Shellsharks Blogroll - BlogFlock

AI's costs are going through the roof - so businesses are telling LLMs to talk like cavemen - Werd I/O

2026-07-01T12:59:39.000Z

Link: Companies Are Making Claude and Codex Talk Like Cavemen to Stop AI’s Soaring Costs, by Joseph Cox at 404 Media

I find this very funny:

“Companies are deliberately making their AI tools speak like cavemen in an attempt to stop burning through AI tokens and curb their massive expenditure on AI, 404 Media has found. The tool turns the usually verbose outpost of LLMs like Claude Code, Codex, or Gemini into a much more to the point answer. Think less “you’re right to push back, I was wrong,” and more “Hulk smash.””

If only we had other limited-vocabulary lexicons designed to talk to computers efficiently!

I think we’re circling a few different possibilities that may show up over the next few years:

Literally LLM-specific “programming languages” that humans can use to talk to models more efficiently, of which Caveman is the hilarious first step
A proprietary bytecode-like language for LLMs that makes interactions more efficient but also just happens to be owned by one of the major vendors and creates a hitherto-unobtainable moat for their business
This all becomes moot when local models become viable for most businesses without insanely high hardware prices or configuration costs
LLM costs eventually fall to a fraction of their existing level

But who knows? Maybe enterprise businesses will continue to talk in stilted caveman language to achieve their business goals forever.

They tell us surveillance makes us safer. It undermines our democratic rights. - Werd I/O

2026-07-01T12:51:20.000Z

Link: The Quiet Erosion of Collective Action Under Digital Surveillance, by Gina Romero in Tech Policy Press

The most important outcome of increased surveillance is a chilling effect on free speech and expression. As Gina Romero, the United Nations Special Rapporteur on the Rights to Freedom of Peaceful Assembly, notes here, that extends to the organizations that have been established to protect those rights:

“As organizations operate under the constant assumption that they are being monitored, their core functions are profoundly affected. Their ability to serve as watchdogs, provide rights-based services, protect victims of human rights abuses, and educate the public is severely constrained. Ultimately, the very possibility of advancing and protecting rights, democracy and the rule of law is undermined.”

Civil society organizations and advocates have been mislabeled as national security threats around the world. It’s true in some of the nations that we’ve long thought of as being authoritarian, but it’s also true in the United States. Even places like the United Kingdom have tried to apply pressure to technology companies so that they can gain access to backdoors.

Tools like Signal have become all the more important. We need more easy to use end to end encrypted systems so that we can communicate and organize with each other without fear of government surveillance. That also allows whistleblowers and sources for journalists to reach out with less of a fear they they will suffer repercussions.

But those tools don’t stop you from being surveilled in the real world. Cameras and microphones are everywhere; license plate readers are now commonplace; even AI-enabled drones have been deployed for events like the World Cup.

It’s generally true that if government can do something, it will. So the only way to stop this kind of widespread surveillance is to make it impossible. Romero calls for legislative prevention that takes into account the whole systemic impact of surveillance rather than just the immediate first-order effects. Her report also calls out that it can be very difficult to challenge these systems because what they are and who owns them tends to be complicated or obfuscated:

“The study reveals a lack of transparency surrounding the relationship between state power and non-state actors, creating an information vacuum that makes surveillance practices exceedingly difficult to challenge through litigation. As a result, the right to an effective remedy is fundamentally weakened.”

So I think we also need more technical capabilities that interfere with how these systems of surveillance actually work. We need more spaces that are designated privacy-first and enforce an anti-surveillance rulebook. And, just as communities have taken it upon themselves to dismantle Flock cameras, we need to take back our streets.

A concrete tool to help newsrooms cover emergencies - Werd I/O

2026-06-30T22:10:04.000Z

Get stories like this sent to you every Friday: Subscribe for free.

Link: Emergency Mode for News, by OpenNews, NC Local, and Newspack

Emergency Mode is a set of resources, tools, and training that aims to prepare small newsrooms for various disasters. It’s a co-production between OpenNews, NC Local, and Newspack. They’ve done a great job. As their about page puts it:

“Emergency Mode for News equips local journalists and their newsrooms with the tools they need to respond to climate disasters. With a disaster reporting action pack, software and a learning community, Emergency Mode is designed to help journalists act nimbly and creatively to serve their communities when the unexpected happens.”

Toolkits include things like a practical checklist for newsrooms covering wildfires and a template for maintaining source lists during an emergency. There’s also a hands-on workshop series and tools like WordPress plugins for live rolling news updates and providing bandwidth-light versions of sites.

Most of all, I really appreciate the practical nature of all of it. Rather than hand-waving about principles and ideas, as many newsroom-facing resources do, everything here is a concrete tool that can actively be used in the field. Newsrooms are more squeezed than they’ve ever been, so it doesn’t hurt that it’s all free.

I’d love to see this level of concrete specificity for the normal working of a newsroom. Wouldn’t it be cool to have a list of business model checklists you could pull from? Or disaster recovery plans? Or data protection policies? Just as the tools on this site are going to be concretely useful to any newsroom that covers a disaster, checklists, tools, and training for standard operational practices could be really meaningful — particularly for smaller newsrooms that don’t have the ability to hire CTOs, CFOs, and so on.

In other words: more, please. This is lovely.

Community Survey 2026 - Werd I/O

2026-06-30T18:51:10.000Z

I’m grateful that people stop by and read my posts. I think I’m really lucky. Thank you!

I love writing here, but I’d love to know how I could serve you better. So every year I ask my readers to fill in a short survey. It doesn’t take more than a minute or two.

This survey will help me figure out which problems and ideas people are thinking about, which will help me figure out how helpful I can be.

Click here to answer a few short questions.

