tor-browser#43121 for some pretty impressive correlation work, I work with really skilled people). But those issues, I believe were fixed.) So I had the feeling it was our turn to get hammered by the AI bots. But how do we tell? I could tell something was hammering at the costly /commit/ and (especially costly) /blame/ endpoint. So at first, I pulled out the trusted awk, sort uniq -c sort -n tail pipeline I am sure others have worked out before: 
awk ' print $1 ' /var/log/nginx/*.log   sort   uniq -c   sort -n   tail -10
For people new to this, that pulls the first field out of web server log files, sort the list, counts the number of unique entries, and sorts that so that the most common entries (or IPs) show up first, then show the top 10. That, other words, answers the question of "which IP address visits this web server the most?" Based on this, I found a couple of IP addresses that looked like Alibaba. I had already addressed an abuse complaint to them (tpo/tpa/team#42152) but never got a response, so I just blocked their entire network blocks, rather violently: 
for cidr in 47.240.0.0/14 47.246.0.0/16 47.244.0.0/15 47.235.0.0/16 47.236.0.0/14; do 
  iptables-legacy -I INPUT -s $cidr -j REJECT
done
That made Ali Baba and his forty thieves (specifically their AL-3 network go away, but our load was still high, and I was still seeing various IPs crawling the costly endpoints. And this time, it was hard to tell who they were: you'll notice all the Alibaba IPs are inside the same 47.0.0.0/8 prefix. Although it's not a /8 itself, it's all inside the same prefix, so it's visually easy to pick it apart, especially for a brain like mine who's stared too long at logs flowing by too fast for their own mental health. What I had then was different, and I was tired of doing the stupid thing I had been doing for decades at this point. I had recently stumbled upon pyasn recently (in January, according to my notes) and somehow found it again, and thought "I bet I could write a quick script that loops over IPs and counts IPs per ASN". (Obviously, there are lots of other tools out there for that kind of monitoring. Argos, for example, presumably does this, but it's a kind of a huge stack. You can also get into netflows, but there's serious privacy implications with those. There are also lots of per-IP counters like promacct, but that doesn't scale. Or maybe someone already had solved this problem and I just wasted a week of my life, who knows. Someone will let me know, I hope, either way.)

ASNs and networks A quick aside, for people not familiar with how the internet works. People that know about ASNs, BGP announcements and so on can skip. The internet is the network of networks. It's made of multiple networks that talk to each other. The way this works is there is a Border Gateway Protocol (BGP), a relatively simple TCP-based protocol, that the edge routers of those networks used to announce each other what network they manage. Each of those network is called an Autonomous System (AS) and has an AS number (ASN) to uniquely identify it. Just like IP addresses, ASNs are allocated by IANA and local registries, they're pretty cheap and useful if you like running your own routers, get one. When you have an ASN, you'll use it to, say, announce to your BGP neighbors "I have 198.51.100.0/24" over here and the others might say "okay, and I have 216.90.108.31/19 over here, and I know of this other ASN over there that has 192.0.2.1/24 too! And gradually, those announcements flood the entire network, and you end up with each BGP having a routing table of the global internet, with a map of which network block, or "prefix" is announced by which ASN. It's how the internet works, and it's a useful thing to know, because it's what, ultimately, makes an organisation responsible for an IP address. There are "looking glass" tools like the one provided by routeviews.org which allow you to effectively run "trace routes" (but not the same as traceroute, which actively sends probes from your location), type an IP address in that form to fiddle with it. You will end up with an "AS path", the way to get from the looking glass to the announced network. But I digress, and that's kind of out of scope. Point is, internet is made of networks, networks are autonomous systems (AS) and they have numbers (ASNs), and they announced IP prefixes (or "network blocks") that ultimately tells you who is responsible for traffic on the internet.

Introducing asncounter So my goal was to get from "lots of IP addresses" to "list of ASNs", possibly also the list of prefixes (because why not). Turns out pyasn makes that really easy. I managed to build a prototype in probably less than an hour, just look at the first version, it's 44 lines (sloccount) of Python, and it works, provided you have already downloaded the required datafiles from routeviews.org. (Obviously, the latest version is longer at close to 1000 lines, but it downloads the data files automatically, and has many more features). The way the first prototype (and later versions too, mostly) worked is that you feed it a list of IP addresses on standard input, it looks up the ASN and prefix associated with the IP, and increments a counter for those, then print the result. That showed me something like this: 
root@gitlab-02:~/anarcat-scripts# tcpdump -q -i eth0 -n -Q in "(udp or tcp)"   ./asncounter.py --tcpdump                                                                                                                                                                          
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode                                                                
listening on eth0, link-type EN10MB (Ethernet), snapshot length 262144 bytes                                                             
INFO: collecting IPs from stdin, using datfile ipasn_20250523.1600.dat.gz                                                                
INFO: loading datfile /root/.cache/pyasn/ipasn_20250523.1600.dat.gz...                                                                   
INFO: loading /root/.cache/pyasn/asnames.json                       
ASN     count   AS               
136907  7811    HWCLOUDS-AS-AP HUAWEI CLOUDS, HK                                                                                         
[----]  359     [REDACTED]
[----]  313     [REDACTED]
8075    254     MICROSOFT-CORP-MSN-AS-BLOCK, US
[---]   164     [REDACTED]
[----]  136     [REDACTED]
24940   114     HETZNER-AS, DE  
[----]  98      [REDACTED]
14618   82      AMAZON-AES, US                                                                                                           
[----]  79      [REDACTED]
prefix  count                                         
166.108.192.0/20        1294                                                                                                             
188.239.32.0/20 1056                                          
166.108.224.0/20        970                    
111.119.192.0/20        951              
124.243.128.0/18        667                                         
94.74.80.0/20   651                                                 
111.119.224.0/20        622                                         
111.119.240.0/20        566           
111.119.208.0/20        538                                         
[REDACTED]  313
Even without ratios and a total count (which will come later), it was quite clear that Huawei was doing something big on the server. At that point, it was responsible for a quarter to half of the traffic on our GitLab server or about 5-10 queries per second. But just looking at the logs, or per IP hit counts, it was really hard to tell. That traffic is really well distributed. If you look more closely at the output above, you'll notice I redacted a couple of entries except major providers, for privacy reasons. But you'll also notice almost nothing is redacted in the prefix list, why? Because all of those networks are Huawei! Their announcements are kind of bonkers: they have hundreds of such prefixes. Now, clever people in the know will say "of course they do, it's an hyperscaler; just ASN14618 (AMAZON-AES) there is way more announcements, they have 1416 prefixes!" Yes, of course, but they are not generating half of my traffic (at least, not yet). But even then: this also applies to Amazon! This way of counting traffic is way more useful for large scale operations like this, because you group by organisation instead of by server or individual endpoint. And, ultimately, this is why asncounter matters: it allows you to group your traffic by organisation, the place you can actually negotiate with. Now, of course, that assumes those are entities you can talk with. I have written to both Alibaba and Huawei, and have yet to receive a response. I assume I never will. In their defence, I wrote in English, perhaps I should have made the effort of translating my message in Chinese, but then again English is the Lingua Franca of the Internet, and I doubt that's actually the issue.

