oneDNN 3.10 Continues Preparing For Future Intel CPUs With AVX 10.2
Released one Friday was the newest version of oneDNN as this library started off by Intel and now officially under the UXL Foundation umbrella for serving as building blocks for deep learning software… ⌘ Read more

Linux Kernel Patches & Device Tree Additions For PCIe M.2 Connectors
On ACPI-enabled systems Linux users can enjoy PCIe M.2 connected peripherals that “just work” without any extra fuss. But for those relying on Device Tree (DT) handling by the kernel, new patches from Qualcomm are working on representing PCIe M.2 connectors within DT files… ⌘ Read more

‘Nintendo Has Too Many Apps’
The Verge’s Ash Parrish writes: Nintendo has released a new store app on Android and iOS giving users the ability to purchase hardware, accessories, and games for the Switch and Switch 2. When I open my phone and scroll down to the N’s, I get a neat, full row dedicated entirely to Nintendo. That’s four apps: the Switch app, the music app, the Nintendo Today news app, and now the store. (The tally increases to five if … ⌘ Read more

NILFS2 File-System Seeing Renewed Interest With Additional Maintainer
It’s been a while since having anything major to talk about with the NILFS2 file-system but it looks like that could be changing. NILFS2 as a reminder is a log-structured file-system with continuous snapshotting with its NILFS predecessor having been in the mainline kernel for two decades since the mid Linux 2.6 days… ⌘ Read more

Rideshare Giant Grab Moves 200 Macs Out of the Cloud, Expects To Save $2.4 Million
Singaporean super-app company Grab has dumped 200 cloudy Mac Minis and replaced them with physical machines, a move it expects will save $2.4 million over three years. From a report: Grab is Southeast Asia’s leading rideshare and food delivery outfit and therefore needs to build apps for iOS to connect with custom … ⌘ Read more

iOS 26.2 to Allow Third-Party App Stores in Japan Ahead of Regulatory Deadline
Japan’s new competition rules are forcing Apple to open the iPhone to third-party app stores, and iOS 26.2 will quietly flip that switch ahead of the December deadline. MacRumors reports: According to a post shared on X by @Tzzlala, iPhones running the beta in Japan are able to install alternative app stores like AltStore PA … ⌘ Read more

@movq@www.uninformativ.de Gemini liked your opinion very much. Here is how it countered:

The criticism of AI as untrustworthy is a problem of misapplication, not capability.

• AI as a Force Multiplier: AI should be treated as a high-speed drafting and brainstorming tool, not an authority. For experts, it offers an immense speed gain, shifting the work from slow manual creation to fast critical editing and verification.
  
• The Rise of AI Literacy: Users must develop a new skill—AI literacy—to critically evaluate and verify AI’s probabilistic output. This skill, along with improving citation features in AI tools, mitigates the “gaslighting” effect.
  

The fear of skill loss is based on a misunderstanding of how technology changes the nature of work; it’s skill evolution, not erosion.

• Shifting Focus to High-Level Skills: Just as the calculator shifted focus from manual math to complex problem-solving, AI shifts the focus from writing boilerplate code to architectural design and prompt engineering. It handles repetitive tasks, freeing humans for creative and complex challenges.
  
• Accessibility and Empowerment: AI serves as a powerful democratizing tool, offering personalized tutoring and automation to people who lack deep expertise. While dependency is a risk, this accessibility empowers a wider segment of the population previously limited by skill barriers.
  

The legal and technical flaws are issues of governance and ethical practice, not reasons to reject the core technology.

• Need for Better Bot Governance: Destructive scraping is a failure of ethical web behavior and can be solved with better bot identification, rate limits, and protocols (like enhanced robots.txt). The solution is to demand digital citizenship from AI companies, not to stop AI development.

Fwupd 2.0.17 Released With More Hardware Support & Features
Days after the Linux Vendor Firmware Service celebrated 135 million firmware downloads, a new version of the Fwupd utility is now available for firmware updating systems and peripherals under Linux… ⌘ Read more

LXQt 2.3 Released With Improved Wayland Support
LXQt 2.3 is out today as the newest release of this lightwight, Qt-based desktop environment… ⌘ Read more

Won a bunch of games of Solitaire and then rearranged the cards for maximum negative points, to distract me from the horrors.