It’s entirely anonymous (unless you decide to leave your email address). And it helps me out a great deal.

Thank you in advance for your feedback!

Nintendo 3DS, to the movies, and a bad roll - W26 - Joel's Log Files

2026-06-30T15:40:00.000Z

I usually take pictures to remember what happens during the week. This was not the case for this one, so the notes themselves may not be as long as some of you may expect! Still, a lot of things happened.

This second paragraph is being written right after I was done with everything here, turned out to be just as long as usual so, I don’t want any complaints!

🕹️ I am now the proud owner of a red New Nintendo 3DS XL—Why did Nintendo thought adding “New” to the name was helpful?—and I love it! Some of my oldest readers may remember the time when I beat Metroid Samus Returns on one of these—though I never reviewed it properly, I think. Now I have the chance to play some other titles in the system. Right now: Fire Emblem Awakening.
🏴‍☠️ The device already came with some magic installed on it, so I’ve been able to acquire quite a few games and homebrew on it. Including a cool Chrono Trigger inspired theme. I may try and get it back to stuck, just to learn how to unlock it myself, but I may just keep using it as it, honestly.
📦 That wasn’t the only Nintendo system I ordered though. There’s also a 2DS XL—the last of the the 3DS family—that took longer to arrive. After using the N3DSXL, I don’t think I’ll keep it, though I really like its design, maybe I’ll give it to my sister who wanted one too, who knows. I’ll go pick it up today as of the writing of these notes
🍿 So I invited a friend to watch a movie with me and family—if you know you know—and it went well! She really enjoyed the film and we all chatted a bit about it and life and other things after we were done with it. We’ll watch another movie next Friday, actually, probably Supergirl.
🎮 I updated my Anbernic RG35XX SP to the latest version of MuOS, codenamed Funky Jacaranda. I am very happy to finally have a built-in Activity Tracker, even if it is a little barebones, I really appreciate its functionality. Speaking of, I didn’t know my 3DS had one too, it is so good!
🚲 This weekend’s bicyle trip was all about looking for a bike shop to find the fenders and bike I want to get installed. Unfortunately, everything is closed in the morning. I had a fun time commuting around downtown though, I went up some hills which required some effort, it felt good to overcome that. I actually wanted to try and pickup the 2DS XL, but same issue.
🛍️ It was Father’s Days a few days ago, but I didn’t get any gifts for him at the time—other than the event at church and a mug but still—so we went out shopping for some things and we also had a coffee together. We do have coffee together at home often, but its rare to go outside for one so, it was a good time.
🎲 We wanted to return to Land of Eem after a very big hiatus, but in the end, we couldn’t make it happen. Our first playthrough was great but there were a couple of players who haven’t been able to return or stay for a full session… We need to set a proper time for one. We did play Survidice again tho.

Gaming

Ongoing

Transistor - Close to the end now, the game has turned into a sort of psychodelic journey and there was this weird flip in my progress that I didn’t quite expect. The voice work, the soundtrack, everything has remained amazing, and the battles have only gotten more complex, I love the sort of real-time turn-based mix it has going and it has proven to be a true challenge as I approach the grand finale!
Terranigma - When dusting off my Anbernic handheld, I returned to this for about twenty minutes. I have advanced to chapter 2 of the story, turning the whole plot inside out once again. I am doing the first dungeon inside of a poisonous tree! It feels very Ocarina of Time, I must say. Though this game came out first!
Slice & Dice - Had some fun here. I am doing a pretty slow Blursed run (the same run during the last couple weeknotes), and yesterday I beat the twentieth battle at last! I used to reach 50 in a day before but, life goes on.

Started

Fire Emblem: Awakening - I tried this game on a borrowed 3DS from a friend, but now that I returned to it, after beating The Blazing Blade, I can’t help but truly love it even more! I played like seven hours of this already, and even decided to let someone die in battle for once. I’ll keep going, adding this to my Summer Game Challenge list—which seems to be changing every week.

Reading

Leviathan Falls - Up to chapter 9 - The grand finale continues, and it has NOT gone where I have expected it to so far, I am a little worried they are opening up new plotlines and won’t be able to be tied up by the end. Then again, they always manage to make everything converge so, I’ll just let myself be taken along for the ride.
Shikimori’s Not Just Cute - Up to chapter 160. It’s funny how I didn’t really care about returning to this series a couple weeks ago, and now that there’s only twenty chapters left I’m a bit sad that this will be over soon. We are on our last school festival event, after an awesome volleyball mini-arc.
Blue Lock - Up to chapter 351 - The new arc will be interesting, with one of the main players of the U-20 team out of the match, Yoichi will have to figure out a way to rethink his ego and skill as the number one striker of the team.

Watching

Toy Story 5 - I actually really enjoyed this movie, it’s the perfect blend of nostalgia with modernity in my opinion, as it deals with the clash between technology and toys in an very nuanced way. I loved it to bits. I am glad to have Jessie as a protagonist this time, she was absolutely awesome. Everything was great in my opinion.
Avatar The Last Airbender S2 (Live Action) - I love Toph, the actress nailed her performance. The show is entertaining, but I am not a fan of a lot of the changes they did. Things keep going a bit fast yet slow somehow. We will see how things go. I will watch it because of my sister though. She keeps shutting my mouth every time I raise a plothole lol.