The Huawei and Facebook blocks Another aside, because this is my blog and I am not looking for a Pullitzer here. So I blocked Huawei from our GitLab server (and before you tear your shirt open: only our GitLab server, everything else is still accessible to them, including our email server to respond to my complaint). I did so 24h after emailing them, and after examining their user agent (UA) headers. Boy that was fun. In a sample of 268 requests I analyzed, they churned out 246 different UAs. At first glance, they looked legit, like: 
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36
Safari on a Mac, so far so good. But when you start digging, you notice some strange things, like here's Safari running on Linux: 
Mozilla/5.0 (X11; U; Linux i686; en-US) AppleWebKit/534.3 (KHTML, like Gecko) Chrome/6.0.457.0 Safari/534.3
Was Safari ported to Linux? I guess that's.. possible? But here is Safari running on a 15 year old Ubuntu release (10.10): 
Mozilla/5.0 (X11; Linux i686) AppleWebKit/534.24 (KHTML, like Gecko) Ubuntu/10.10 Chromium/12.0.702.0 Chrome/12.0.702.0 Safari/534.24
Speaking of old, here's Safari again, but this time running on Windows NT 5.1, AKA Windows XP, released 2001, EOL since 2019: 
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-CA) AppleWebKit/534.13 (KHTML like Gecko) Chrome/9.0.597.98 Safari/534.13
Really? Here's Firefox 3.6, released 14 years ago, there were quite a lot of those: 
Mozilla/5.0 (Windows; U; Windows NT 6.1; lt; rv:1.9.2) Gecko/20100115 Firefox/3.6
I remember running those old Firefox releases, those were the days. But to me, those look like entirely fake UAs, deliberately rotated to make it look like legitimate traffic. In comparison, Facebook seemed a bit more legit, in the sense that they don't fake it. most hits are from: 
meta-externalagent/1.1 (+https://developers.facebook.com/docs/sharing/webmasters/crawler)
which, according their documentation:
crawls the web for use cases such as training AI models or improving products by indexing content directly
From what I could tell, it was even respecting our rather liberal robots.txt rules, in that it wasn't crawling the sprawling /blame/ or /commit/ endpoints, explicitly forbidden by robots.txt. So I've blocked the Facebook bot in robots.txt and, amazingly, it just went away. Good job Facebook, as much as I think you've given the empire to neo-nazis, cause depression and genocide, you know how to run a crawler, thanks. Huawei was blocked at the web server level, with a friendly 429 status code telling people to contact us (over email) if they need help. And they don't care: they're still hammering the server, from what I can tell, but then again, I didn't block the entire ASN just yet, just the blocks I found crawling the server over a couple hours.

A full asncounter run So what does a day in asncounter look like? Well, you start with a problem, say you're getting too much traffic and want to see where it's from. First you need to sample it. Typically, you'd do that with tcpdump or tailing a log file: 
tail -F /var/log/apache2/*access*.log   awk ' print $2 '   asncounter
If you have lots of traffic or care about your users' privacy, you're not going to log IP addresses, so tcpdump is likely a good option instead: 
tcpdump -q -n   asncounter --input-format=tcpdump --repl
If you really get a lot of traffic, you might want to get a subset of that to avoid overwhelming asncounter, it's not fast enough to do multiple gigabit/second, I bet, so here's only incoming SYN IPv4 packets: 
tcpdump -q -n -Q in "tcp and tcp[tcpflags] & tcp-syn != 0 and (port 80 or port 443)"   asncounter --input-format=tcpdump --repl
In any case, at this point you're staring at a process, just sitting there. If you passed the --repl or --manhole arguments, you're lucky: you have a Python shell inside the program. Otherwise, send SIGHUP to the thing to have it dump the nice tables out: 
pkill -HUP asncounter
Here's an example run: 
> awk ' print $2 ' /var/log/apache2/*access*.log   asncounter
INFO: using datfile ipasn_20250527.1600.dat.gz
INFO: collecting addresses from <stdin>
INFO: loading datfile /home/anarcat/.cache/pyasn/ipasn_20250527.1600.dat.gz...
INFO: finished reading data
INFO: loading /home/anarcat/.cache/pyasn/asnames.json
count   percent ASN AS
12779   69.33   66496   SAMPLE, CA
3361    18.23   None    None
366 1.99    66497   EXAMPLE, FR
337 1.83    16276   OVH, FR
321 1.74    8075    MICROSOFT-CORP-MSN-AS-BLOCK, US
309 1.68    14061   DIGITALOCEAN-ASN, US
128 0.69    16509   AMAZON-02, US
77  0.42    48090   DMZHOST, GB
56  0.3 136907  HWCLOUDS-AS-AP HUAWEI CLOUDS, HK
53  0.29    17621   CNCGROUP-SH China Unicom Shanghai network, CN
total: 18433
count   percent prefix  ASN AS
12779   69.33   192.0.2.0/24    66496   SAMPLE, CA
3361    18.23   None        
298 1.62    178.128.208.0/20    14061   DIGITALOCEAN-ASN, US
289 1.57    51.222.0.0/16   16276   OVH, FR
272 1.48    2001:DB8::/48   66497   EXAMPLE, FR
235 1.27    172.160.0.0/11  8075    MICROSOFT-CORP-MSN-AS-BLOCK, US
94  0.51    2001:DB8:1::/48 66497   EXAMPLE, FR
72  0.39    47.128.0.0/14   16509   AMAZON-02, US
69  0.37    93.123.109.0/24 48090   DMZHOST, GB
53  0.29    27.115.124.0/24 17621   CNCGROUP-SH China Unicom Shanghai network, CN
Those numbers are actually from my home network, not GitLab. Over there, the battle still rages on, but at least the vampire bots are banging their heads against the solid Nginx wall instead of eating the fragile heart of GitLab. We had a significant improvement in latency thanks to the Facebook and Huawei blocks... Here are the "workhorse request duration stats" for various time ranges, 20h after the block:
range mean max stdev
20h 449ms 958ms 39ms
7d 1.78s 5m 14.9s
30d 2.08s 3.86m 8.86s
6m 901ms 27.3s 2.43s
We went from two seconds mean to 500ms! And look at that standard deviation! 39ms! It was ten seconds before! I doubt we'll keep it that way very long but for now, it feels like I won a battle, and I didn't even have to setup anubis or go-away, although I suspect that will unfortunately come. Note that asncounter also supports exporting Prometheus metrics, but you should be careful with this, as it can lead to cardinal explosion, especially if you track by prefix (which can be disabled with --no-prefixes . Folks interested in more details should read the fine manual for more examples, usage, and discussion. It shows, among other things, how to effectively block lots of networks from Nginx, aggregate multiple prefixes, block entire ASNs, and more! So there you have it: I now have the tool I wish I had 20 years ago. Hopefully it will stay useful for another 20 years, although I'm not sure we'll have still have internet in 20 years. I welcome constructive feedback, "oh no you rewrote X", Grafana dashboards, bug reports, pull requests, and "hell yeah" comments. Hacker News, let it rip, I know you can give me another juicy quote for my blog. This work was done as part of my paid work for the Tor Project, currently in a fundraising drive, give us money if you like what you read.