(Still ended up with >0 points on OS/2, because don’t ask me.)

https://www.uninformativ.de/desktop/2025%2D11%2D04%2D%2Dkatriawm%2Dsolitaire.png

@movq@www.uninformativ.de streamlining jenny.vim?

index adc0db9..cb54abc 100644
--- a/vim/ftdetect/jenny.vim
+++ b/vim/ftdetect/jenny.vim
@@ -1 +1,2 @@
 au BufNewFile,BufRead jenny-posting.eml setl completefunc=jenny#CompleteMentions fo-=t wrap
+au BufRead,BufNewFile jenny-posting.eml normal $

I experimented with a 2.4x7mm aluminium rivet I had on hand. As expected, it was quite a bit long. Using my pliers wrench, I was able to crush it down by quite some bit. I should have taken a photo right after the hand riveter for comparison. Now, it’s much smoother and the chance of cutting my hand open is reduced by quite a bit. But breaking the burr with a few file strokes is still necessary. I should get 2.4x4mm rivets and try with them. I reckon they would be more suited for my 0.5mm sheet metal.

With the pliers wrench again, I was able to also crush down the chopped off 3mm copper nail and form a second head. That was surprisingly easy. Now, I need to figure out how to efficiently make a head on the remaining copper nail shaft, so that I can use this again.

Both are rock solid, there’s absolutely no movement at all between the two sheet metal cutoffs.

https://lyse.isobeef.org/tmp/nietenexperiment/

Okay, they are also offering 2.8x25mm copper nails. Which I actually do have a single one here. :-)

My hardware collection also includes a few brass-like looking screws that I could repurpose into rivets. But I reckon I have to upgrade my burner first. I’m not a metal worker by any means, so I could be totally wrong, but I imagine that some heat is necessary to loosen the work-hardening effect when beating on them. I will do some experiments on Saturday and report back.

@itsericwoodward@itsericwoodward.com No worries, all good, mate! We all have to start somewhere. Other software requests my feed several orders of magnitude more often.

I can confirm, the User-Agent header appears to be fixed. \o/

Two other things I noticed, though:

1. There’s now an OPTIONS request for my feed coming from something that claims to be Firefox, pointing to your feed URL in the query. No clue what this is about. In any case, it’s rejected with a 405 Method Not Allowed.

2. Not that these few requests bother me at all, but you might wanna implement caching next with either the If-Modified-Since or If-None-Match request headers. This way, if the feed hasn’t changed, the web server can reply with a 304 Not Modified and no body at all, saving unnecessary traffic. But again, this is really not an issue for me at all. I just wanted to make sure you’re aware of it, that’s all. It might be even already on your agenda. Or you might decide to never do anything about it, which is also fine for me. :-)

Please don’t hate me today; I’m a bit grumpy and have too many reasons to be upset:

• 2 counts of pushing and trying to get the simplest things done at work (that for some reason are made more difficult than they should be)
  
• This whole Chat Control bullshit
  
• And some other person things going on that have been ongoing for 72 days and counting 🤬

@prologic@twtxt.net I know we won’t ever convince each other of the other’s favorite addressing scheme. :-D But I wanna address (haha) your concerns:

1. I don’t see any difference between the two schemes regarding link rot and migration. If the URL changes, both approaches are equally terrible as the feed URL is part of the hashed value and reference of some sort in the location-based scheme. It doesn’t matter.

2. The same is true for duplication and forks. Even today, the “cannonical URL” has to be chosen to build the hash. That’s exactly the same with location-based addressing. Why would a mirror only duplicate stuff with location- but not content-based addressing? I really fail to see that. Also, who is using mirrors or relays anyway? I don’t know of any such software to be honest.

3. If there is a spam feed, I just unfollow it. Done. Not a concern for me at all. Not the slightest bit. And the byte verification is THE source of all broken threads when the conversation start is edited. Yes, this can be viewed as a feature, but how many times was it actually a feature and not more behaving as an anti-feature in terms of user experience?

4. I don’t get your argument. If the feed in question is offline, one can simply look in local caches and see if there is a message at that particular time, just like looking up a hash. Where’s the difference? Except that the lookup key is longer or compound or whatever depending on the cache format.

5. Even a new hashing algorithm requires work on clients etc. It’s not that you get some backwards-compatibility for free. It just cannot be backwards-compatible in my opinion, no matter which approach we take. That’s why I believe some magic time for the switch causes the least amount of trouble. You leave the old world untouched and working.

