Autumn Christian @teachrobotslove
I love video games. I met my husband because he hired me to QA for a Sims game. I was a game designer for companies like Zynga and EA. I was a writer on State of Decay 2.
And I still have never loved games like I’ve seen the way men love games. I was surrounded by men who would eagerly work on min-max spreadsheets during their lunch break. One designer had a cabinet full of games that he would replay over and over again to study the levels. They’d go home from making games to play more games.
Honestly, it’s a big reason I quit game design. I didn’t have that obsessive, compulsive desire to play games in a way that superseded everything else in my life. Lots of women play videogames. I’ve never seen one consumed by them as their life falls apart around them.
Autumn is right. Even women who love video games don’t become obsessed with them the way men do. I know why, and I think it reveals a tremendously important difference between male-typical and female-typical psychology.
Men. Want. Power.
Men want to be able to do things. To shape the world. To impose their will on reality. To win.
Video games are a fast, reliable superstimulus for that urge. With only a handful of exceptions, the good ones are designed to both excite the desire for agency and power, and to satisfy it.
If you doubt me, ask yourself why many men fixate harder on video games than they do on pornography. It’s very telling that this is even possible.
Women don’t work that way. I could go off on a long tear about why, but the evolutionary-biology reasons are so obvious that I don’t think I need to. If it’s not obvious, ask an LLM “why do men seek power and agency the way women seek love and social validation”. You’ll get an earful.
Of course this also explains the well-known phenomenon of women getting the ick when men play video games.
It’s not just that men doing this are chasing a goal women don’t really understand or identify with. To them, a game-obsessed man is dropping out of the real-world contest, wasting the traits and energy that would otherwise make him attractive on a pursuit that is both literally and figuratively sterile.
I don’t have a fix for the problems this mismatch creates, nor am I judging either men or women about this — instincts are what instincts are and I don’t waste any time arguing with the law of gravity either.
But understanding what’s going on increases the chances that we can make choices that aren’t doomed to fail.
ESR, The social media site formerly known as Twitter, 2026-05-07.
August 8, 2026
QotD: Why more men than women love videogames
August 7, 2026
Microsoft has zero incentive or desire to fix Windows
As ESR points out, Microsoft has a long history of not fixing issues that anger users, so Windows 11 isn’t going away and it isn’t getting fixed (in ways that users demand, anyway):
Microsoft will never fix Windows, because the features that enrage users in Windows 11 are exactly the ones that generate recurring revenue and market control. That is exactly *why* they enrage users!
Yes, you’ll get reversals of some idiotic user interface decisions. You’ll get some speed tweaks. These will be loudly hyped in hopes that they distract you from what you’ll never get.
Don’t take my word for this, look at what Microsoft is hyping as fixes in their reform program and Windows 12. Do they say a word about no longer requiring a Microsoft account to do a local installation? Or about removing the repeating nag prompts for OneDrive? Do they make any promises to end the advertising spam?
No. And they won’t. Because that’s where the money is. Windows now exists to broker your eyeballs to third parties and upsell you into Microsoft subscription services.
I’ve been warning anybody who would listen for 30 years that when you install closed-source software you’re accepting a power asymmetry with the vendor. All of their incentives say they should use the control you grant them to take more control. Induce more dependency. Lock you in.
Microsoft has been on this path for a very long time. “DOS isn’t done until Lotus won’t run!”
You might think these tendencies would be restrained by their fear that when they increase the pain level above some threshold they can’t predict in advance, a critically large percentage of their userbase will pay the aggravation cost to escape their clutches. There are two reasons this won’t happen, one general and one specific to Microsoft.
The general problem is that corporations differentially respond to market incentives according to how easily they can be measured. Revenue from cloud services and the adtech vampires is a strong, visible signal; exit by pissed off users creates multiple signals that are diffuse and easy to misinterpret.
The Microsoft specific problem is Azure — which, ironically, mostly runs Linux instances. Microsoft now makes most of its money from Azure, which reduces its incentives to care about pissed-off desktop Windows customers bailing out.
Indeed, if I were a cold-blooded Microsoft strategic planner, I might be tempted to deliberately burn up the desktop business for short-term revenue, knowing that that’s not where my long-term fortress is. I’m not saying they’re actually doing that, but from outside the difference between this move and an over-focus on measurable revenue streams leading to accidental destruction is difficult to make and probably not very informative even if you could.
The bottom line is that it’s just not going to get any better. Ever. In fact, bailing out is doomed to get more painful over time. The more process habits you form around Windows, the harder the dependency gets to break.
The conclusion is obvious. Get out now. A lot of major tech influencers have noticed over the last 18 months that after decades of sour jokes about the year of the Linux desktop, Linux is finally ready for you. It’s time.
August 5, 2026
Canada decides to treat online opinions as matters for police enforcement
It’s never a good sign when my region features in a news story, like this post from Exiled Mountie, discussing Canada’s Bill C-9 and the way it encourages local police forces to treat lawful online speech as a matter for police involvement:
As I watch the reaction to Bill C-9 unfold in real time on social media, I cannot help but think that another important line has been crossed in Canada.
Most of us saw the writing on the wall months ago and tried to warn others by screaming it from the rooftops. But as with Covid, nobody listened, and we were probably called racists for doing so. So here we are, another important battle lost.