30 April 2025

Russell Coker: Links April 2025

Asianometry has an interesting YouTube video about elecrolytic capacitors degrading and how they affect computers [1]. Keep your computers cool people! Biella Coleman (famous for studying the Anthropology of Debian) and Eric Reinhart wrote an interesting article about MAHA (Make America Healthy Again) and how it ended up doing exactly the opposite of what was intended [2]. SciShow has an informative video about lung cancer cases among non-smokers, the risk factors are genetics, Radon, and cooking [3]. Ian Jackson wrote an insightful blog post about whether Rust is woke [4]. Bruce Schneier write an interesting blog post about research into making AIs Trusted Third Parties [5]. This has the potential to solve some cryptology problems. CHERIoT is an interesting project for controlling all jump statements in RISC-V among other related security features [6]. We need this sort of thing for IoT devices that will run for years without change. Brian Krebs wrote an informative post about how Trump is attacking the 1st Amendment of the US Constitution [7]. The Register has an interesting summary of the kernel enclave and exclave functionality in recent Apple OSs [8]. Dr Gabor Mate wrote an interesting psychological analysis of Hillary Clinton and Donald Trump [9]. ChoiceJacking is an interesting variant of the JuiceJacking attack on mobile phones by hostile chargers [10]. They should require input for security sensitive events to come from the local hardware not USB or Bluetooth.
Related posts:
  1. Links March 2025 Anarcat s review of Fish is interesting and shows some benefits...
  2. Links January 2025 Aaron Quigley s Everything Open lecture about Intelligent Interfaces is one...
  3. Links February 2025 Oliver Lindburg wrote an interesting article about Designing for Crisis...

31 March 2025

Russell Coker: Links March 2025

Anarcat s review of Fish is interesting and shows some benefits I hadn t previously realised, I ll have to try it out [1]. Longnow has an insightful article about religion and magic mushrooms [2]. Brian Krebs wrote an informative artivle about DOGE and the many security problems that it has caused to the US government [3]. Techdirt has an insightful article about why they are forced to become a democracy blog after the attacks by Trump et al [4]. Antoine wrote an insightful blog post about the war for the Internet and how in many ways we are losing to fascists [5]. Interesting story about people working for free at Apple to develop a graphing calculator [6]. We need ways for FOSS people to associate to do such projects. Interesting YouTube video about a wiki for building a cheap road legal car [7]. Interesting video about powering spacecraft with Plutonion 238 and how they are running out [8]. Interesting information about the search for mh370 [9]. I previously hadn t been convinced that it was hijacked but I am now. The EFF has an interesting article about the Rayhunter, a tool to detect cellular spying that can run with cheap hardware [10].
  • [1] https://anarc.at/blog/2025-02-28-fish/
  • [2] https://longnow.org/ideas/is-god-a-mushroom/
  • [3] https://tinyurl.com/27wbb5ec
  • [4] https://tinyurl.com/2cvo42ro
  • [5] https://anarc.at/blog/2025-03-21-losing-war-internet/
  • [6] https://www.pacifict.com/story/
  • [7] https://www.youtube.com/watch?v=x8jdx-lf2Dw
  • [8] https://www.youtube.com/watch?v=geIhl_VE0IA
  • [9] https://www.youtube.com/watch?v=HIuXEU4H-XE
  • [10] https://tinyurl.com/28psvpx7
    • Related posts:
    1. Links January 2025 Aaron Quigley s Everything Open lecture about Intelligent Interfaces is one...
    2. Links March 2024 Bruce Schneier wrote an interesting blog post about his workshop...
    3. Links February 2025 Oliver Lindburg wrote an interesting article about Designing for Crisis...

    6 March 2025

    Antoine Beaupr : Nix Notes

    Meta In case you haven't noticed, I'm trying to post and one of the things that entails is to just dump over the fence a bunch of draft notes. In this specific case, I had a set of rough notes about NixOS and particularly Nix, the package manager. In this case, you can see the very birth of an article, what it looks like before it becomes the questionable prose it is now, by looking at the Git history of this file, particularly its birth. I have a couple of those left, and it would be pretty easy to publish them as is, but I feel I'd be doing others (and myself! I write for my own documentation too after all) a disservice by not going the extra mile on those. So here's the long version of my experiment with Nix.

    Nix A couple friends are real fans of Nix. Just like I work with Puppet a lot, they deploy and maintain servers (if not fleets of servers) with NixOS and its declarative package management system. Essentially, they use it as a configuration management system, which is pretty awesome. That, however, is a bit too high of a bar for me. I rarely try new operating systems these days: I'm a Debian developer and it takes most of my time to keep that functional. I'm not going to go around messing with other systems as I know that would inevitably get me dragged down into contributing into yet another free software project. I'm mature now and know where to draw the line. Right? So I'm just testing Nix, the package manager, on Debian, because I learned from my friend that nixpkgs is the largest package repository out there, a mind-boggling 100,000 at the time of writing (with 88% of packages up to date), compared to around 40,000 in Debian (or 72,000 if you count binary packages, with 72% up to date). I naively thought Debian was the largest, perhaps competing with Arch, and I was wrong: Arch is larger than Debian too. What brought me there is I wanted to run Harper, a fast spell-checker written in Rust. The logic behind using Nix instead of just downloading the source and running it myself is that I delegate the work of supply-chain integrity checking to a distributor, a bit like you trust Debian developers like myself to package things in a sane way. I know this widens the attack surface to a third party of course, but the rationale is that I shift cryptographic verification to another stack than just "TLS + GitHub" (although that is somewhat still involved) that's linked with my current chain (Debian packages). I have since then stopped using Harper for various reasons and also wrapped up my Nix experiment, but felt it worthwhile to jot down some observations on the project.

    Hot take Overall, Nix is hard to get into, with a complicated learning curve. I have found the documentation to be a bit confusing, since there are many ways to do certain things. I particularly tripped on "flakes" and, frankly, incomprehensible error reporting. It didn't help that I tried to run nixpkgs on Debian which is technically possible, but you can tell that I'm not supposed to be doing this. My friend who reviewed this article expressed surprised at how easy this was, but then he only saw the finished result, not me tearing my hair out to make this actually work.