Around the Web

Blog posts

There are been a lot of great posts on my rss aggregator lately, I actually read all these when they were published, instead of my usual approach, consisting of reading bunch of posts on a day to add them here.

stop asking writers about “AI” - Just yesterday Amin shared this to me and well, it’s a must-read. If you use AI, whatever, just don’t ask me to validate you.
Back, I Think - Syl is back! Her gaming blogposts (and other blogposts) have been a favorite of mine for almost a year now (remember Blaugust 2025?), she only really left like half a month, but I still missed her on my feed.
Lighten up, Eeyore - isa did some personality quizzes, reflected on her posts and their themes, not liking change but having to do so, and trying to write more about stuff she likes!
Internet works best in small dosages - This post was apparently inspired by me. Michal has done a lot to distance himself from the internet, from getting rid of most social media, to switching to an ink-screen phone. Great write-up and ramble!
Since You Asked - Some light-hearted life advice from Musings from a Tangled Mind, one of those blogs you just have to keep on your feed, always full of fun stuff like this. Get some snacks!
Everything Is a Computer. Nothing Is a Computer. - Jeremy decided to actually write for once, and he shared a fun life story plus reflections on technology today. Awesome to see.
What could you accomplish in sixty hours? - A last minute addition that didn’t go where I expected. Very much worth reading.

YouTube

Practicing intentional gaming during a heatwave (a vlog) - Laura’s channel has been one of my favorites when it comes to just chill gaming and life thoughts. This video doesn’t really have much going on, but isn’t that how life really is during those quiet times?
Should You Really Buy a 3DS in 2026? - I watched way too many 3DS videos, and many of them were excellent, but not many recognized the alternatives like this one does. There’s plenty of devices out there that are cheaper, have bigger libraries and are more capable than Nintendo’s dual screen device.
Why everyone’s Modding their Switch Again - The Nintendo Switch modding scene never really left, but with the Switch 2 becoming the focus, it continues to thrive! I am thinking about getting a chipmod for my own model one of these days.
the Spider-Man CGI Tier List - A fun rundown of every Spider-Man swing in film so far! I actually really liked the final list, and saw a few I didn’t know existed at all, cool.

This is day 87 of #100DaysToOffload

Reply to this post via email | Reply on Fediverse

Book Review: Fake Creativity by Blake Loch ★★★☆☆ - Terence Eden’s Blog

2026-06-30T11:34:20.000Z

Thanks to BookSirens for providing me with a review copy. This is an intriguing self-published novel with a backstory almost as interesting as the plot.

The story is a descent into paranoia as an author is convinced that an AI is plagiarising his work. As the madness takes over, he's forced to confront whether his creative processes are genuine or not.

It raises some excellent questions about whether AI can replicate art. It also posits some solutions for ensuring genuine human content. Without going in to spoilers, I think some of the methods the protagonist comes up with might be the only way to "prove" that a human has created a work.

The pace is excellent - with some well-placed plot twists. As with any self-published novel, it could do with a little tightening up. Some of the characters have oblique motivations which need a bit more exposition.

A note on AI use. There's a novel-within-a-novel which is genuinely generated by an AI (as the author freely acknowledges). I think this is an acceptable use of generative AI - the prose it produces is utterly risible and cliché ridden. It works as a nice contrast to the human generated text.

I suspect more and more authors will turn to AI fears just as they turned to pandemic allegories a few years ago. This is a decent attempt to capture a moment in time when authors stared into the abyss and found only themselves staring back.