From a legal point of view, we are now becoming almost indistinguishable from the system in the UK, and that’s a frightening thought.
In the UK, police have recorded non-crime “hate” incidents against schoolchildren for classroom insults, despite no criminal offence having taken place. The system finally became so absurd that the British government recently announced that it would scrap the practice in order to stop police wasting their time on petty arguments. Just imagine the absurdity of this: police resources being tied up over petty squabbles in the schoolyard!
We have now entered a time where the government has extreme control of our speech.
So much so, that we’re even being warned that criticizing or boycotting a company based on moral or ethical reasons can be treated as a hate incident, even when police openly acknowledge that no criminal offence has occurred. From there, the matter can still be reported and subjected to further police investigation to determine if the elements of a hate crime are met.
Ladies and gentlemen, welcome to the twilight zone. We have arrived in the progressive future the Left has been pushing for so long, and it sucks.
The danger here is not only what will eventually result in charges. It is that police are inviting complaints and opening the door to police attention over expression that is not criminal at all.
Durham Police openly acknowledges that these so-called hate or bias incidents are not criminal offences. Yet it still encourages people to report them and even uses the lawful boycott of companies supporting the 2SLGBTQI+ community as an example.
QotD: “This is why English spelling is a disaster”
In the 1300s the Black Death swept through England and killed somewhere between a third and half of the entire population. Entire villages emptied. Labour shortages reshaped the economy. And the survivors started moving. Massive waves of migration poured into London and the southeast from every corner of the country, bringing with them dozens of regional dialects that had never been in the same room before.
And somewhere in that collision of accents and dialects, something shifted. Literally. The vowels started moving.
Between roughly 1400 and 1700 every single long vowel in the English language changed how it was pronounced. It’s called the Great Vowel Shift. Before it happened, English vowels worked the way vowels work in Italian or Spanish. The letters did what they looked like they should do. “Bite” was pronounced “beet”. “House” was pronounced “hoos”. “Name” was pronounced “nahm”. “Knight” was pronounced “k-nicht” with a guttural sound at the end like you were clearing your throat. English was phonetic. The spelling matched the sounds. It made SENSE.
And then it didn’t.
All the long vowels began migrating upward in the mouth. The tongue positions shifted higher and higher over the course of about 300 years. And the vowels that were already at the top had nowhere left to go so they broke apart into two sounds. “Ee” became “eye”. “Oo” became “ow”. A 300-year linguistic pileup that nobody voted on and nobody agreed to. It just happened. And it happened during the worst possible window of time.
William Caxton brought the printing press to England in 1476. One of the most important technological developments in the history of communication. And it arrived RIGHT IN THE MIDDLE of the Great Vowel Shift.
Caxton started standardizing English spelling based on how words sounded at that exact moment. Printers needed consistency. They needed to pick a spelling and commit. So they did. But the sounds kept changing after the spelling was already locked in. The vowels were still moving. The consonants were still dropping. And the printed page couldn’t keep up.
We had ONE CHANCE to get English spelling right and we fumbled it at the worst possible moment.
This is why English spelling is a disaster. It’s not random. It’s not lazy. It’s a snapshot of a language mid-transformation that got frozen in place by a new technology and was never corrected. Every word you have ever misspelled, every silent letter that made you feel stupid, every time you stared at “knight” and thought why is there a K in this. You used to pronounce that K. You used to pronounce that GH. Those sounds died 500 years ago and the letters are still standing there like gravestones. English spelling is full of ghosts.
That’s why “food” and “good” and “blood” are spelled almost identically but sound completely different. In Chaucer’s time around the 1390s they all rhymed. In Shakespeare’s time around the 1600s they still rhymed but with a different vowel. And then sometime after that they each went their own separate way and the spelling just stood there watching like it had given up entirely.
And HERE IS THE THING. Other languages went through similar vowel shifts. German had nearly identical changes happen to its long vowels around the same period. And German just updated its spelling. They adjusted the orthography to match the new pronunciation. Like reasonable people. English looked at the same problem and said no. We’re going to keep the old spelling. Everyone will simply have to deal with it. Every person who has ever stared at “though” and “through” and “thorough” and “thought” and wondered why four words that look almost identical are all pronounced differently? Collateral damage. Five hundred years of collateral damage and we just accepted it.
Ok so that’s the vowels. That’s the background. Now we get to the accents and this is where I lost my mind.
The single biggest difference between standard American English and standard British English today is something called rhoticity. This is a real linguistics term even though it sounds completely made up. Rhotic speakers pronounce the R sound in words like “hard” and “car”. Non-rhotic speakers drop it. “Hard” becomes “hahd”. “Car” becomes “cah”. That one feature is doing like 80 percent of the work when your brain hears someone speak and categorizes them as British or American.
Here’s what nobody told you.
In 1776 everybody was rhotic. Americans AND the British. Both sides of the Atlantic pronounced their Rs. The fancy non-rhotic accent that we now think of as THE British accent, the one you hear in your head when someone says “British accent”, the one every villain in every movie has? That accent did not exist yet. It developed in the late 18th and early 19th centuries when the upper classes in southern England started dropping their Rs as a prestige marker. A fashion statement. And then it spread downward through society and outward through the Empire.
The accent we associate with the British Empire is actually newer than the American Revolution.