If these are general concerns, I’m completely with you. But I don’t think that they only apply to location-based addressing. That’s how I interpreted your message. I could be wrong. Happy to read your explanations. :-)

Why do I care about this?

1. The load will become a problem at some point.
   
2. These crawlers and the current “AI” in general are breaking the rules. I am supposed to be paying for every little thing, I get sued for “piracy”. But apparently, these rules only apply to me. If I had more money, I could break them. Fuck that.
   
3. I simply don’t want it. Period.

This probably means that I can no longer host my own website. I don’t want to deploy something like Anubis, because that ruins the whole thing: I want it to be accessible from ancient browsers, like OS/2 or Windows 3.11.

I’ll keep an eye on it for a while. Maybe try to block some IPs.

Sooner or later, I’ll take the website down and shift everything to Gopher.

@lyse@lyse.isobeef.org “Advanced”, well, probably more “mature”. There aren’t a ton of crazy features and that icon thing is the largest code addition in the last 10 years. %)

Speaking of OS/2 … I just realized that Windows 3.x didn’t have icons, either. If I’m not mistaken, this only got added in Windows 95. In other words, OS/2 had this feature before Windows did, because at least OS/2 2.1 from 1993 had icons. Who would have thunk.

(Now I kind of want to know which system really introduced this feature.)

@movq@www.uninformativ.de Ah, okay! That’s why it’s in such an advanced state. :-)

Nice, I never came in contact with OS/2.

@lyse@lyse.isobeef.org Oh, huh, maybe it was just my GNOME 2 themes back then that didn’t show the icon. 🤔

I like the looks of your window manager. That’s using Wayland, right?

Oh, no. It’s still X11. All my recent Wayland comments resulted from me trying to switch, but I think it’s still too early. Being unable to use QEMU (because it can’t capture the mouse pointer) is a pretty big blocker for me. This is completely broken, it just happens to be unnoticeable with modern guest OSes, so it’s probably not a priority for devs.

(Not to mention that I would have to fork and substantially extend dwl in order to “replicate” my X11 WM. And then, after having done that, I’d have to follow upstream Wayland development, for which I don’t have the resources. Things would need to slow down before I can do that.)

all that wasted space of the windows not making use of the full screen!!!1

Heh. I’ve been using tiling WMs for ~15 years now, so it’s actually kind of refreshing to see something different for a change. 😅

Probably close to the older Windowses.

That particular theme is a ripoff of OS/2 Warp 3: https://movq.de/v/6c2a948882/s.png 😅

We ran some similar brownish color scheme (don’t recall its name) on Win95 or Win98

Oh god. Yeah, I wasn’t a fan of those, either. 🥴

@movq@www.uninformativ.de According to this screenshot, KDE still shows good old application icons: https://upload.wikimedia.org/wikipedia/commons/9/94/KDE_Plasma_5.21_Breeze_Twilight_screenshot.png

And GNOME used to have them, too: https://upload.wikimedia.org/wikipedia/commons/9/9f/Gnome-2-22_%284%29.png

I like the looks of your window manager. That’s using Wayland, right? The only thing on this screenshot to critique is all that wasted space of the windows not making use of the full screen!!!1 At least the file browser. 8-)

This drives me nuts when my workmates share their screens. I really don’t get it how people can work like that. You can’t even read the whole line in the IDE or log viewer with all the expanded side bars. And then there’s 200 pixels on the left and another 300 pixels on the right where the desktop wallpaper shows. Gnaa! There’s the other extreme end when somebody shares their ultra wide screen and I just have a “regularish” 16:10 monitor and don’t see shit, because it’s resized way too tiny to fit my width. Good times. :-D

Sorry for going off on a tangent here. :-) Back to your WM: It has the right mix of being subtle and still similar to motif. Probably close to the older Windowses. My memory doesn’t serve me well, but I think they actually got it fairly good in my opinion. Your purple active window title looks killer. It just fits so well. This brown one (https://www.uninformativ.de/blog/postings/2025-07-22/0/leafpads.png) gives me also classic vibes. Awww. We ran some similar brownish color scheme (don’t recall its name) on Win95 or Win98 for some time on the family computer. I remember other people visting us not liking these colors. :-D

I was drafting support for showing “application icons” in my window manager, i.e. the Firefox icon in the titlebar:

https://movq.de/v/0034cc1384/s.png

Then I realized: Wait a minute, lots of applications don’t set an icon? And lots of other window managers don’t show these icons, either? Openbox, pekwm, Xfce, fvwm, no icons.