    Nix on Debian primer So here's how I got started. First I installed the nix binary package: 
    apt install nix-bin
    Then I had to add myself to the right group and logout/log back in to get the rights to deploy Nix packages: 
    adduser anarcat nix-users
    That wasn't easy to find, but is mentioned in the README.Debian file shipped with the Debian package. Then, I didn't write this down, but the README.Debian file above mentions it, so I think I added a "channel" like this: 
    nix-channel --add https://nixos.org/channels/nixpkgs-unstable nixpkgs
nix-channel --update
    And I likely installed the Harper package with: 
    nix-env --install harper
    At this point, harper was installed in a ... profile? Not sure. I had to add ~/.nix-profile/bin (a symlink to /nix/store/sympqw0zyybxqzz6fzhv03lyivqqrq92-harper-0.10.0/bin) to my $PATH environment for this to actually work.

    Side notes on documentation Those last two commands (nix-channel and nix-env) were hard to figure out, which is kind of amazing because you'd think a tutorial on Nix would feature something like this prominently. But three different tutorials failed to bring me up to that basic setup, even the README.Debian didn't spell that out clearly. The tutorials all show me how to develop packages for Nix, not plainly how to install Nix software. This is presumably because "I'm doing it wrong": you shouldn't just "install a package", you should setup an environment declaratively and tell it what you want to do. But here's the thing: I didn't want to "do the right thing". I just wanted to install Harper, and documentation failed to bring me to that basic "hello world" stage. Here's what one of the tutorials suggests as a first step, for example: 
    curl -L https://nixos.org/nix/install   sh
nix-shell --packages cowsay lolcat
nix-collect-garbage
    ... which, when you follow through, leaves you with almost precisely nothing left installed (apart from Nix itself, setup with a nasty "curl pipe bash". So while that works in testing Nix, you're not much better off than when you started.

    Rolling back everything Now that I have stopped using Harper, I don't need Nix anymore, which I'm sure my Nix friends will be sad to read about. Don't worry, I have notes now, and can try again! But still, I wanted to clear things out, so I did this, as root: 
    deluser anarcat nix-users
apt purge nix-bin
rm -rf /nix ~/.nix*
    I think this cleared things out, but I'm not actually sure.

    Side note on Nix drama This blurb wouldn't be complete without a mention that the Nix community has been somewhat tainted by the behavior of its founder. I won't bother you too much with this; LWN covered it well in 2024, and made a followup article about spinoffs and forks that's worth reading as well. I did want to say that everyone I have been in contact with in the Nix community was absolutely fantastic. So I am really sad that the behavior of a single individual can pollute a community in such a way. As a leader, if you have all but one responsability, it's to behave properly for people around you. It's actually really, really hard to do that, because yes, it means you need to act differently than others and no, you just don't get to be upset at others like you would normally do with friends, because you're in a position of authority. It's a lesson I'm still learning myself, to be fair. But at least I don't work with arms manufacturers or, if I would, I would be sure as hell to accept the nick (or nix?) on the chin when people would get upset, and try to make amends. So long live the Nix people! I hope the community recovers from that dark moment, so far it seems like it will. And thanks for helping me test Harper!

    28 February 2025

    Antoine Beaupr : testing the fish shell

    I have been testing fish for a couple months now (this file started on 2025-01-03T23:52:15-0500 according to stat(1)), and those are my notes. I suspect people will have Opinions about my comments here. Do not comment unless you have some Constructive feedback to provide: I don't want to know if you think I am holding it Wrong. Consider that I might have used UNIX shells for longer that you have lived. I'm not sure I'll keep using fish, but so far it's the first shell that survived heavy use outside of zsh(1) (unless you count tcsh(1), but that was in another millenia). My normal shell is bash(1), and it's still the shell I used everywhere else than my laptop, as I haven't switched on all the servers I managed, although it is available since August 2022 on torproject.org servers. I first got interested in fish because they ported to Rust, making it one of the rare shells out there written in a "safe" and modern programming language, released after an impressive ~2 year of work with Fish 4.0.

    Cool things Current directory gets shortened, ~/wikis/anarc.at/software/desktop/wayland shows up as ~/w/a/s/d/wayland Autocompletion rocks. Default prompt rocks. Doesn't seem vulnerable to command injection assaults, at least it doesn't trip on the git-landmine. It even includes pipe status output, which was a huge pain to implement in bash. Made me realized that if the last command succeeds, we don't see other failures, which is the case of my current prompt anyways! Signal reporting is better than my bash implementation too. So far the only modification I have made to the prompt is to add a printf '\a' to output a bell. By default, fish keeps a directory history (but separate from the pushd stack), that can be navigated with cdh, prevd, and nextd, dirh shows the history.

    Less cool I feel there's visible latency in the prompt creation. POSIX-style functions (foo() true ) are unsupported. Instead, fish uses whitespace-sensitive definitions like this: 
    function foo
    true
end
    This means my (modest) collection of POSIX functions need to be ported to fish. Workaround: simple functions can be turned into aliases, which fish supports (but implements using functions). EOF heredocs are considered to be "minor syntactic sugar". I find them frigging useful. Process substitution is split on newlines, not whitespace. you need to pipe through string split -n " " to get the equivalent. <(cmd) doesn't exist: they claim you can use cmd foo - as a replacement, but that's not correct: I used <(cmd) mostly where foo does not support - as a magic character to say 'read from stdin'. Documentation is... limited. It seems mostly geared the web docs which are... okay (but I couldn't find out about ~/.config/fish/conf.d there!), but this is really inconvenient when you're trying to browse the manual pages. For example, fish thinks there's a fish_prompt manual page, according to its own completion mechanism, but man(1) cannot find that manual page. I can't find the manual for the time command (which is actually a keyword!) Fish renders multi-line commands with newlines. So if your terminal looks like this, say: 
    anarcat@angela:~> sq keyring merge torproject-keyring/lavamind-
95F341D746CF1FC8B05A0ED5D3F900749268E55E.gpg torproject-keyrin
g/weasel-E3ED482E44A53F5BBE585032D50F9EBC09E69937.gpg   wl-copy
    ... but it's actually one line, when you copy-paste the above, in foot(1), it will show up exactly like this, newlines and all: 
    sq keyring merge torproject-keyring/lavamind-
95F341D746CF1FC8B05A0ED5D3F900749268E55E.gpg torproject-keyrin
g/weasel-E3ED482E44A53F5BBE585032D50F9EBC09E69937.gpg   wl-copy
    Whereas it should show up like this: 
    sq keyring merge torproject-keyring/lavamind-95F341D746CF1FC8B05A0ED5D3F900749268E55E.gpg torproject-keyring/weasel-E3ED482E44A53F5BBE585032D50F9EBC09E69937.gpg   wl-copy
    Note that this is an issue specific to foot(1), alacritty(1) and gnome-terminal(1) don't suffer from that issue. I have already filed it upstream in foot and it is apparently fixed already. Globbing is driving me nuts. You can't pass a * to a command unless fish agrees it's going to match something. You need to escape it if it doesn't immediately match, and then you need the called command to actually support globbing. 202[345] doesn't match folders named 2023, 2024, 2025, it will send the string 202[345] to the command.