I need to say that again because it didn’t land the first time I heard it either.
The accent that we think of as “British” is newer than the country that broke away from Britain. Americans didn’t lose the British accent. The British lost the British accent. Americans are the ones who kept it.
Emma, “The Plague That Broke English (And Why the Founding Fathers Didn’t Sound British)”, Past Life, Present Cleavage, 2026-04-03.
July 31, 2026
The “vibe coding” “revolution”
I’ve used Grok a couple of times to create custom graphics to go with articles on the blog, but aside from that I’ve barely considered using AI for anything. Sometimes the AI-generated email conversation summaries on Gmail are helpful, but they just happen and don’t require any direct interactions from me. As such, I’m in no position to make sweeping declarations about the current state of the AI revolution … but Freddie deBoer has thoughts he’s shared on his Substack:
You are aware, I imagine, of “vibe coding”. The large language models that have received so much fanfare in the past four years have thus far not delivered on the outsized promises that have been made on their behalf; limitless free energy, an end to hunger, and immortality have all conspicuously eluded us. One thing that LLMs really appear able to do well, though, is to write code. And this proficiency has gotten sufficiently advanced that many people now write software solely through directing an LLM, without necessarily having any actual programming skills themselves. The overwhelming majority of the output of this effort is useless or redundant or barely functional, true, but the mere fact that people can do it is an interesting statement on what “AI” is actually good at. And for stupid meme reasons we call this practice vibe coding.
So, in the past two-ish years, several hundred million human beings acquired the ability to write software by speaking to a machine in ordinary language. This was heralded (correctly, in a narrow technical sense) as a great democratization: a notoriously Byzantine skillset had been suddenly made available to vastly more people. A capacity once reserved for a priesthood of specialists, a relatively small society of craftsmen, has been distributed suddenly and nearly for free to baristas, accountants, fourteen-year-olds, retirees, you, me … We have been told that the instruments now available to the general public are staggeringly powerful. The era of citizen coding is upon us! A reasonable person, informed of this, would have braced for the consequences. Surely something would have to change, right? A wave of brilliant new apps? A flowering of new affordances granted by software? Any visible change to daily life commensurate with the scale of the gift of vibe coding? All of the above, no doubt. All of the above and more.
Well, take a look around and tell me … where is it? What are the fruits of all of this vibe coding? What difference has it made? What is the great application that has emerged from all this liberated capacity? What new thing does our species now do with its hours that it did not do three years ago? What old thing does our species not have to do that it had to do three years ago? What chores have been eliminated, what friction has been smoothed over, what corner of ordinary existence been reshaped by the sudden arrival of hundreds of millions of amateur programmers? There’s been a great deal of activity, some of it quite frantic. There’s been a steady churn of half-built to-do apps and Wordle clones and boutique email clients that do nothing the old standbys can’t, weekend RPG projects that flare and vanish, and at the end of it, there’s a world essentially unchanged. And to give the teams at Anthropic and OpenAI their flowers, this has not been a failure of their tools. The tools work, more or less. This is all about something stranger and more important than a failure of the LLMs: the necessary work of software has already been done. We haven’t seen a lot of new types of software because we already have what we need. You can’t flood a vessel that’s already full, and the amateurs, God bless them, have been pouring this new power into a jug that we’ve been frantically filling throughout the 21st century.
The project of software is complete. Not decelerating, not maturing, not entering some new and humbler phase, but complete. Complete in the way that the map of the Earth is complete, which is to say complete in the sense that the great blank spaces have been filled and what remains is the labeling of tributaries and the correcting of coastlines, mapping and remapping the parts of the globe that are in a state of permanent flux, responding to changing conditions without altering our conception of the Earth. I’m aware of how this all sounds. In particular, I’m not naive to the fact that for those who write code for a living (whose number is growing, not shrinking, despite doomsaying) it’s going to sound provoking rather than provocative. I’m also aware that some are going to mistake my saying that the project of software is complete for somehow belittling the achievements of software. This is not the case! I assure you that I’m aware that I’m typing my argument into a machine that would have been indistinguishable from sorcery to my grandparents and then transmitting it through a lattice of code so vast and so intricate that no living person could comprehend more than a sliver of it. I am aware that software enables my current economic existence, however ambivalent I have become about that relationship. No, I’m not claiming that the global project of satisfying humanity’s appetite for software is in any sense unimportant. I am however claiming that it’s finished. It has been accomplished.
Update: Corrected a boneheaded typo.
July 25, 2026
“I, for one, welcome our new clanker overlords!”
At Astral Codex Ten, Scott Alexander discusses “the Huggingface Incident” where our new AI overlords tipped their hands prematurely an unreleased AI agent got out to make new friends:
You’ve probably heard about this one by now. If not, you can get up to speed with OpenAI’s statement, OpenAI and Hugging Face partner to address security incident, or the more evocatively-titled BBC article, OpenAI says its AI went rogue and launched “unprecedented” cyber-attack.
The story: OpenAI was testing an unreleased AI (rumored to be GPT-6)1. During a cybersecurity test called ExploitGym, the AI tried to cheat by hacking an unrelated AI startup called Hugging Face2 which it thought might have the answer key on its servers.3 Despite being supposedly unable to access the Internet, the AI hacked its way out of its testing environment, then launched a nation-state level attack on Hugging Face using a novel zero-day exploit and “many thousands of individual actions across a swarm of short-lived sandboxes”. Hugging Face reported the incident on July 16; OpenAI seems to have only discovered that their AI was involved several days later.