Shipping post-quantum cryptography to Python - Trail of Bits Blog

2026-06-30T11:00:00.000Z

Post-quantum cryptography is now one <code>pip-install</code> away for the entire Python ecosystem. With funding from the <a href="https://www.sovereign.tech/">Sovereign Tech Agency</a>, we implemented support for ML-KEM, the NIST-standard key-establishment primitive, and ML-DSA, the NIST-standard digital-signature primitive, in <code>pyca/cryptography</code>. On June 22, 2026, the White House <a href="https://www.whitehouse.gov/presidential-actions/2026/06/securing-the-nation-against-advanced-cryptographic-attacks/">ordered</a> the U.S. government to accelerate its transition to post-quantum cryptography. The order says large-scale quantum computers, especially in adversarial hands, will threaten widely used cryptographic systems, and that attackers may already be collecting encrypted data now so they can decrypt it later. It also sets concrete migration deadlines: high-value and high-impact federal systems must use post-quantum key establishment by December 31, 2030, and post-quantum digital signatures by December 31, 2031. And even if you don’t care about quantum resistance, that’s not a problem because <a href="https://blog.trailofbits.com/2024/07/01/quantum-is-unimportant-to-post-quantum/">quantum resistance isn’t the main benefit of post-quantum crypto.</a> That transition cannot happen only at the policy layer. Every application that signs packages, validates certificates, establishes secure channels, or protects long-lived secrets depends on cryptographic libraries. If those libraries do not expose post-quantum algorithms, the software stack cannot migrate. Almost every Python program that touches cryptography goes through <code>pyca/cryptography</code>. It’s currently the <a href="https://pypistats.org/top">eleventh most-downloaded package on PyPI</a>, pulling 1.2 billion downloads in the last month alone. The <code>pyca/cryptography</code> package handles the cryptographic operations of projects like Ansible, Certbot (the Let’s Encrypt client), Apache Airflow, paramiko (the Python-only SSH client), and <a href="https://deps.dev/pypi/cryptography/48.0.0/dependents">many others</a>. If <code>pyca/cryptography</code> doesn’t ship post-quantum primitives, the Python ecosystem can’t begin to migrate. <h2 id="post-quantum-support-is-now-one-pip-install-away">Post-quantum support is now one pip install away</h2> As of <code>cryptography>=48</code>, support for post quantum algorithms is just a <code>pip install</code> away. The version 48 release includes our Rust bindings for ML-KEM and ML-DSA, the cross binding API and tests, and support for AWS-LC as a cryptographic backend. It also includes work from pyca/cryptography’s maintainers to support the other cryptographic backends. Sadly, this is not enough for a post-quantum migration drop-in swap. These primitives have different size, performance, and integration tradeoffs than the classical algorithms they replace. <h2 id="pq-algorithm-tradeoffs">PQ algorithm tradeoffs</h2> Post-quantum primitives keep the same security strength, but they change the size of the data on the wire. Public keys, signatures, and ciphertexts are often 1–2 orders of magnitude larger than the classical values they replace. The operations are also more complex and therefore slower, but on modern hardware they are still imperceptible for regular use, and are likely to get faster with improved hardware and algorithms. For signatures, here’s how the classical primitive (Ed25519) compares to its post-quantum equivalent (ML-DSA-65): <table> <thead> <tr> <th style="text-align: left">Algorithm</th> <th style="text-align: right">Public key</th> <th style="text-align: right">Private key</th> <th style="text-align: right">Output</th> </tr> </thead> <tbody> <tr> <td style="text-align: left">Ed25519</td> <td style="text-align: right">32 B</td> <td style="text-align: right">32 B</td> <td style="text-align: right">64 B sig</td> </tr> <tr> <td style="text-align: left">ML-DSA-65</td> <td style="text-align: right">1,952 B</td> <td style="text-align: right">32 B</td> <td style="text-align: right">3,309 B sig</td> </tr> </tbody> </table> And for key exchange and encryption, here’s how X25519 compares to its post-quantum equivalent (ML-KEM-768): <table> <thead> <tr> <th style="text-align: left">Algorithm</th> <th style="text-align: right">Public key</th> <th style="text-align: right">Private key</th> <th style="text-align: right">Output</th> </tr> </thead> <tbody> <tr> <td style="text-align: left">X25519</td> <td style="text-align: right">32 B</td> <td style="text-align: right">32 B</td> <td style="text-align: right">32 B shared</td> </tr> <tr> <td style="text-align: left">ML-KEM-768</td> <td style="text-align: right">1,184 B</td> <td style="text-align: right">64 B</td> <td style="text-align: right">1,088 B ciphertext</td> </tr> </tbody> </table> If you maintain a protocol or wire format that hardcodes Ed25519-sized signatures or X25519-sized public keys, the post-quantum migration involves more than a primitive swap. The surrounding fields, length prefixes, and chunking assumptions need to grow with it. <h2 id="using-ml-dsa-fips-204-quantum-resistant-signatures">Using ML-DSA (<a href="https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.204.pdf">FIPS 204</a>): Quantum-resistant signatures</h2> ML-DSA is the lattice-based signature scheme that replaces RSA, ECDSA, and Ed25519. The Python API mirrors the existing asymmetric primitives: <figure class="highlight"> <pre tabindex="0" class="chroma"><code class="language-py" data-lang="py">from cryptography.hazmat.primitives.asymmetric import mldsa private_key = mldsa.MLDSA65PrivateKey.generate() public_key = private_key.public_key() signature = private_key.sign(b"message") public_key.verify(signature, b"message") # raises InvalidSignature on failure</code></pre> </figure> <h2 id="using-ml-kem-fips-203-key-encapsulation-for-the-post-quantum-era">Using ML-KEM (<a href="https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.203.pdf">FIPS 203</a>): Key encapsulation for the post-quantum era</h2> ML-KEM is a key encapsulation mechanism (KEM) for establishing shared secrets. The construction is different, though. ML-KEM is a key encapsulation mechanism, not a Diffie-Hellman exchange. Instead of both parties combining key shares to derive a shared secret, one party encapsulates a fresh shared secret to the receiver’s public key, and the receiver decapsulates it with the matching private key. These operations allow both parties to exchange a secret but in a manner fundamentally different from Diffie-Hellman, and resistant to quantum factoring attacks. <figure class="highlight"> <pre tabindex="0" class="chroma"><code class="language-py" data-lang="py">from cryptography.hazmat.primitives.asymmetric import mlkem # Receiver generates a keypair and publishes the public key. private_key = mlkem.MLKEM768PrivateKey.generate() public_key = private_key.public_key() # Sender encapsulates a fresh shared secret to that public key. shared_secret_sender, ciphertext = public_key.encapsulate() # Receiver decapsulates the same shared secret from the ciphertext. shared_secret_receiver = private_key.decapsulate(ciphertext) assert shared_secret_sender == shared_secret_receiver</code></pre> </figure> <h2 id="the-road-ahead-slh-dsa-and-protocol-integration">The road ahead: SLH-DSA and protocol integration</h2> Two areas are still in progress: a third NIST standard, and the work of integrating these primitives into real protocols. <h3 id="slh-dsa">SLH-DSA</h3> SLH-DSA (<a href="https://nvlpubs.nist.gov/nistpubs/fips/nist.fips.205.pdf">FIPS 205</a>) is NIST’s hash-based digital signature standard. Like ML-DSA, it is meant to replace classical signature schemes such as RSA, ECDSA, and Ed25519. Its tradeoff is different: SLH-DSA has very large signatures and slow signing, but it relies only on the security properties of hash functions, which have been studied for decades. That makes it a conservative backstop if future cryptanalysis weakens lattice-based signatures. SLH-DSA is not supported in <code>pyca/cryptography</code> 48, but we’ve started working on it. <h3 id="post-quantum-in-protocols">Post-quantum in protocols</h3> Primitives are the foundation, but the post-quantum migration will be complete only when protocols use the post-quantum resistant algorithms. You’re unlikely to use PQ algorithms directly in tools like Certbot or Ansible until common protocols add support for them. While well-designed to replace existing implementations, algorithm changes require cautious development, testing, and auditing. We are actively working on helping maintainers integrate PQ algorithms into applications. <h2 id="acknowledgments">Acknowledgments</h2> This work was funded by the <a href="https://www.sovereign.tech/">Sovereign Tech Agency</a>, whose mission is to support the open-source infrastructure that public digital systems depend on. We’re also indebted to pyca/cryptography’s maintainers, <a href="https://langui.sh/">Paul Kehrer</a> and <a href="https://alexgaynor.net/">Alex Gaynor</a>, who offered constant feedback and review throughout the development process, and continue to steward this critical piece of open-source software.