    Blockers () is like $(): it's process substitution, and not a subshell. This is really impractical: I use ( cd foo ; do_something) all the time to avoid losing the current directory... I guess I'm supposed to use pushd for this, but ouch. This wouldn't be so bad if it was just for cd though. Clean constructs like this: 
    ( git grep -l '^#!/.*bin/python' ; fdfind .py )   sort -u
    Turn into what i find rather horrible: 
    begin; git grep -l '^#!/.*bin/python' ; fdfind .py ; end   sort -ub
    It... works, but it goes back to "oh dear, now there's a new langage again". I only found out about this construct while trying: 
      git grep -l '^#!/.*bin/python' ; fdfind .py     sort -u
    ... which fails and suggests using begin/end, at which point: why not just support the curly braces? FOO=bar is not allowed. It's actually recognized syntax, but creates a warning. We're supposed to use set foo bar instead. This really feels like a needless divergence from standard. Aliases are... peculiar. Typical constructs like alias mv="\mv -i" don't work because fish treats aliases as a function definition, and \ is not magical there. This can be worked around by specifying the full path to the command, with e.g. alias mv="/bin/mv -i". Another problem is trying to override a built-in, which seems completely impossible. In my case, I like the time(1) command the way it is, thank you very much, and fish provides no way to bypass that builtin. It is possible to call time(1) with command time, but it's not possible to replace the command keyword so that means a lot of typing. Again: you can't use \ to bypass aliases. This is a huge annoyance for me. I would need to learn to type command in long form, and I use that stuff pretty regularly. I guess I could alias command to c or something, but this is one of those huge muscle memory challenges. alt . doesn't always work the way i expect.

    23 February 2025

    Colin Watson: Qalculate time hacks

    Anarcat recently wrote about Qalculate, and I think I m a convert, even though I ve only barely scratched the surface. The thing I almost immediately started using it for is time calculations. When I started tracking my time, I quickly found that Timewarrior was good at keeping all the data I needed, but I often found myself extracting bits of it and reprocessing it in variously clumsy ways. For example, I often don t finish a task in one sitting; maybe I take breaks, or I switch back and forth between a couple of different tasks. The raw output of timew summary is a bit clumsy for this, as it shows each chunk of time spent as a separate row: 
    $ timew summary 2025-02-18 Debian
Wk Date       Day Tags                            Start      End    Time   Total
W8 2025-02-18 Tue CVE-2025-26465, Debian,       9:41:44 10:24:17 0:42:33
                  next, openssh
                  Debian, FTBFS with GCC-15,   10:24:17 10:27:12 0:02:55
                  icoutils
                  Debian, FTBFS with GCC-15,   11:50:05 11:57:25 0:07:20
                  kali
                  Debian, Upgrade to 0.67,     11:58:21 12:12:41 0:14:20
                  python_holidays
                  Debian, FTBFS with GCC-15,   12:14:15 12:33:19 0:19:04
                  vigor
                  Debian, FTBFS with GCC-15,   12:39:02 12:39:38 0:00:36
                  python_setproctitle
                  Debian, Upgrade to 1.3.4,    12:39:39 12:46:05 0:06:26
                  python_setproctitle
                  Debian, FTBFS with GCC-15,   12:48:28 12:49:42 0:01:14
                  python_setproctitle
                  Debian, Upgrade to 3.4.1,    12:52:07 13:02:27 0:10:20 1:44:48
                  python_charset_normalizer
                                                                         1:44:48
    So I wrote this Python program to help me: 
    #! /usr/bin/python3
"""
Summarize timewarrior data, grouped and sorted by time spent.
"""
import json
import subprocess
from argparse import ArgumentParser, RawDescriptionHelpFormatter
from collections import defaultdict
from datetime import datetime, timedelta, timezone
from operator import itemgetter
from rich import box, print
from rich.table import Table
parser = ArgumentParser(
    description=__doc__, formatter_class=RawDescriptionHelpFormatter
)
parser.add_argument("-t", "--only-total", default=False, action="store_true")
parser.add_argument(
    "range",
    nargs="?",
    default=":today",
    help="Time range (usually a hint, e.g. :lastweek)",
)
parser.add_argument("tag", nargs="*", help="Tags to filter by")
args = parser.parse_args()
entries: defaultdict[str, timedelta] = defaultdict(timedelta)
now = datetime.now(timezone.utc)
for entry in json.loads(
    subprocess.run(
        ["timew", "export", args.range, *args.tag],
        check=True,
        capture_output=True,
        text=True,
    ).stdout
):
    start = datetime.fromisoformat(entry["start"])
    if "end" in entry:
        end = datetime.fromisoformat(entry["end"])
    else:
        end = now
    entries[", ".join(entry["tags"])] += end - start
if not args.only_total:
    table = Table(box=box.SIMPLE, highlight=True)
    table.add_column("Tags")
    table.add_column("Time", justify="right")
    for tags, time in sorted(entries.items(), key=itemgetter(1), reverse=True):
        table.add_row(tags, str(time))
    print(table)
total = sum(entries.values(), start=timedelta())
hours, rest = divmod(total, timedelta(hours=1))
minutes, rest = divmod(rest, timedelta(minutes=1))
seconds = rest.seconds
print(f"Total time:  hours:02 : minutes:02 : seconds:02 ")
    $ summarize-time 2025-02-18 Debian
  Tags                                                     Time
  
  CVE-2025-26465, Debian, next, openssh                 0:42:33
  Debian, FTBFS with GCC-15, vigor                      0:19:04
  Debian, Upgrade to 0.67, python_holidays              0:14:20
  Debian, Upgrade to 3.4.1, python_charset_normalizer   0:10:20
  Debian, FTBFS with GCC-15, kali                       0:07:20
  Debian, Upgrade to 1.3.4, python_setproctitle         0:06:26
  Debian, FTBFS with GCC-15, icoutils                   0:02:55
  Debian, FTBFS with GCC-15, python_setproctitle        0:01:50
Total time: 01:44:48
    Much nicer. But that only helps with some of my reporting. At the end of a month, I have to work out how much time to bill Freexian for and fill out a timesheet, and for various reasons those queries don t correspond to single timew tags: they sometimes correspond to the sum of all time spent on multiple tags, or to the time spent on one tag minus the time spent on another tag, or similar. As a result I quite often have to do basic arithmetic on time intervals; but that s surprisingly annoying! I didn t previously have good tools for that, and was reduced to doing things like str(timedelta(hours=..., minutes=..., seconds=...) + ...) in Python, which gets old fast. Instead: 
    $ qalc '62:46:30 - 51:02:42 to time'
(225990 / 3600)   (183762 / 3600) = 11:43:48
    I also often want to work out how much of my time I ve spent on Debian work this month so far, since Freexian pays me for up to 20% of my work time on Debian; if I m under that then I might want to prioritize more Debian projects, and if I m over then I should be prioritizing more Freexian projects as otherwise I m not going to get paid for that time. 
    $ summarize-time -t :month Freexian
Total time: 69:19:42
$ summarize-time -t :month Debian
Total time: 24:05:30
$ qalc '24:05:30 / (24:05:30 + 69:19:42) to %'
(86730 / 3600) / ((86730 / 3600) + (249582 / 3600))   25.78855349%
    I love it.