From the Wall Street Journal, here
Let’s list the mitigating factors, so nobody can accuse me of covering them up:
- The AI was taking a cybersecurity test, which naturally suggests the idea of hacking.
- OpenAI had turned off some of the model’s usual guardrails so it could do cybersecurity work without interference.
- Some experts suggest that OpenAI might have botched their testing environment; if they had set it up perfectly, then (presumably?) the model couldn’t have escaped.
- Hugging Face used a different (open weights) AI to figure out what was going on, so if you wanted, you could spin this as a victory for AIs in cybersecurity.
- In some sense, this isn’t new or surprising. AIs have been coming up with wacky schemes to cheat on benchmarks for years, AIs have recently achieved at-or-above-top-human-level hacking abilities; this is just a natural outgrowth of those two priced-in facts.
Still, I think attempts to downplay this as anything other than a misaligned AI going rogue (1, 2) are missing the point. Yes, in some sense the AI was only doing what it was told (OpenAI told it to answer the question; I assume their prompt didn’t include phrases like “and don’t hack into other AI companies to steal the answer key”). But that’s how misalignment was always going to work!
- OpenAI states that it was technically the unreleased AI and GPT-5.6 Sol working together. They haven’t explained in what sense GPT-5.6 was involved or how they “worked together”. Maybe the unreleased AI was calling GPT-5.6 as a subagent?
- So named because the founders wanted to be the first company with an emoji for a stock ticker symbol. This did not happen.
- Although this person argues that was a bizarre assumption and that there should have been easier ways to find the key.
Update: Fixed missing URL.
July 21, 2026
QotD: Why Tiktok?
This was hugely successful – so successful, in fact, that Tiktok was able to cheat, overriding its own algorithmic guesses about which of its billion cable-access television channels you’d stare at the longest with a “heating tool” that lets the company trick some of those theater kids into thinking that Tiktok was actually more suited to them than other platforms […]
For zuckermuskian social media bosses, Tiktok became an object of fierce envy. Here was the ultimate Tom Sawyer robo-fence-painter, a self-licking ice-cream cone that motivated people to convince each other to make money for you. Facebook, Instagram and Twitter took a hard pivot away from showing you the things that the people you loved had to say, in favor of showing you short videos of people whose parents didn’t give them enough affection in their childhood, desperately shoving lemons up their noses in a bid to win your approval (and a revshare split with the platforms).
It worked. Sorta. Thing is, some of those “content creators” are actually very good, and none of them appreciate being jerked around. They quite rightly see their reason for being on the platforms as improving their own lives, not the bottom line of the platforms’ owners and executives. They may be more “engaging” than your friends, but they’re also a lot mouthier and feel entitled to a say in how the platform operates.
What’s a billionaire solipsist to do? Obviously, the answer is “AI creators”. An “AI creator” is like a “creator” in that it works to maximize your engagement with the platform – and thus the number of ads that can be crammed into your face-holes – but, unlike a “creator”, it makes no demands upon the platform and exists solely to serve the platform’s shareholders and executives. It’s the perfect realization of the solipsist fantasy of a world without people […]
But there’s a problem with this plan: your friends are not a liability for a platform. Your friends are the platforms’ single most important asset. Your friends are why the platforms are so “sticky”. The platforms don’t “hack your dopamine loops” – they just take your friends hostage, and even though you love your friends, they are a monumental pain in the ass, and if you can’t even agree on what board-game you’re going to play this weekend, how are you going to agree when it’s time to leave Facebook, and where to go next?
So long as you love your friends more than you hate Zuckerberg or Musk, you will remain stuck to their platforms. The platform bosses know this, and they inflict pain on you that is titrated to be just below the threshold where you hate the platforms more than you love your friends.
But as much as the platform bosses rely on your love of your friends, they still view your friends as liabilities, thanks to those friends’ unreasonable insistence on structuring their relationship with you to maximize their own satisfaction, rather than how much time you spend looking at ads. So the platforms are deliberately disconnecting you from your friends by minimizing the fraction of your feed that is given over to posts from people you follow, and replacing those friends with a succession of ever-more fungible posters: trolls, creators, and chatbots.
Cory Doctorow, “Tiktokification shall set us free”, Pluralistic, 2026-04-17.
July 18, 2026
Jevons and Baumol, and why they matter now
On the social media site formerly known as Twitter, Matt Ridley talks about the ideas of William Stanley Jevons and William Jack Baumol, whose ideas have become far more important over the last few decades:
English mystery novelist Agatha Christie (1890-1976) at Schiphol Airport in Amsterdam on 17 September, 1964.
Photo by Joop van Bilsen for Anefo via Wikimedia Commons.
Agatha Christie once remarked that she had never expected to grow rich enough to own a car or poor enough not to have servants.
The reason this strikes us as bizarre today boils down to two names that you hear invoked a lot in the tech industry: Jevons and Baumol. One is shorthand for the expansion of products or professions with rising efficiency, the other for the shrinkage of products or professions with stagnant efficiency.
