2025-04-03 07:20:48
I love blogging! Writing about a couple of things that interest me every day or so is one of the highlights of my life. That so many of you now read my ramblings as well is that much more lovely. I don’t dare take it for granted.
Unfortunately, there are also topics that impact me, but that I don’t want to give any more oxygen. For a few years it was blockchain, in part because I had to deal with so much of it from potential clients at work. It was an unmitigated disaster, and didn’t even resolve any of the problems it purported to solve. Fraud, transparency, accountability, lack of government intervention (insofar as people thought this was a desirable trait), it was all a wash… and a predictable one too for anyone with the most basic understanding of mathematics, finance, and history.
I felt this compulsion to write massive long screeds about blockchain, about crypto-“currency”, about the utter nonsense that was NFTs, and of their negative impacts and pain they wrought. How the tech had done more to jade me on the tech industry than even its nonchalant, cavalier attitude to digital privacy in the twenty-first century.
But during the peak it felt as though I was shouting into the void. Discussing the topic left me feeling frustrated, despondent, and tired. You have to look out for number one first, so the saying goes, so I made the executive decision to not air anything more about it, for my own sake. It was a selfish decision given that we needed as many voices pushing back on this crap as we could get, but it was one I felt I had to make.
🌲 🌲 🌲
genAI is the next chapter in this calamitous cavalcade of computerised crapulance. This technology, whatever perceived or incidental benefit it has, slops out such enormous and uncaptured externalities that it affects all of us. The slave labour in Africa used to classify data. Our servers creaking under the load of their indiscriminate scrapers, to say nothing of our costs. Open source projects—the ones that deliver so much of the code necessary to make them work—being washed of their licences and clogged with spam. Creative outlets being forced to close under mountains of spam and genAI slop. I waste so much of my job now dealing with automated abuse alerts it beggars belief. I could link to examples of each of these, but I’m tired, and I’m sure you’ve read the same stories.
The financials underpinning these models and companies also aren’t sustainable. The poster children of the movement are burning through billions of dollars of investor funds and capital a year, and have yet to turn a profit for this alleged transformative tech. Some of their backers have even scaled back their data centre rollouts. The fundamentals simply aren’t there, though my ethical concerns and frustrations would still exist even if they did solve the financial problems they claim.
It also didn’t need to be this way; the idea of inevitability is a con peddled by those who want to push their narrow, self-serving view of the future. The billions being spent researching and developing this tech could literally have been allocated in any other number of ways. Say, I dunno, like building out more clean energy, or throwing slavish amounts of money at RDNA to solve malaria, or… here’s an idea… paying the artists, musicians, and writers who’s work makes our lives worth living. What about those open source tools you use, how about throwing them a bone?
AI researches themselves are also fed up. I’ve received email from many of you frustrated that legitimate machine learning applications in scientific and medical research are being lumped in with slop generators and shitty “summarising” tools that absolutely don’t. Much as the blockchain bros tainted the innocuous and useful term crypto, so too have these charlatans with the concept of AI. And don’t get me started on the gamers who just want a bit of escapism after a hard and unforgiving day at work without scalped GPUs used instead to shit out industrial art simulacrums that are an offence to life itself. My god Miyazaki, I’m so, so sorry.
genAI is the most selfish technology I’ve ever seen in my lifetime. I resent the mental CPU cycles I’ve spent thinking about it, how sycophantic and uncritical so many bloggers and journalists have been at discussing the tech’s problems (I’ve had to unsubscribe to a dishearteningly large number of them), and am frustrated that it reached such a boiling point again that I’ve felt compelled to write about it too.
Life is too short for this crap!
Back to BSD, coffee, retrocomputers, and travel soon; promise.
By Ruben Schade in Sydney, 2025-04-03.
2025-04-02 05:41:10
Most of you are likely aware of the Ship of Theseus thought experiment. If not, it grapples with the question of whether something is still the same object if all its parts have been replaced over time. In the original example, imagine if the sails were replaced with a new canvas, then over time the hull due to rotting wood. Is it the same ship, if it shares nothing physical with the original?