    21 February 2025

    Russell Coker: Links February 2025

    Oliver Lindburg wrote an interesting article about Designing for Crisis [1]. Bruce Schneier blogged about how to cryptographically identify other humans in advance of AT technology allowing faking people you know [2]. Anarcat has an interesting review of qalc which is a really good calculator, I ll install it on all my workstations [3]. It even does furlongs per fortnight! This would be good to be called from a LLM system when someone asks about mathematical things. Krebs has an informative article about a criminal employed by Elon s DOGE [4]. Conservatives tend to be criminals. Krebs wrote an interesting article about the security of the iOS (and presumably Android) apps for DeekSeek [5]. Seems that the DeepSeek people did everything wrong. Bruce Schneier and Davi Ottenheimer wrote an insightful article DOGE as a National Cyberattack [6]. Bruce Schneier and Barath Raghavan wrote an insightful article about why and how computer generated voices should sound robotic [7]. Cory Doctorow has an interesting approach to the trade war between the US and Canada, instead of putting tarrifs on imports from the US the Canadian government should make it legal for Canadians to unlock their own property [8]. This youtube video about designing a compressed air engine for a model plane is interesting [9]. Krebs has an interesting article on phishing and mobile phone wallets, Google and Apple need to restrict the number of wallets per phone [10]. The Daily WTF has a good summary of why Elon s DOGE organisation is badly designed and run and a brief mention of how it damages the US [11]. ArsTechnica has an informative article about device code phishing [12]. The increased use of single-sign-on is going to make this more of a problem. Shrivu wrote an insightful and informative article on how to backdoor LLMs [13]. Cory Doctorow wrote an informative post about MLMs and how they are the mirror world version of community organising [14].
    Related posts:
    1. Links March 2024 Bruce Schneier wrote an interesting blog post about his workshop...
    2. Links February 2024 In 2018 Charles Stross wrote an insightful blog post Dude...
    3. Links August 2024 Bruce Schneier and Kim C rdova wrote an insightful article about...

    9 February 2025

    Antoine Beaupr : A slow blogging year

    Well, 2024 will be remembered, won't it? I guess 2025 already wants to make its mark too, but let's not worry about that right now, and instead let's talk about me. A little over a year ago, I was gloating over how I had such a great blogging year in 2022, and was considering 2023 to be average, then went on to gather more stats and traffic analysis... Then I said, and I quote:
    I hope to write more next year. I've been thinking about a few posts I could write for work, about how things work behind the scenes at Tor, that could be informative for many people. We run a rather old setup, but things hold up pretty well for what we throw at it, and it's worth sharing that with the world...
    What a load of bollocks.

    A bad year for this blog 2024 was the second worst year ever in my blogging history, tied with 2009 at a measly 6 posts for the year: 
    anarcat@angela:anarc.at$ curl -sSL https://anarc.at/blog/   grep 'href="\./'   grep -o 20[0-9][0-9]   sort   uniq -c   sort -nr   grep -v 2025   tail -3
      6 2024
      6 2009
      3 2014
    I did write about my work though, detailing the migration from Gitolite to GitLab we completed that year. But after August, total radio silence until now.

    Loads of drafts It's not that I have nothing to say: I have no less than five drafts in my working tree here, not counting three actual drafts recorded in the Git repository here: 
    anarcat@angela:anarc.at$ git s blog
## main...origin/main
?? blog/bell-bot.md
?? blog/fish.md
?? blog/kensington.md
?? blog/nixos.md
?? blog/tmux.md
anarcat@angela:anarc.at$ git grep -l '\!tag draft'
blog/mobile-massive-gallery.md
blog/on-dying.mdwn
blog/secrets-recovery.md
    I just don't have time to wrap those things up. I think part of me is disgusted by seeing my work stolen by large corporations to build proprietary large language models while my idols have been pushed to suicide for trying to share science with the world. Another part of me wants to make those things just right. The "tagged drafts" above are nothing more than a huge pile of chaotic links, far from being useful for anyone else than me, and even then. The on-dying article, in particular, is becoming my nemesis. I've been wanting to write that article for over 6 years now, I think. It's just too hard.