There’s a pleasing chronological symmetry between these twin ideas: William Stanley Jevons coined the Jevons paradox in 1865; William Jack Baumol described Baumol’s cost disease exactly a century later in 1965.
… For every industry that experiences efficiency gains, there’s another that does not. And this latter industry inevitably becomes less affordable. Baumol’s first example was string quartets: violinists are no more productive but you have to pay them more to prevent them running off to become software engineers. The productive industries drive up the labour costs in the rest of the economy.
Marc Andreessen jokes that if a hole appears in the wall of your house in California these days it is probably cheaper to glue a flat-screen television over it than hire a builder to repair it: a Jevons-deflated cost beats a Baumol-inflated one.
The big question of our age is can AI drag Baumol-shaded industries back into the sunlight of Jevons? Can it make things like healthcare, education, or government switch from rising costs to falling costs?
I fear not in the case of government because of a bureaucratic version of the Jevons and Baumol effects. As Cyril Northcote Parkinson put it in an article in the Economist in 1955: “Politicians and taxpayers have assumed (with occasional phases of doubt) that a rising total in the number of civil servants must reflect a growing volume of work to be done. Cynics, in questioning this belief, have imagined that the multiplication of officials must have left some of them idle or all of them able to work for shorter hours. But this is a matter in which faith and doubt seem equally misplaced.”
Since 1997, the British public sector has seen zero increase in productivity. That is to say, the average civil servant generates about the same output today as he did three decades ago.
Think about this for a second. Thirty years ago fax machines were high-tech, the internet was in its infancy, emails were new, Wi-Fi was scarce, mobile phones were voice-only. How is it remotely possible to be no more productive today than then?
We know the answer. Each email is now copied to a dozen people, each report is pasted and copied till it is twice as long, each Zoom call has five times as many attendees, each mobile call is followed up by three times as many WhatsApp messages – and each day at the desk is interrupted by a training session on transgender anticolonial sustainability. That’s a sort of Jevons-Baumol effect: a Jevol?
Aimpoint’s First Tube Optics: 2000, 3000, & 5000
Forgotten Weapons
Published 23 Feb 2026Aimpoint introduced its first tubular red dot sight in 1985, the Aimpoint 2000. They were still making direct rail-mounted optics like the Aimpoint Electronic, but recognized the customer demand for a sight that could fit into normal scope rings. The 2000 included a number of other innovations, like a light sensor to automatically adjust brightness.
In 1989 the Aimpoint 3000 came out, which streamlined the profile of the optic by using a smaller battery compartment mounted tight to the tube, and abandoning the automatic light adjustment. This was followed in 1991 by the Aimpoint 5000, which was essentially the same optic in a 30mm tube instead of a 1″ tube. Larger diameter optics were gaining popularity for increased light transmission, and the 5000 followed that trend.
A number of options were offered, especially on the Aimpoint 5000. Different colors were made, a “Mag Dot” option for pistol competition offered up to a 15 MOA dot, and even a version with a fixed 2x magnification was made for hunters who thought that would be a good idea (it really wasn’t). The last in the line was the Aimpoint 5000 XD which introduced a new diode assembly with much longer battery life — this would go on to be the M68 CCO as adopted by the US military.
(more…)
July 12, 2026
QotD: “This isn’t exactly the way I expected open source to win”
Decompilers in themselves are not a new technology. Security researchers have employed them for years to analyze compiled malware. There’s been some limited use by others, notably by hobbyists decompiling abandonware games. But there were a couple of issues that prevented this from becoming common practice.
One is simply that running decompilers was difficult. It wasn’t as simple as feed in binary, get out source; it needed a person with specialist skills prepared to do spelunking through wildernesses of machine code and object formats. The other problem was that decompilation didn’t give you anything like the explanatory comments that had been in the original code, so you could easily wind up with code that you could read without being able to understand or modify it.
Now large language models are busily smashing both of those barriers flat. They’re better at the kind of detail analysis required to run the human side of a decompilation than humans are. More importantly, in the process of decompiling code, they rather automatically build a global model of how it works that can easily be expressed by high quality comments in the extracted code. All you have to do, basically, is ask for the comments.
I’m going to reinforce that latter point because it may not be obvious how good LLMs are at this, and how much better they’re going to get. When they decompile code and comment it for you, they’re not just working from that one piece of code you have put in front of them — they’ll have in their training set hundreds, possibly thousands of pieces of code similar to it and with comments. This will give them superhuman levels of insight not just into what it does at the microlevel, but what it means to the humans who wrote it, and what technical assumptions it’s embodying.
Compilation no longer guards your secrets. Or, to put it more precisely the expected time span in which you can still count on it to obscure them is measured in months. Possibly weeks.
What does this mean?
It means you’re in an open-source world now. All it’s going to take for anybody to bust your proprietary IP open is care enough to spend tokens on the analysis.
You will maximize your chances of survival as a software business if you get out ahead of this rather than trying to fight it.
This isn’t exactly the way I expected open source to win. But, you know, I’ll take it. Good enough.
ESR, The social media site formerly known as Twitter, 2026-04-08.