How people answer that question is interesting. My circle of family, friends, and colleagues over time are split between saying it is and it isn’t a new ship. Maybe have a think yourself where you land here. Or sail.
I’m sympathetic to both views. I can see the ship isn’t physically configured with any of the original components, even if replacements have been created that mimic their materials, behaviour, and appearance. I can see how puritans would claim it’s not the same ship from the moment a single thing is replaced, let alone half or most of the components. That’s why when people buy something precious to them, they may keep an extra copy locked away in “mint” condition.
Personally, is a word with ten letters. I lean towards saying it’s the same ship. It’s occupying the same space in about the same way. This is informed by the fact that some of those replacement parts did share the boat with original components at some point. These are two key attributes that separate a Ship of Theseus from a replica, which is its own distinct object with its own history. Granted this mental gymnastics’ only works if one projects context onto to the object, otherwise a ship with replaced components could be confused for a duplicate.
I started down this train of thought again recently after contrasing my recent experience seeing the NSWGR 3526 steam locomotive, and my lifelong dream purchase of a Commodore C116. Yes, I did it, I managed to think about Edwardian engineering and retrocomputers in the same sentence, and without any steampunk of any kind.
The steam locomotive is more than a century old, and has had numerous components rebuilt, upgraded, or entirely replaced. The locomotive could have sat static in a museum without fixes, but would never have operated again. This would have satisfied the objective Ship of Theseus, but the sights, sounds, and experience of witnessing that locomotive operating as she would have a hundred years ago was additive itself. The restoration work was done respectfully, and to an extent that people standing around us who were alive when she was in mainland service couldn’t wipe the smiles off their faces. This was the locomotive they grew up with.
By contrast, the Commodore 116 is entirely stock. It has had no modifications done, no upgrades, no replacements, and yet is fully functional in 2025 (a shocking achievement when you consider most computers aren’t designed to outlive a two-year warranty). I do admit it makes me feel great knowing this is the same physical basket of components that would have been assembled, bought, and used in 1984. Had there been a C116 that underwent significant rework, and a stock C116, I would likely choose the latter if I could afford it.
But keen-eyed readers among you may have spotted a small cheat in the photo of the C116: A RetroTINK 2x upscaler. I don’t have an original Commodore CRT (nor any CRTs on account of them giving me headaches today). But I’m able to use this machine, in its original condition, thanks to being able to pipe S-Video and audio out of the C116 and into a converter box. While the C116 is the original, how I’m using it certainly isn’t.
Despite the steam loco being modified, and the C116 running stock, a case could be made that the experience of the former is more authentic than the latter. Which to me gets to the core issue underlying the Ship of Theseus. To use another overused phrase, seeing is believing.
Uh oh, I can sense a part two.
By Ruben Schade in Sydney, 2025-04-02.
2025-04-01 06:51:01
After months of saving and indecision, we’ve finally booked the hotels and flights:
Flying into Tōkyō, and spending a few days there. We generally prefer flying to Kansai because Ōsaka is our favourite Japanese city, but we got some jaw-dropping rates with JAL. Maybe it has something to do with the Expo in Ōsaka right now.
Taking the Hikari Shinkansen to Nagoya (it’s a bit cheaper than the Nozomi, and it’s pretty close) and staying there a week. I’ve only ever zipped past, so I’m keen to explore all the museums and the port.
Travelling on one of those slow tourist trains up to Takayama in the mountains of Gifu Prefecture, and spending a few days hiking and exploring. This is the part I’m looking forward to the most.
Travelling further north to Takaoka in Toyama Prefecture to explore the Sea of Japan coast for a week for the first time. The plan is to do day trips to Imizu and greater Toyama city further east. There are a lot of… second hand hardware, book, and music stores up there too (cough).