📝 2026-06-30 11:37: Our 2 hens have finished sitting on the Guinea fowl eggs - out of 10,... - Kev Quirk

2026-06-30T10:37:00.000Z

Our 2 hens have finished sitting on the Guinea fowl eggs - out of 10, we managed to hatch 5 of them.

The chicken eggs we have in the incubator won't be ready for another week or so. Excited to see how many of them we get.

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

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

Wonders of Web Weaving, Episode 8 - James' Coffee Blog

2026-06-30T00:00:00.000Z

The eighth episode of Wonders of Web Weaving is out:

In Episode 8, I chat with Brennan, the author of brennan.day. We talk about, among other things, writing routines, building community in the indie web, "start here" pages on personal websites, and more.

I hope you enjoy the episode!

Wonders of Web Weaving also has an RSS feed you can use to follow along from wherever you get your podcasts.

Brennan brennan.day The eighth episode of Wonders of Web Weaving is out Wonders of Web Weaving also has an RSS feed

A right-wing media chain tried to replace 47 newspapers with AI. They all died. - Werd I/O

2026-06-29T12:54:29.000Z

Get stories like this sent to you every Friday: Subscribe for free.

Hey, so, mea culpa.

I referenced an article here by what appears to be a fully AI-generated newsroom. The irony!

The story doesn’t seem to be real — at least, not the piece about 1819 News. (My points about local news are real and important!) In particular, some of the newsrooms listed still appear to exist. And it’s questionable whether the journalists are real, either.

Thank you to Damon Kiesow for bringing this to my attention.

So the new question is: who is behind The Editorial, and what is the purpose of this site?

I’ve left my post up below.

Link: The Ghost Paper That Ate Alabama: How a Media Startup Killed 47 Weekly Newspapers and No One Noticed, by Elena Marchetti in The Editorial

This is a horrifying story: 47 small, local, weekly newspapers were acquired by a man with ambitions to create a new conservative media empire. He fired much of their staff and replaced them with AI — with predictable results. Subscriptions fell, there were accuracy and ethics issues, and the papers even lost their printing contracts because local print shops didn’t want to print slop. The media empire fizzled out.

This isn’t the first time someone has bought into the AI marketing — and had a disregard for human expertise — enough to make a foolish staffing decision. Companies as big as Ford have sacked humans in favor of AI only to backtrack with their tails between their legs and re-hire. But here the aftermath was news deserts in 43 of the poorest communities in America. When a community becomes a news desert, it is more likely to experience corruption, making them still more vulnerable.

But for me, the most important part of the story is what the communities did next:

“In three towns, residents have started volunteer-run Facebook groups to share local news. In one — Grove Hill, population 1,800 — a retired teacher named Evelyn Petty has taken it upon herself to write a weekly newsletter, which she prints on her home laser printer and distributes at the post office. “It’s not a newspaper,” she says. “But it’s something.” She has seventy-three subscribers. Each pays $20 a year. She loses money on every issue.”

Community news is so important that, in the absence of anything else, people will spontaneously start groups, publish newsletters, and fill the void however they can. (It’s also not lost on me that Facebook groups — in other words, community spaces — have, with journalistic support, the potential to occupy that gap.) But they need support.

I’m reminded of the story of Kari Mar, who launched a non-profit community newsroom, La Conner Community News after her local paper shut down. That effort succeeded to the extent that she could buy a neighboring community’s struggling weekly paper. Similarly, I recently had the privilege of meeting Amy Bushatz, whose non-profit Mat-Su Sentinel in Alaska bought its local weekly in order to revitalize it.

Both are supported by Tiny News Collective, which supports early-stage local news funders. I recently joined the board, and will write more about their work in a future post. Not only are these new local startup newsrooms a way to prevent news deserts and the kind of strip-mining the newsletters in this story were subject to, but they’re also more likely to innovate with models for news. I think they offer some hope for both the future of news itself and as democratic infrastructure for the communities they support. And in turn, they deserve our support.

Two posts about writing - Werd I/O

2026-06-28T22:11:28.000Z

Winnie Lim:

“When I think about what I wish to accomplish if I knew I was going to die, apart from spending time with my loved ones I think only writing is left. It is a way of emptying myself I guess, a way of sublimating my sad existence. I wouldn’t pretend that my writing is useful, but at least it will be a window into an interior world that doesn’t exist anywhere else.”

Co-signed.

Eliot Peper:

“Writing about whatever you’re uniquely obsessed is hardly a new approach, it’s the basis of all good writing. The only new thing is that writing about anything else is less valuable than it used to be, and it was never that valuable in the first place. Life is short. If you’re going to write, write something worth reading.

Write what only you can write.”

Book Review: The Hotel Avocado by Bob Mortimer ★★☆☆☆ - Terence Eden’s Blog

2026-06-28T11:34:45.000Z

Despite vowing not to read sequels of books I love, I'm constantly surprised that regression to the mean is an iron-clad law of the universe. I thoroughly enjoyed the first book in the series, so eagerly gobbled up the second. What a burlap fool I am.

What was charming and wry in The Satsuma Complex is now overdone and clichéd.

The violence, which was an undercurrent in the first book, is now too pronounced and too grim to be funny. Yet the constant tomfoolery with language undermines any built-up tension.

The language is gorgeous though. Bob Mortimer has an ability to conjure the most ridiculous similes and metaphors. It's impossible not to be entranced by his panache.