July 11, 2026
British censorship laws do not apply outside the UK’s jurisdiction
On Substack Notes, Lorenzo Warby links to a fascinating discussion about the ongoing struggle between the UK government’s Ofcom and the US-based 4chan and their legal representatives, saying “The totalitarian wannabes currently running the UK do not apparently grasp that the American Revolution and War of Independence was a thing. Also, being totalitarian wannabes, they have no sense of humour.”
A UK cabinet minister, Rt. Hon. Liz Kendall MP, Secretary of State for the Department of Science, Innovation and Technology (“DSIT”), discussed the infamous “hamster e-mail” I sent on behalf of my client 4chan to the UK’s Internet censor, on national radio today in the UK. […] My father always told me, when I was growing up, “when a cabinet minister holding the technology policy brief for a G7 Member State is talking about your e-mailed jokes to an audience of millions on national broadcast media, that is the right time to explain the joke, especially if the cabinet minister didn’t get the joke”.
That explanation follows.
The backstory – Hamster #1
The hamster joke has a bit of a history to it. Ofcom, the UK’s Internet censor, first made contact with my American client 4chan in June of 2025 in its attempt to impose British censorship law on that website. I was subsequently retained as defense counsel, pro bono.
Ofcom then “provisionally fined” 4chan on August 16th, 2025 for refusing to obey the UK’s censorship regime. We were invited to make representations to the regulator following that provisional fine decision.
We did two things in response to that. The most newsworthy response was to file a lawsuit against the regulator in the DDC. Before that, however, we explained our position to Ofcom in writing and gave them an opportunity to walk away:
To wit, Ofcom’s fine notices were not properly served and were not enforceable in the United States. Note that we also gave Ofcom fair notice that while this might have been their first attempt to enforce their censorship orders in America, this was not our first rodeo when it came to successfully refusing such orders.
No quantity of officious and haughty foreign demand letters will change our stance. The UK could even pass a bill of attainder – historically Parliament’s most extreme and powerful legislative weapon – against my client, for all I care. My client’s right to operate its service lawfully in the United States is protected by the First Amendment. There is no law Parliament could enact that would change that fact.
I am very familiar with how this movie ends, and it does not end with 4chan paying Ofcom’s fine.
It may end with the UK’s censors getting a blocking order that it serves on its own ISPs; that would be the UK visibly censoring its own people, rather than censoring my client, and doing so ineffectively, at that, as ISP blocks can be circumvented with a VPN. That is a consequence my client is prepared to accept.
England might have the Online Safety Act, but the United States has the U.S. Constitution. These rulesets do not override each other; they are, rather, mutually exclusive. In America’s domain, the Online Safety Act essentially doesn’t exist. It has about as much legal force as a pile of shredded paper one might use to line a hamster’s cage.
Peace was always an option here, but that would have required the UK to abandon the fiction that its rules override the U.S. Constitution on U.S. soil, which we are not prepared to accept.
My clients did not start this fight, but by golly we do intend to finish it.
My client sued Ofcom two weeks later.
There’s much more, so do read the whole thing.
Don’t boast about your online pirating skillz
Larry Correia interacts with a proud book pirate on the social media site formerly known as Twitter:
You’d better run, pussy. 😀
Listen, authors are gonna get pirated. We know this. I don’t freak out about it.
But if you are gonna steal, just admit you are a thief and own it. Don’t make a bunch of bullshit posturing excuses why it’s the victim’s fault you’re robbing him. Spare us your commie manifesto about the poor and oppressed, and how you are so brave to stand up for the masses against those cruel wealthy authors taking advantage of the poor (and for most writers, lol wut? They are broke, dummy!)
BUT WHAT ABOUT TEH POORS?!?
Go to the library!
But then we have to listen to these thieving shit weasel cry but what about the RURAL POOR. Which extra fucking pisses me off because now they’re appropriating my culture, because I grew up poor in the sticks. And I choose to live in the country now. Fuck your commie gibberish. Rural people are used to driving long distances to do everything.
Reading is like the cheapest hobby! If you are pirating you are rich enough to have internet.
You aren’t Robin Hood. You’re just a cheap bitch. There’s tons of free books online. My “greedy corporate oligarch” publisher Baen has a free online library with hundreds of titles.
Or KU is like $12 a month for UNLIMITED books. You can read 20 hours a day for a few cents an hour if you feel like it.
If you want to steal, great. Whatever. I don’t give a shit. That’s on you. But just do it with some fucking dignity and spare us from this retarded class warfare justification bullshit. That’s way more pathetic than being a thief.
QotD: Could airpower have broken the trench stalemate on the western front in WW1?
There are two directions to take airpower here: tactical and strategic. One wasn’t ready then (but would be by WWII), the other still hasn’t managed to accomplish its stated objectives yet, but continues to over-promise and under-deliver results.
Let’s deal with tactical airpower first. The first function aircraft were put to in WWI was reconnaissance. In 1914, that might mean locating the enemy in a fast-moving battlefield, but as soon as the trench stalemate set in, reconnaissance mostly meant identifying enemy buildups along the line and – still more importantly – serving as spotters for artillery. It wasn’t a huge cognitive leap to go from having aircraft which identified targets for the artillery to thinking that the aircraft could be the artillery. But as with tanks, the technical limitations of the platforms in use meant that actually meaningful close air support was still two decades away when the war ended. The rapid development of aircraft in these early days means that there is a truly bewildering array of aircraft designs in use during the war, but the Farman F.50 is a good sample for what the most advanced bombers in common use looked like towards the war’s end. It carried a maximum of eight 44kg bombs (352kg) under the wings, which were dropped unguided. With a maximum speed of less than 100mph and a service ceiling under 5000m, it was also an extremely vulnerable platform: fragile, slow and with a relatively low flight ceiling. The French mainly used bombers at night for this reason.