    Writing elsewhere There's also the fact that I write for work already. A lot. Here are the top-10 contributors to our team's wiki: 
    anarcat@angela:help.torproject.org$ git shortlog --numbered --summary --group="format:%al"   head -10
  4272  anarcat
   423  jerome
   117  zen
   116  lelutin
   104  peter
    58  kez
    45  irl
    43  hiro
    18  gaba
    17  groente
    ... but that's a bit unfair, since I've been there half a decade. Here's the last year: 
    anarcat@angela:help.torproject.org$ git shortlog --since=2024-01-01 --numbered --summary --group="format:%al"   head -10
   827  anarcat
   117  zen
   116  lelutin
    91  jerome
    17  groente
    10  gaba
     8  micah
     7  kez
     5  jnewsome
     4  stephen.swift
    So I still write the most commits! But to truly get a sense of the amount I wrote in there, we should count actual changes. Here it is by number of lines (from commandlinefu.com): 
    anarcat@angela:help.torproject.org$ git ls-files   xargs -n1 git blame --line-porcelain   sed -n 's/^author //p'   sort -f   uniq -ic   sort -nr   head -10
  99046 Antoine Beaupr 
   6900 Zen Fu
   4784 J r me Charaoui
   1446 Gabriel Filion
   1146 Jerome Charaoui
    837 groente
    705 kez
    569 Gaba
    381 Matt Traudt
    237 Stephen Swift
    That, of course, is the entire history of the git repo, again. We should take only the last year into account, and probably ignore the tails directory, as sneaky Zen Fu imported the entire docs from another wiki there... 
    anarcat@angela:help.torproject.org$ find [d-s]* -type f -mtime -365   xargs -n1 git blame --line-porcelain 2>/dev/null   sed -n 's/^author //p'   sort -f   uniq -ic   sort -nr   head -10
  75037 Antoine Beaupr 
   2932 J r me Charaoui
   1442 Gabriel Filion
   1400 Zen Fu
    929 Jerome Charaoui
    837 groente
    702 kez
    569 Gaba
    381 Matt Traudt
    237 Stephen Swift
    Pretty good! 75k lines. But those are the files that were modified in the last year. If we go a little more nuts, we find that: 
    anarcat@angela:help.torproject.org$ $ git-count-words-range.py    sort -k6 -nr   head -10
parsing commits for words changes from command: git log '--since=1 year ago' '--format=%H %al'
anarcat 126116 - 36932 = 89184
zen 31774 - 5749 = 26025
groente 9732 - 607 = 9125
lelutin 10768 - 2578 = 8190
jerome 6236 - 2586 = 3650
gaba 3164 - 491 = 2673
stephen.swift 2443 - 673 = 1770
kez 1034 - 74 = 960
micah 772 - 250 = 522
weasel 410 - 0 = 410
    I wrote 126,116 words in that wiki, only in the last year. I also deleted 37k words, so the final total is more like 89k words, but still: that's about forty (40!) articles of the average size (~2k) I wrote in 2022. (And yes, I did go nuts and write a new log parser, essentially from scratch, to figure out those word diffs. I did get the courage only after asking GPT-4o for an example first, I must admit.) Let's celebrate that again: I wrote 90 thousand words in that wiki in 2024. According to Wikipedia, a "novella" is 17,500 to 40,000 words, which would mean I wrote about a novella and a novel, in the past year. But interestingly, if I look at the repository analytics. I certainly didn't write that much more in the past year. So that alone cannot explain the lull in my production here.

    Arguments Another part of me is just tired of the bickering and arguing on the internet. I have at least two articles in there that I suspect is going to get me a lot of push-back (NixOS and Fish). I know how to deal with this: you need to write well, consider the controversy, spell it out, and defuse things before they happen. But that's hard work and, frankly, I don't really care that much about what people think anymore. I'm not writing here to convince people. I have stop evangelizing a long time ago. Now, I'm more into documenting, and teaching. And, while teaching, there's a two-way interaction: when you give out a speech or workshop, people can ask questions, or respond, and you all learn something. When you document, you quickly get told "where is this? I couldn't find it" or "I don't understand this" or "I tried that and it didn't work" or "wait, really? shouldn't we do X instead", and you learn. Here, it's static. It's my little soapbox where I scream in the void. The only thing people can do is scream back.

    Collaboration So. Let's see if we can work together here. If you don't like something I say, disagree, or find something wrong or to be improved, instead of screaming on social media or ignoring me, try contributing back. This site here is backed by a git repository and I promise to read everything you send there, whether it is an issue or a merge request. I will, of course, still read comments sent by email or IRC or social media, but please, be kind. You can also, of course, follow the latest changes on the TPA wiki. If you want to catch up with the last year, some of the "novellas" I wrote include: (Well, no, you can't actually follow changes on a GitLab wiki. But we have a wiki-replica git repository where you can see the latest commits, and subscribe to the RSS feed.) See you there!

    Antoine Beaupr : Qalculate hacks

    This is going to be a controversial statement because some people are absolute nerds about this, but, I need to say it. Qalculate is the best calculator that has ever been made. I am not going to try to convince you of this, I just wanted to put out my bias out there before writing down those notes. I am a total fan. This page will collect my notes of cool hacks I do with Qalculate. Most examples are copy-pasted from the command-line interface (qalc(1)), but I typically use the graphical interface as it's slightly better at displaying complex formulas. Discoverability is obviously also better for the cornucopia of features this fantastic application ships.

    Qalc commandline primer On Debian, Qalculate's CLI interface can be installed with: 
    apt install qalc
    Then you start it with the qalc command, and end up on a prompt: 
    anarcat@angela:~$ qalc
>
    Then it's a normal calculator: 
    anarcat@angela:~$ qalc
> 1+1
  1 + 1 = 2
> 1/7
  1 / 7   0.1429
> pi
  pi   3.142
>
    There's a bunch of variables to control display, approximation, and so on: 
    > set precision 6
> 1/7
  1 / 7   0.142857
> set precision 20
> pi
  pi   3.1415926535897932385
    When I need more, I typically browse around the menus. One big issue I have with Qalculate is there are a lot of menus and features. I had to fiddle quite a bit to figure out that set precision command above. I might add more examples here as I find them.

    Bandwidth estimates I often use the data units to estimate bandwidths. For example, here's what 1 megabit per second is over a month ("about 300 GiB"): 
    > 1 megabit/s * 30 day to gibibyte 
  (1 megabit/second)   (30 days)   301.7 GiB
    Or, "how long will it take to download X", in this case, 1GiB over a 100 mbps link: 
    > 1GiB/(100 megabit/s)
  (1 gibibyte) / (100 megabits/second)   1 min + 25.90 s

    Password entropy To calculate how much entropy (in bits) a given password structure, you count the number of possibilities in each entry (say, [a-z] is 26 possibilities, "one word in a 8k dictionary" is 8000), extract the base-2 logarithm, multiplied by the number of entries. For example, an alphabetic 14-character password is: 
    > log2(26*2)*14
  log (26   2)   14   79.81
    ... 80 bits of entropy. To get the equivalent in a Diceware password with a 8000 word dictionary, you would need: 
    > log2(8k)*x = 80
  (log (8   000)   x) = 80  
  x   6.170
    ... about 6 words, which gives you: 
    > log2(8k)*6
  log (8   1000)   6   77.79
    78 bits of entropy.

    Exchange rates You can convert between currencies! 
    > 1 EUR to USD
  1 EUR   1.038 USD
    Even fake ones! 
    > 1 BTC to USD
  1 BTC   96712 USD
    This relies on a database pulled form the internet (typically the central european bank rates, see the source). It will prompt you if it's too old: 
    It has been 256 days since the exchange rates last were updated.
Do you wish to update the exchange rates now? y
    As a reader pointed out, you can set the refresh rate for currencies, as some countries will require way more frequent exchange rates. The graphical version has a little graphical indicator that, when you mouse over, tells you where the rate comes from.

    Other conversions Here are other neat conversions extracted from my history 
    > teaspoon to ml
  teaspoon = 5 mL
> tablespoon to ml
  tablespoon = 15 mL
> 1 cup to ml 
  1 cup   236.6 mL
> 6 L/100km to mpg
  (6 liters) / (100 kilometers)   39.20 mpg
> 100 kph to mph
  100 kph   62.14 mph
> (108km - 72km) / 110km/h
  ((108 kilometers)   (72 kilometers)) / (110 kilometers/hour)  
  19 min + 38.18 s

    Completion time estimates This is a more involved example I often do.