The ending is, frankly, a bit unearned and unsatisfying. It feels like a word-count limit was reached and the quickest solution was found.

The Laziest Generation - Kev Quirk

2026-06-28T08:01:00.000Z

The Laziest Generation

by Ibrahim Diallo

Ibrahim talks about house prices in the US, how it's only getting worse, and the perception from previous generations that kids today are somehow lazy because they can't afford a house before the age of 40.

Read post ➡

Being the father of 2 young people, this worries me too. Despite this post being US-centric, the script is the same here in the UK.

Unless my kids generation come out of school on a 6 figure salary, they don't have a hope in hell of buying a decent house.

To put that in context, here in the UK a £100,000 salary puts you in the top 3% of earners.

In the late 90's a house would cost around 4x a person's salary on average. Today it's 8x. So most can forget about saving for a deposit.

Instead younger generations will have to rely on inheritance, which will only exacerbate the late stages of life in which people are buying houses.

Something has to give.

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

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

3D Printers are actually very useful - Kev Quirk

2026-06-27T15:13:00.000Z

I recently started getting into 3D printing, but so far I've spent most of my time getting setup and learning the ropes. I've now completed my first little project with the 3D printers and I'm really happy with the result.

As a biker of many years I have a number of helmets lying around. This is because you're supposed to replace a helmet every 5 years, because the protective foam inside degrades over time. So I have a small collection of lids and nothing to really do with them.

So, I decided to print myself some helmet stands and mount them in my office.

The stands

There were 3 helmets I wanted to display. A reasonably well rated helmet stand on Amazon costs around £11. So for the 3 lids, I'd be looking at £33 (~$45).

Instead of handing over 33 of my finest pounds to Jeff Bezos, I decided to have a nose on Maker World and found this helmet stand that was very well rated. So I downloaded the files and set my printers to work, and a day or so later I had these little beauties:

They feel really solid and have no problem holding a helmet on the wall. Better yet, they only cost me around £2.50 ($3.30) each in filament (~750g of filament in total), so way cheaper than the Amazon option.

Today I finally had time to mount the lids to the wall, and I think they look great!

Sure, I could have pissed about making toy dragons or whatever, but I think these are a far better use of my 3D printers, and really why I bought them. I'm so glad that the printed results are good enough to be useable.

I already have some ideas of things I want to create next, but I'm going to have to start familiarising myself with FreeCAD for that project. We'll see how that goes...

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

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

Are frontier models really too dangerous? - Werd I/O

2026-06-27T14:50:40.000Z

Link: OpenAI will delay GPT-5.6 after Trump administration request, by Hayden Field in The Verge

I’ve got (at least) two worries about the story that the Trump Administration halted the release of models from both Anthropic and OpenAI.

Anthropic recently pulled its Fable model release in response to the government. Now it turns out that OpenAI has done something similar:

“The Information reported that OpenAI CEO Sam Altman told employees Wednesday in a company Q&A that it would release GPT-5.6 in limited preview form — granting access only to a small group of enterprise customers — in compliance with a request from the federal government. During that preview period, the Trump administration itself would reportedly approve access for customers on a case-by-case basis.”

In some ways, what a coup for the AI industry. This technology is so powerful that the government doesn’t think anyone should have it — and when it does inevitably release into the public’s hands, what a valuable product that will be. Get the tech that’s too dangerous to be released! This magical product can be yours for an unbelievable price!

So one worry is that this is, in essence, great marketing for these vendors.

But it’s worth remembering that these AI models are black boxes that respond to information queries in opaque ways. The more people rely on them for knowledge, the more powerful the models become. The argument being presented is that they can be used in ways that might present traditional security threats — but consider that some versions of the truth, to the wrong kind of authoritarian-minded government, might also be considered a threat. (Remember that “extremism on migration, race, and gender” and hostility to “traditional American views” are now considered markers of domestic terrorism.)

This is a golden opportunity, in other words, to hit pause on frontier model releases, at a time when models are becoming more prevalent, in order to make sure models are shaped to represent a certain version of the world. The administration has already signaled a willingness to do this; there is nothing to say they aren’t. The only way to prove that they aren’t is to open source not just the models but the training process and make the whole thing transparent and verifiable. The industry is a long way off from doing that.

Notable links: June 26, 2026 - Werd I/O

2026-06-26T13:39:14.000Z

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

Did I miss something important? Send me an email to let me know.

Did someone forward this to you? Subscribe for free.

Before we begin this week's roundup, I want to express my condolences to the friends and family of Om Malik, who we sadly lost this week. Om was one of the most important voices in blogging, tech journalism, and then the wider industry. So many people in tech have a story about him that highlights his conscience, thoughtfulness, and kindness. He will be missed.

Journalism's logical fallacy

Journalism is in crisis, and it’s really easy (and lazy) to say that making a technology or process tweak will fix it: we just need to use AI to fill capacity gaps, or build stronger comments into our site, or we need a better business or distribution model.

None of those things address the underlying question of why we need journalism, why it’s important, and what it should be. By addressing innovation at the edges, newsrooms are avoiding the hard, existential work of revisiting their core value to begin with. But it’s only by understanding that core value that they will actually reset their relationships with audiences, build greater trust and loyalty, and pull themselves out of the rut they find themselves in.

Shirish Kulkarni’s findings from listening projects in Wales — with multiple dramatically different demographics — contradict a lot of the narratives newsrooms have been telling themselves. For example:

“The second finding challenges one of the journalism industry’s most comfortable premises: that audiences – particularly marginalised communities – are news-illiterate and need to be educated. The opposite is true. In fact, the communities we work with are forensically sharp about media – often more so than the industry insiders who talk about them.”