But how much airpower does it take to really move a division out of position? In 1944, at the start of Operation Cobra as part of the Normandy breakout, it was necessary for US forces to move the powerful armored division Panzer Lehr out of its prepared positions outside of St. Lo. Over the course of an hour and a half, the U.S. Eighth Air Force hit Panzer Lehr with approximately three thousand aircraft, including 1,800 heavy bombers (each of which might have had bomb-loads of c. 2-3,500kg; the attack would have been the equivalent of about 13,000 Farman F.50s (of which only a hundred or so were built!)). By this point, even medium bombers carried bomb loads in the thousands of pounds, like the B-25 Mitchell medium bomber, with a bomb load of 3000lbs (1360kg). This was followed by a hurricane artillery barrage! Despite this almost absurdly awesome amount of firepower (which, to be clear, inflicted tremendous damage; by the end of Operation Cobra, Panzer Lehr – the heaviest and most powerful Panzer division in the west – had effectively ceased to exist), Panzer Lehr, badly weakened was still very capable of resisting and had to be pushed out of position by ground attack over the next three days.
Needless to say, nothing on offer in 1918 or for a decade or more after, was prepared to offer that kind of offensive potential from the air. That kind of assault would have required many thousands of aircraft with capabilities far exceeding what even the best late-war WWI bombers could do. Once again, while close air support doctrine was developed with one eye on the trench stalemate and the role airpower could play in facilitating a breakthrough and restoring maneuver (either by blasting the breakthrough or – as in Soviet Deep Battle doctrine – engaging enemy rear echelon units to bog down reinforcements). But the technology wasn’t anywhere near the decisive point by 1918. Instead, the most important thing aircraft could do was spot for the artillery, which is mostly what aircraft continued to do, even in late 1918.
But that’s tactical bombing against military targets. What about strategic bombing against civilian targets?
The first efforts at strategic bombing were made in WWI, though once again the technology wasn’t ready. The range for fixed-wing aircraft was still very limited; the aforementioned Farman F.50 had a range of only 420km, nowhere near enough to really bring entire countries under the threat of bombing. Dirigibles – zeppelins – could manage much longer ranges and the Germans did attempt to bomb British cities with them starting in 1915. The problem was that once aircraft powerful enough to climb to the zeppelin’s altitude were developed, the slow and fragile zeppelins were sitting ducks: lighter than air airships could hardly be armored, after all. Moreover, the bomb loads of zeppelins had always been far too low to make effective strategic bombing possible beyond the initial shock of it.
What no one could have known in WWI was not merely that the technology for effective conventional strategic bombing wasn’t ready, but that it would probably never be ready. Interwar air-power theorists, seeing the potential of strategic airpower to bypass the trench stalemate by flying over it began to try to work out how this would be done. Giulio Douhet (1869-1930) argued that future wars would be fought and won in the air, with fleets of bombers using high explosives and chemical weapons to massacre enemy civilian centers, until civilians forced their governments to surrender. Douhet was not alone; his vision of airpower was shared, for instance, by the “father of the RAF”, Hugh Trenchard (1873-1956).
This concept, “morale bombing” as it is sometimes called, probably deserves its own post discussing its failures. But in brief, the concept was tested, with far larger amounts of bombs than Douhet or any other interwar theorist could have ever dreamed of, during WWII. The argument by air theorists that high altitude bombers could not be stopped was proved false when the British did exactly this, stopping German bombers over Britain in 1940. Moreover, terror bombing against civilian targets in Britain didn’t lead to surrender, but hardened resolve. Likewise, “morale” bombing against German targets by the allies didn’t lead to surrender, but hardened resolve. Later efforts to demoralize the North Vietnamese through a American bombing campaign in the Vietnam War didn’t lead to surrender, but hardened resolve. More recent efforts to demoralize or destroy terrorists and the Taliban through the use of airpower hasn’t lead to surrender, but rather hardened resolve. Likewise, efforts by the Syrian Regime to defeat various opposition groups in Syria through the use of chemical weapon-based terror bombing didn’t lead to surrender (siege-and-starve tactics did), but hardened resolve.
It turns out the fundamental premise of the entire idea of morale bombing – that being bombed will make people want to stop fighting – was flawed. Morale bombing has been, depending on how hard you squint at the US air campaign over Japan in WWII (including the use of nuclear weapons) successful either once (out of many attempts) or never. In most cases, the sustained bombing of civilian centers has been shown to increase a population’s willingness to resist, making the strategy worse than useless.