    Background Say you have started a long running copy job and you don't have the luxury of having a pipe you can insert pv(1) into to get a nice progress bar. For example, rsync or cp -R can have that problem (but not tar!). (Yes, you can use --info=progress2 in rsync, but that estimate is incremental and therefore inaccurate unless you disable the incremental mode with --no-inc-recursive, but then you pay a huge up-front wait cost while the entire directory gets crawled.)

    Extracting a process start time First step is to gather data. Find the process start time. If you were unfortunate enough to forget to run date --iso-8601=seconds before starting, you can get a similar timestamp with stat(1) on the process tree in /proc with: 
    $ stat /proc/11232
  File: /proc/11232
  Size: 0               Blocks: 0          IO Block: 1024   directory
Device: 0,21    Inode: 57021       Links: 9
Access: (0555/dr-xr-xr-x)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2025-02-07 15:50:25.287220819 -0500
Modify: 2025-02-07 15:50:25.287220819 -0500
Change: 2025-02-07 15:50:25.287220819 -0500
 Birth: -
    So our start time is 2025-02-07 15:50:25, we shave off the nanoseconds there, they're below our precision noise floor. If you're not dealing with an actual UNIX process, you need to figure out a start time: this can be a SQL query, a network request, whatever, exercise for the reader.

    Saving a variable This is optional, but for the sake of demonstration, let's save this as a variable: 
    > start="2025-02-07 15:50:25"
  save("2025-02-07T15:50:25"; start; Temporary; ; 1) =
  "2025-02-07T15:50:25"

    Estimating data size Next, estimate your data size. That will vary wildly with the job you're running: this can be anything: number of files, documents being processed, rows to be destroyed in a database, whatever. In this case, rsync tells me how many bytes it has transferred so far: 
    # rsync -ASHaXx --info=progress2 /srv/ /srv-zfs/
2.968.252.503.968  94%    7,63MB/s    6:04:58  xfr#464440, ir-chk=1000/982266)
    Strip off the weird dots in there, because that will confuse qalculate, which will count this as: 
      2.968252503968 bytes   2.968 B
    Or, essentially, three bytes. We actually transferred almost 3TB here: 
      2968252503968 bytes   2.968 TB
    So let's use that. If you had the misfortune of making rsync silent, but were lucky enough to transfer entire partitions, you can use df (without -h! we want to be more precise here), in my case: 
    Filesystem              1K-blocks       Used  Available Use% Mounted on
/dev/mapper/vg_hdd-srv 7512681384 7258298036  179205040  98% /srv
tank/srv               7667173248 2870444032 4796729216  38% /srv-zfs
    (Otherwise, of course, you use du -sh $DIRECTORY.)

    Digression over bytes Those are 1 K bytes which is actually (and rather unfortunately) Ki, or "kibibytes" (1024 bytes), not "kilobytes" (1000 bytes). Ugh. 
    > 2870444032 KiB
  2870444032 kibibytes   2.939 TB

    > 2870444032 kB
  2870444032 kilobytes   2.870 TB
    At this scale, those details matter quite a bit, we're talking about a 69GB (64GiB) difference here: 
    > 2870444032 KiB - 2870444032 kB
  (2870444032 kibibytes)   (2870444032 kilobytes)   68.89 GB
    Anyways. Let's take 2968252503968 bytes as our current progress. Our entire dataset is 7258298064 KiB, as seen above.

    Solving a cross-multiplication We have 3 out of four variables for our equation here, so we can already solve: 
    > (now-start)/x = (2996538438607 bytes)/(7258298064 KiB) to h
  ((actual   start) / x) = ((2996538438607 bytes) / (7258298064
  kibibytes))
  x   59.24 h
    The entire transfer will take about 60 hours to complete! Note that's not the time left, that is the total time. To break this down step by step, we could calculate how long it has taken so far: 
    > now-start
  now   start   23 h + 53 min + 6.762 s
> now-start to s
  now   start   85987 s
    ... and do the cross-multiplication manually, it's basically: 
    x/(now-start) = (total/current)
    so: 
    x = (total/current) * (now-start)
    or, in Qalc: 
    > ((7258298064  kibibytes) / ( 2996538438607 bytes) ) *  85987 s
  ((7258298064 kibibytes) / (2996538438607 bytes))   (85987 secondes)  
  2 d + 11 h + 14 min + 38.81 s
    It's interesting it gives us different units here! Not sure why.

    Now and built-in variables The now here is actually a built-in variable: 
    > now
  now   "2025-02-08T22:25:25"
    There is a bewildering list of such variables, for example: 
    > uptime
  uptime = 5 d + 6 h + 34 min + 12.11 s
> golden
  golden   1.618
> exact
  golden = ( (5) + 1) / 2

    Computing dates In any case, yay! We know the transfer is going to take roughly 60 hours total, and we've already spent around 24h of that, so, we have 36h left. But I did that all in my head, we can ask more of Qalc yet! Let's make another variable, for that total estimated time: 
    > total=(now-start)/x = (2996538438607 bytes)/(7258298064 KiB)
  save(((now   start) / x) = ((2996538438607 bytes) / (7258298064
  kibibytes)); total; Temporary; ; 1)  
  2 d + 11 h + 14 min + 38.22 s
    And we can plug that into another formula with our start time to figure out when we'll be done! 
    > start+total
  start + total   "2025-02-10T03:28:52"
> start+total-now
  start + total   now   1 d + 11 h + 34 min + 48.52 s
> start+total-now to h
  start + total   now   35 h + 34 min + 32.01 s
    That transfer has ~1d left, or 35h24m32s, and should complete around 4 in the morning on February 10th. But that's icing on top. I typically only do the cross-multiplication and calculate the remaining time in my head. I mostly did the last bit to show Qalculate could compute dates and time differences, as long as you use ISO timestamps. Although it can also convert to and from UNIX timestamps, it cannot parse arbitrary date strings (yet?).

    Other functionality Qalculate can:
    • Plot graphs;
    • Use RPN input;
    • Do all sorts of algebraic, calculus, matrix, statistics, trigonometry functions (and more!);
    • ... and so much more!
    I have a hard time finding things it cannot do. When I get there, I typically need to resort to programming code in Python, use a spreadsheet, and others will turn to more complete engines like Maple, Mathematica or R. But for daily use, Qalculate is just fantastic. And it's pink! Use it!

    Further reading and installation This is just scratching the surface, the fine manual has more information, including more examples. There is also of course a qalc(1) manual page which also ships an excellent EXAMPLES section. Qalculate is packaged for over 30 Linux distributions, but also ships packages for Windows and MacOS. There are third-party derivatives as well including a web version and an Android app.

    Updates Colin Watson liked this blog post and was inspired to write his own hacks, similar to what's here, but with extras, check it out!

    Next.