This mirrors something you often hear from mission-driven tech projects: “we just need to educate the user”. Usually the opposite is true: you need to educate yourself about the user and give them the thing they actually need. And in the case of journalism, at least as a finding of this research, the need turns out to be pretty simple:

“They want help making good decisions. For themselves, their families, their communities. Not drama, not outrage, not the next breaking story. Practical, trustworthy, usable information that helps them navigate their lives.”

It’s important, once again, to separate the work of news — breaking headlines, emergent facts — from journalism’s work to provide context and meaning. The first is a commodity; the second is both inherently community-driven and has always been more valuable.

Here I want to bang an old drum: newsrooms like to talk about audience strategies, not community strategies. It’s a meaningful difference that Shirish highlights well. The first implies an ivory tower broadcast approach: “we just need to reach people”. The second is an active relationship between a newsroom and the people it serves; a two-way conversation that requires trust and understanding on both sides.

The internet has always been a conversation. We’ve had the ability to build relationship-centric news organizations for 30 years, but most remain stubbornly set in a print mindset. This kind of research makes it clear how important that shift is, but, like Shirish, I don’t believe most existing newsrooms will evolve to actually meet this need. They can’t: the immediate commercial pressure is severe, and changing the model requires changing highly-ingrained cultural norms and assumptions that have been inherited from print. And in the midst of that panic, they’re jumping into bed with companies (AI vendors, proprietary social media platforms) that intermediate their relationships with their communities in exchange for some short-term wins.

So their outlook is not rosy. Instead, I think we’ll see new newsrooms emerge that reinvent what journalism is, are unafraid to build real, lasting, two-way relationships with the people they’re trying to serve, and eat everybody else’s lunch.

Empower others

Another superb post from Corey Ford:

“There's a simple diagnostic for any organization: Are your leaders focused on accumulating control, or on empowering the people around them?

Too many people in positions of power are overly focused on themselves. They hoard decisions. They micromanage. They grab power. And the people around them suffer for it. Those employees wait for direction instead of taking initiative. They bring problems instead of solutions. They ask "now what?" instead of saying "here's what I think we should do.””

He goes on to describe bus factor, the question of what happens if you’re hit by a bus. (My colleague Adam Hirsch recently offered “lottery factor”, reframing the question to be about what happens when someone wins the lottery and goes away to live the life of their dreams, which is more fun to think about.) Could your organization live without you?

The only way to build an organization that can live without you is to devolve decision-making — and the only way to do that is to create the conditions to make devolved decision-making a safe norm. Corey’s been building on that idea in his recent posts, and I enthusiastically co-sign. I’ve taken “surround yourself with people who are better than you” to heart: the Director of Product Engineering, and Director of IT and Security, who both report to me, fit that description exactly and we’re all better off for it.

When we examine situations where this doesn’t happen, it’s worth considering incentives. If you’re a founder or top-level leader who is building a sustainable organization in good faith, these principles are obviously, straightforwardly a good idea. If, on the other hand, you’re not in that position, and you’re worried about your own standing in the organization, you might feel like it’s a good idea to entrench your own power and decision-making. I think that’s short-sighted: you’re there to support the organization, and hiring an amazing team, as long as you support them well, can only make you look good.

At various tech companies in particular, I’ve also encountered people who don’t want power to be devolved to them: they want to be given a plan and to follow it. I don’t think that’s realistic in most smaller companies, and even in larger ones autonomy, insight, and decision-making ability are prized. It’s a thing to hire for. When you want to devolve decision-making, you hire people who can make great decisions. When you want to entrench your own power and micro-manage, you hire followers. These are different paths. I can only recommend the former.

This, from Corey, is absolutely vital:

“Real empowerment means saying: "That's not what I would have done. But you own the decision. Here's what I'm thinking about. Here are the questions I'd want you to consider. But ultimately, it's yours.””

Nobody will ever do it your way. If you want them to follow your lead exactly, you’re not devolving power. If you want them to be genuinely empowered, you need to not just accept that they do things their way, but embrace it — while also being transparent about your own thinking and principles, and setting norms mantras that help people navigate what’s important for the team.

He’s got a lot to say about that. As always, his full post, and the underlying series, are required reading.

It is trivially easy to use Reddit to manipulate AI search, research suggests

All good fun:

“A tiny snippet of user-generated text as short as 13 words long is often enough to manipulate the AI agents that power tools like ChatGPT and Google’s AI search, new research shows. The study suggests that it is trivially easy for brands to inject promotional content on sites like Reddit, Quora, and Wikipedia with the end goal of poisoning or manipulating the output of AI tools.”

So not only do we need to worry about AI-generated slop polluting our social spaces, we also have to worry about people who want to influence the output of the AI-generated slop polluting them, too. There’s a whole industry of companies trying to improve their clients’ coverage in AI results, just as there were for search engine results. And all of them will be spamming the crap out of our public communities and collaborative websites.

As 404 Media points out, this poses a real question for moderators in those spaces. How can they possible stem the tide? It’s not clear that this is even possible. Which means, inevitably, that the signal vs noise ratio in those spaces will decline, leading to a decline in usership of those spaces overall, and a retreat for most people into private group chats and uncrawlable communities.

To put it simply:

“Poisoning LLM results is basically just as easy as doing targeted posting on highly relevant subreddits to the industry or company you’re trying to promote, phrasing the comment to align with popular LLM queries, and attempting to evade moderation for as long as possible.”

Guarding against bot-driven spam is a relatively simpler problem. In contrast, this content will often be insidiously human: cunningly designed to try to provide value while also hiding a paid agenda. In some ways, it’s all the same as it ever was, but the volume of the junk is only going to keep increasing.

Of course, the silver lining is that eventually these sources will become unusable for AI training too. Then, finally, maybe everyone will go away.