Looks like macOS doesn’t show them, either?!

Has this grown out of fashion? Is this purely a Windows / OS/2 thing?

Only figured this out yesterday:

pinentry, which is used to safely enter a password on Linux, has several frontends. There’s a GTK one, a Qt one, even an ncurses one, and so on.

GnuPG also uses pinentry. And you can configure your frontend of choice here in gpg-agent.conf.

But what happens when you don’t configure it? What’s the default?

Turns out, pinentry is a shellscript wrapper and it’s not even that long. Here it is in full:

#!/bin/bash

# Run user-defined and site-defined pre-exec hooks.
[[ -r "${XDG_CONFIG_HOME:-$HOME/.config}"/pinentry/preexec ]] && \
        . "${XDG_CONFIG_HOME:-$HOME/.config}"/pinentry/preexec
[[ -r /etc/pinentry/preexec ]] && . /etc/pinentry/preexec

# Guess preferred backend based on environment.
backends=(curses tty)
if [[ -n "$DISPLAY" || -n "$WAYLAND_DISPLAY" ]]; then
        case "$XDG_CURRENT_DESKTOP" in
        KDE|LXQT|LXQt)
                backends=(qt qt5 gnome3 gtk curses tty)
                ;;
        *)
                backends=(gnome3 gtk qt qt5 curses tty)
                ;;
        esac
fi

for backend in "${backends[@]}"
do
        lddout=$(ldd "/usr/bin/pinentry-$backend" 2>/dev/null) || continue
        [[ "$lddout" == *'not found'* ]] && continue
        exec "/usr/bin/pinentry-$backend" "$@"
done

exit 1

Preexec, okay, then some auto-detection to use a toolkit matching your desktop environment …

… and then it invokes ldd? To find out if all the required libraries are installed for the auto-detected frontend?

Oof. I was sitting here wondering why it would use pinentry-gtk on one machine and pinentry-gnome3 on another, when both machines had the exact same configs. Yeah, but different libraries were installed. One machine was missing gcr, which is needed for pinentry-gnome3, so that machine (and that one alone) spawned pinentry-gtk …

@prologic@twtxt.net interesting, a Chinese pickup truck. Hmm, I would very interested to know your thoughts about it 2-3 years from now.

@bender@twtxt.net That was one of the inputs into my research 🧐 So that’s already factored in. We bought our new truck (2025 GWM Canon) recently to replace the ‘ol 2nd hand Nissan Navara we bought that just had too many things go wrong with it, and I don’t have time or energy to learn to be a diesel mechanic haha 🤣 – So yes, the SCT-16 has a Tare (unladen weight) of 2150Kg and a maximum legal (ATM) weight of 2,800Kg.

@prologic@twtxt.net that looks like a beautiful camper! What kind of truck do you have to pull it? That could be the next thing you might need to focus on. I mean, 2,800kg gross is not feather light!

The WM_CLASS Property is used on X11 to assign rules to certain windows, e.g. “this is a GIMP window, it should appear on workspace number 16.” It consists of two fields, name and class.

Wayland (or rather, the XDG shell protocol – core Wayland knows nothing about this) only has a single field called app_id.

When you run X11 programs under Wayland, you use XWayland, which is baked into most compositors. Then you have to deal with all three fields.

Some compositors map name to app_id, others map class to app_id, and even others directly expose the original name and class.

Apparently, there is no consensus.

@movq@www.uninformativ.de Yeah, it’s a shitshow. MS overconfirms all my prejudices constantly.

Ignoring e-mail after lunch works great, though. :-)

Our timetracking is offline for over a week because of reasons. The responsible bunglers are falling by the skin of their teeth: https://lyse.isobeef.org/tmp/timetracking.png

1. The error message neither includes the timeframe nor a link to an announcement article.
   
2. The HTML page needs to download JS in order to display the fucking error message.
   
3. Proper HTTP status codes are clearly only for big losers.
   
4. Despite being down, heaps of resources are still fetched.
   

I find it really fascinating how one can screw up on so many levels. This is developed inhouse, I’m just so glad that we’re not a software engineering company. Oh wait. How embarrassing.