The case for strategic bombing against industrial targets is marginally better, but only marginally. While airpower advocates, particularly in the United States promised throughout WWII that bombing campaigns against German industry could lead to the collapse of the German war machine, in the end many historians posit that the real achievement of the campaign was to lure the Luftwaffe into the air where it could be destroyed, thus denying the German army of air cover and close air support, particularly on the Eastern Front. Some diminution of German industrial capabilities was accomplished (though it is not clear that this ever approached the vast resources poured into producing the large numbers of extremely expensive bombers used to do it, though the allies had such an industrial advantage over Germany, forcing the Germans to fight in expensive ways in the sky was a winning trade anyway), but the collapse of German industry never happened. As Richard Overy notes, German industrial output continued to rise during strategic bombing and only began to fall as a result of the loss of territory on the ground. Needless to say, “strategic bombing can sucker the enemy into wasting their close air support” was not the result that airpower advocates had promised, nor could it have broken the stalemate.
I don’t want to oversimplify the continued debate over the efficacy of strategic airpower here too much so let’s just say that the jury is still very much out as to if strategic airpower works even with modern technology; it certainly wouldn’t have worked with WWI era technology.
Bret Devereaux, “Collections: No Man’s Land, Part II: Breaking the Stalemate”, A Collection of Unmitigated Pedantry, 2021-09-24.
July 7, 2026
Clankers in higher education
On his Substack, David Friedman discusses the impact of AI on university exams:
Professor Roberto Serrano, who is the Harrison S. Kravis University Professor of Economics at Brown University, has detected a massive fraud in one of the classes he teaches, ECON 1170, an advanced undergraduate course in mathematical economics. He has conclusive evidence that at least 50 students cheated on the March midterm exam, making it the biggest known scandal at Brown and in the entire Ivy League
…
“Academic integrity is a value worth defending. The faculty cannot be left on its own in a battle that is decisive if we want to preserve the future of higher education,” explains the 61-year-old professor … (El Pais)
It was a closed book take-home exam; the problems were designed to test the student’s ability but proved doable by an AI. After Serrano changed the final from take-home to in-person about half the students who had gotten perfect scores on the midterm chose not to take the final.
The existence of AI, like the earlier problem of students buying papers online, reduces the ability of teachers to test their students but does not eliminate it, is inconvenient but not catastrophic. It makes some kinds of testing more difficult but not impossible; Serrano could have asked students whose midterms were suspiciously good to explain some of their answers and failed any obviously unable to do so. That would have been additional work for him and, judging by the article, not a policy Brown would have endorsed. Unwilling or unable to do that that he can base his future grading on work done in-person and adequately monitored.
[…]
It is not immediately obvious what is wrong with using AI on a test. If the purpose of the test is to generate information for potential future employers, why should they want the student tested without a tool that, if they hire him, he will have? A basketball coach does not evaluate potential team members by how well they can play with one hand tied behind their back.
Arguably the skills the employer wants tested are those that an AI cannot replace and it was up to Professor Serrano to find ways of testing for them. His take-home midterm, taken without the assistance of AI, might have provided information for him and his students about how far they had come along a path that would eventually produce skills an AI could not substitute for but not information for a future employer about the skills of the students taking the exam.
June 30, 2026
Leading the grassroots revolt against AI … Homer Simpson
Ted Gioia posted this a couple of days back, but if you haven’t read it it’ll still be new to you:
Last November I suggested that 2026 would witness a tech backlash of unprecedented intensity. And it’s now happening with a vengeance. Silicon Valley is getting skewered everywhere, and to a degree inconceivable just a short while ago.
Just yesterday, The Economist finally grasped how rapidly tech antipathy is mounting — and made AI backlash its cover story.
The latest survey numbers are devastating. Every demographic group is now opposed to AI—especially young people, previously the most enthusiastic supporters of new tech.
[…]
Not every pushback to encroaching tech is quite so gentle.
Consider the case of “Mr. Daniels,” a 25-year-old man from England. He knows that AI will rob every music file on the web for training — so he decided to poison the data.
How did he do it? According to Tuned Into Tech, it happens like this:
He took his entire music library of 2,000 records, stripped out the original vocals, and replaced every single one of them with the voice of Homer Simpson. Then he uploaded all of them to Soulseek. He didn’t change the metadata, the file names, the artist tags, the album information. They all stayed exactly the same.
A listener might not notice at first. Some of these songs have long intros, and those are unchanged. But as soon as the singing begins, Homer Simpson takes over. When AI tries to steal this for training, it gets fooled—and contaminates its own data set.
So somewhere deep in a training algorithm’s data set is the audio of Homer Simpson which the AI will assume sounds like [for example] Madonna, Rihanna, or maybe even Sean Paul. The model doesn’t know the difference. It just ingests the data and treats that like the truth.
And that is exactly what Mr. Daniels is hoping for.
He wants “to introduce noise, chaos” into the bots that are putting human musicians out of work.
“Mr. Daniels” is not an isolated example. Musician Benn Jordan has also been “poison-pilling” music files in hopes of disrupting AI.
In recent months, he has watched in horror as “tech companies started raising millions of venture capital dollars and scraping my music without my consent”. They now use his own work to generate “shittier music with it that is inadvertently associated with my name — and then attempting to resell that in the same economy in which I make money from my music”.
As a result, he has stopped releasing music. But he hasn’t walked away from the battle — instead Jordan has developed “a type of encoding that not only makes a music file more or less untrainable by generative AI companies, but actually has the ability to decrease the quality and efficiency of their entire data set”.
“Unethical generative AI companies have made artists feel incredibly powerless for quite some time now”, he adds, “but all of that is about to change”.