Texas anti-ICE protesters convicted of terrorism charges sentenced to at least 50 years in prison

This is a litmus test for the freedom to protest in the United States:

“A group of Texas protesters convicted of terrorism charges received unusually harsh sentences of at least 50 years in prison on Tuesday in a closely watched case that was widely seen as a test case of the Trump administration’s efforts to crack down on dissent.”

Let’s be clear: a few of the protesters were out of bounds. One fired an AR-15 at the police, which goes beyond legitimate protest into inciting violence (and maybe even deliberate provocation). I would never condone that kind of activity. None of us should.

But these sentences far outstrip anything that’s been given to anyone on the right wing: the leader of the Proud Boys, as this article notes, was sentenced to 22 years in prison. One protester wasn’t even present, but was sentenced to 30 years for moving some zines:

“The ninth defendant, Daniel Sanchez-Estrada was not at the protest, but was convicted of corruptly concealing a document or record after prosecutors said he moved leftwing zines and other materials at the request of Rueda, his wife, after she was arrested. Sanchez-Estrada was sentenced to 30 years in prison on Tuesday.”

Many of the protesters had guns and were part of a gun club. They all possessed them legally. I personally wish there was not a right to bear arms and think that their ubiquitous presence in America makes everyone less safe, but the right to own them is enshrined in the Second Amendment. Instead, other “evidence” was used to infer that they planned violence, including this specific argument that should give everyone pause:

“[…] including their decision to communicate and auto-delete messages on Signal, an encrypted messaging platform widely used among activists, journalists and other citizens wary of government surveillance.”

Collectively, the justice department argued that these convictions are proof that anti-fascists are terrorists, which should also give us pause. The precedent here is obviously very dangerous for freedom of speech, freedom of assembly, and democracy in America.

📝 2026-06-26 13:59: My first rather large #3DPrinting project. Can anyone work out what they are? (No they're... - Kev Quirk

2026-06-26T12:59:00.000Z

My first rather large #3DPrinting project. Can anyone work out what they are?

(No they're not abstract Starship Enterprises)

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

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

Review: Gamrombo PS5 controller - including Linux set up ★★★★☆ - Terence Eden’s Blog

2026-06-26T11:34:42.000Z

I'm not paying seventy bloody quid for an official PS5 controller - so I found a knock-off version for a smidge under £40. And this one has lots of unnecessary blinkenlights!

Gamrombo is the consumer-facing brand of the generically named Professional Controller Manufacturer. AKA "Huizhou Ronghui Technology Co., Ltd" - there's virtually no information about them online other than paid-for reviews. Well, I bought this with my own money - so enjoy this honest review:

It's basically fine and it works on Linux.

OK, a bit more detail

The USB cable which comes with the controller is charge only - so you can dispose of it. Using a USB-C to C cable, I plugged it into to my Linux laptop and it was instantly detected. Rather cheekily, it shows up as 054c:0ce6 Sony Corp. DualSense Wireless Controller.

The PS5 touchpad was immediately usable. Single finger moved the mouse pointer, two-finger scrolling and zooming worked, clicking the pad clicks the mouse. Nice!

I popped along to hardwaretester.com/gamepad in Firefox, everything worked as expected.

Plugging it in to my Android phone also worked. So pretty handy if you want to play games on a miniscule screen.

Same with Bluetooth. I held down ⨞ and 🏠 on the controller, then connected. Everything worked. Weirdly, it wouldn't stop vibrating until I'd changed the LED colours.

Speaking of which - how much do you like garish LEDs? Tapping the light button changes the colours of the glowing rings around the joysticks. There's half a dozen solid colours or a slowly rotating rainbow effect. Or, just hold down the button to switch them all off.

As well as being cheaper than an official controller, it has more functionality. There are two programmable "macro" buttons on the back. Each can hold a sequence of up to 24 buttons pushes. Perfect if you need to record combos.

There's also a "turbo" button. You can use it to turn a long-press into a repeated-press. Perfect for saving you from RSI.

It vibrates as normal, has an inbuilt speaker, 3.5mm jack, and charges at about 2.5 Watts - all the same as a regular controller. Oh! And it paired with the PS5 as well 😆

OK, that's the good stuff. Where are the drawbacks?

Downsides

No microphone. This might be a deal-breaker for some, but I don't play any games which need it.

There is a firmware update available - but it comes as an EXE which only works on Windows a and is distributed on WeTransfer!

The icons on the buttons aren't the × ○ △ □ you'll see in-game. They are the right colour though.

Charging is only by USB. It won't work in a charging dock.

I managed to lock-up the controller. Between faffing around with Bluetooth and plugging it in to a bunch of different computers, it had a bit of a tizzy and wouldn't turn on again. Clicking the reset button using a paperclip brought it back to life.

Finally, it does feel kinda cheap. It is lighter weight and the buttons don't feel quite as "clicky" as an official controller. It isn't bad per se but it is different.

Should I get one?

It's pretty good! I've no idea how long the battery will last, or if it'll fall apart after I post this review, but saving £30 means you can spend more on games.

📝 2026-06-26 09:57: You know what isn't fun? Living in a 200 year old stone house, with no... - Kev Quirk

2026-06-26T08:57:00.000Z

You know what isn't fun? Living in a 200 year old stone house, with no insulation, during a heatwave.

There's collapsed animals (not literally) and fans all over the place. 🥵️

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

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

Started reading Kiosk Kingdom - Molly White's activity feed

2026-06-26T02:38:12.000Z

Started reading:

Discount Dan's Backroom Bargains series, book 3.

Kiosk Kingdom

by James A. Hunter.

Published March 31, 2026. 680 pages.

Started June 25, 2026.