We’re entering the “too hot to think”-season in 3, 2, 1 … and we’re live!

Saw this on Mastodon:

https://racingbunny.com/@mookie/114718466149264471

18 rules of Software Engineering

1. You will regret complexity when on-call
   
2. Stop falling in love with your own code
   
3. Everything is a trade-off. There’s no “best” 3. Every line of code you write is a liability 4. Document your decisions and designs
   
4. Everyone hates code they didn’t write
   
5. Don’t use unnecessary dependencies
   
6. Coding standards prevent arguments
   
7. Write meaningful commit messages
   
8. Don’t ever stop learning new things
   
9. Code reviews spread knowledge
   
10. Always build for maintainability
    
11. Ask for help when you’re stuck
    
12. Fix root causes, not symptoms
    
13. Software is never completed
    
14. Estimates are not promises
    
15. Ship early, iterate often
    
16. Keep. It. Simple.
    

Solid list, even though 14 is up for debate in my opinion: Software can be completed. You have a use case / problem, you solve that problem, done. Your software is completed now. There might still be bugs and they should be fixed – but this doesn’t “add” to the program. Don’t use “software is never done” as an excuse to keep adding and adding stuff to your code.

@aelaraji@aelaraji.com I use Alt+. all the time, it’s great. 👌

FWIW, another thing I often use is !! to recall the entire previous command line:

$ find -iname '*foo*'
./This is a foo file.txt

$ cat "$(!!)"
cat "$(find -iname '*foo*')"
This is just a test.

Yep!

Or:

$ ls -al subdir
ls: cannot open directory 'subdir': Permission denied

$ sudo !!
sudo ls -al subdir
total 0
drwx------ 2 root root  60 Jun 20 19:39 .
drwx------ 7 jess jess 360 Jun 20 19:39 ..
-rw-r--r-- 1 root root   0 Jun 20 19:39 nothing-to-see

@prologic@twtxt.net … or just bullshit.

I’m Alex, COO at ColdIQ. Built a $4.5M ARR business in under 2 years.

Some “C-level” guy telling people what to do, yeah, I have my doubts.

On my blog: Real Life in Star Trek, Descent, part 2 https://john.colagioia.net/blog/2025/05/29/descent-2.html #scifi #startrek #closereading

@prologic@twtxt.net Mosaic (2.7) works fine, I maintain that package in the AUR and test my website regularly. 😅

https://movq.de/v/d7787d6fb1/s.png

Maybe you’ll enjoy this as well:

I still have one of my first modems, a Creatix LC 144 VF:

I think this was the modem that I used when I first connected to the internet, but I’m not sure.

I plugged it in again and it still works:

The firmware appears to be from 1994, which sounds about right. I don’t think we had internet access before that. We certainly did use local mailboxes, though. (Or BBS’s, as you might call them.)

I now want to actually use that modem again. For the moment, I can only use a phone to dial into it, I lack a second modem to actually establish a connection. Here’s a video:

Not spectacular, but the modem does answer after me entering ATA.

I bought another cheap old modem on eBay and am now waiting for it to arrive. Once it’s here, I want to simulate an actual dial-up session, hopefully from OS/2 or Windows 3.x.

One of the nicest things about Go is the language itself, comparing Go to other popular languages in terms of the complexity to learn to be proficient in:

• Go: 25 keywords (Stack Overflow); CSP-style concurrency (goroutines & channels)
  
• Python 2: 30 keywords (TutorialsPoint); GIL-bound threads & multiprocessing (Wikipedia)
  
• Python 3: 35 keywords (Initial Commit); GIL-bound threads, asyncio & multiprocessing (Wikipedia, DEV Community)
  
• Java: 50 keywords (Stack Overflow); threads + java.util.concurrent (Wikipedia)
  
• C++: 82 keywords (Stack Overflow); std::thread, atomics & futures (en.cppreference.com)
  
• JavaScript: 38 keywords (Stack Overflow); single-threaded event loop & async/await, Web Workers (Wikipedia)
  
• Ruby: 42 keywords (Stack Overflow); GIL-bound threads (MRI), fibers & processes (Wikipedia)
  

@bender@twtxt.net 5, 4, 3, 2, 1 🤣

@bender@twtxt.net I’m not sure this is accurate, if you lookup mine:

$ whois shortcircuit.net.au 2>&1 | grep -i creat
created:      1986-03-05

I think this has to be the registrar’s creation date no? 🤔

On my blog: Free Culture Book Club — Pilogy, part 2 https://john.colagioia.net/blog/2025/05/03/pilogy-2.html #freeculture #bookclub

Can you automate the drawing with a script? On X11, you can:

#!/bin/sh

# Position the pointer at the center of the dot, then run this script.

sleep 1

start=$(xdotool getmouselocation --shell)
eval $start

r=400
steps=100
down=0

for step in $(seq $((steps + 1)) )
do
    # pi = 4 * atan(1)
    new_x=$(printf '%s + %s * c(%s / %s * 2 * (4 * a(1)))\n' $X $r $step $steps | bc -l)
    new_y=$(printf '%s + %s * s(%s / %s * 2 * (4 * a(1)))\n' $Y $r $step $steps | bc -l)

    xte "mousemove ${new_x%%.*} ${new_y%%.*}"
    if ! (( down ))
    then
        xte 'mousedown 1'
        down=1
    fi
done

xte 'mouseup 1'
xte "mousemove $X $Y"

Interestingly, you can abuse the scoring system (not manually, only with a script). Since the mouse jumps to the locations along the circle, you can just use very few steps and still get a great score because every step you make is very accurate – but the result looks funny:

🥴

@xuu@txt.sour.is or @kat@yarn.girlonthemoon.xyz Do either of you have time this weekend to test upgrading your pod to the new cacher branch? 🤔 It is recommended you take a full backup of you pod beforehand, just in case. Keen to get this branch merged and to cut a new release finally after >2 years 🤣

@prologic@twtxt.net This shi_ is as fun as it is frustrating! 😆 the bot is poking at me from a different ASN now, Alibaba’s.

1. Short term solution: I’ve geo-locked my Timeline instance since I’m the only one using it (and I only do so for reading twts when I’m away from terminal).
   
2. Long term: I took a look at your Caddy WAF but couldn’t figure things out on my own; until then, I’ll be poking at Caddy-Defender, maybe throw in a Crowdsec for lols… #FUN

I use restic and Backblaze B2 for offline backup storage at a cost of $6/TB/month. I don’t backup my entire ~20TB NAS and its datasets however, so I’m only paying about ~$2/month right now. I only backup the most important things I cannot afford to lose or annot re-created.

@prologic@twtxt.net in this case it isn’t vendor lock-in. I believe they do it because the carrier “eats” the costs (the interest part of the instalments). The phones are fully unlocked.

I will return the two we got for parents, and re-buy them again, this time on an instalment plan. Why pay $2,000 up front when you can split the cost in convenient monthly payments, all interest free, right?

I updated wordwrap.[ch] to more closely match the interface for string(2); it’s now just that plus a margin. I also updated litclock and marquee to match. http://a.9srv.net/src/index.html

@prologic@twtxt.net, from IRC:

1. Saving preferences is failing. Specifically trying to save “Open Links” on the same window. For sure it isn’t happening. Check errors on browser’s console.
   
2. Search results pagination is broken. Search for “twtxt.net” and see it. Also, picking oldest/newest makes no difference on that search query.

So I re-write this shell alias that I used all the time alias dkv="docker rm" to be a much safer shell function:

dkv() {
  if [[ "$1" == "rm" && -n "$2" ]]; then
    read -r -p "Are you sure you want to delete volume '$2'? [Y/n] " confirm
    confirm=${confirm:-Y}
    if [[ "$confirm" =~ ^[Yy]$ ]]; then
      # Disable history
      set +o history

      # Delete the volume
      docker volume rm "$2"

      # Re-enable history
      set -o history
    else
      echo "Aborted."
    fi
  else
    docker volume "$@"
  fi
}

I’ve just ordered a new toy! A ProDesk 600 G4 Mini with a Core i5-8500T, 32Go of DDR4 RAM and 256Go SSD storage. A cheaper alternative to an 8GB RPi5 + Argon one v3 m.2 RPi case kit (NVME not included) 🤷. It should be here by Friday 🤞

On my blog: Free Culture Book Club — Nose Ears, part 2 https://john.colagioia.net/blog/2025/04/05/nose-ears-2.html #freeculture #bookclub