Back to Tōkyō on the Hokuriku Shinkansen to hang out for a few more days, then flying out.
We still badly want to explore Hiroshima and Okinawa further south; not least because we know a few of you living down there! But the budget and timing didn’t work out this time.
By Ruben Schade in Sydney, 2025-04-01.
2025-04-01 06:01:43
I read a lot of technical newsletters, product announcements, feature update notices, and related messages as part of my job. I guard my personal email like a hawk, to the point where I have a filter that matches on the word unsubscribe. At work, I have to keep tabs on our suppliers, customers, partners, leads, and other parties I can include in a list of things on a blog that I keep tabs on.
Receiving these messages, and even helping work write ours, gives me insight into what makes a good newsletter, and a bad one.
Here are some examples of what not to do:
Avoid massive banner images, or banner images at all if you can avoid them. Email clients all block images by default now, so we just get a massive blank space before your text starts.
Don’t venture into spam territory. I don’t need to hear from your business almost every day.
Don’t use clickbait language. I know your marketing team is telling you to, but respect your readers instead.
Don’t send cutesy messages shortly after a massive outage or customer-impacting event. Like a fine cheese, it grates.
Don’t abuse your contact lists by subscribing us to your new newsletter we didn’t ask for. This is poor form.
And what you should do:
Have a useful subject line. Don’t say $FOOBAR Newsletter, say $FOOBAR Newsletter: Feature X, Update Y.
Be brief. Respect the time of your readers. They can click through for more information.
Remind us what you do, or what you device or service does, even if just briefly. Don’t say $GLAVEN, say $GLAVEN, the cloud orchestration tool. It may be obvious to you, but it’s best to assume people receiving your newsletter are doing so for the first time, or forgot who you are.
Use appropriate alt text for images. As mentioned above, nobody’s email client downloads images automatically now, so we see that alt text first. The number of times that’s just a filename, a string of jibberish, a UUID, or an internal note that probably shouldn’t have gone out is surprising.
If you include a copyright date, make sure it’s the right year. I still get emails from a US-based telco that says “2021” in the footer.
Give us a clear, unambiguous unsubscribe link. I trust people who make this easier, unlike certain Redmond-based newsletters that I’ve since trained my spam filters against.
By Ruben Schade in Sydney, 2025-04-01.
2025-03-31 18:07:00
In all my years of collecting, writing software for, tinkering with, and learning about 8-bit home computer hardware from Commodore, I never, never thought I’d see the day where a 1984 Commodore 116 would be on the table with my other 264 machines! Almost two decades of saved eBay searches finally pinged me last night, and for a price that was shockingly fair.
Here she is at a local coffee shop after picking her up. Technically a hardware store that people in Sydney might recognise, but I thought the coffee break area had surprisingly good lighting:
Commodore fans may understand the significance of this machine, but it’s also important to me for silly, personal reasons.
The Commodore 116 formed part of the company’s 264 line of machines in 1984. Commodore CEO Jack Tramiel had commissioned a project to build a low-cost series of computers to replace the VIC-20 in Commodore’s home computer lineup, and to compete with Sinclair’s ZX Spectrum. A prototype 364 with a numeric keypad was shown at CES but never released, and Tramiel left the company soon after.
The 264 line was notable for being powered by Dave Diorio’s TED chip, which combined sound and graphics onto the same die. While its two-channel, four-octave sound capabilities and lack of sprites paled in comparison to the Commodore 64’s legendary VIC-II and SID, it sported a wider 121 colour pallete. This was paired with a vastly improved BASIC 3.5 with proper graphics and sound subroutines. It was this combination of features that made me fall in love with the 264s in the early 2000s, as we’ll cover in the next section.
Three machines were ultimately released under the stewardship of Bil Herd of Commodore 128 fame:
The flagship Commodore Plus/4 with 64 KiB of memory. This was the original 264, but repackaged as a business productivity machine with included software in ROM. I’ve said its industrial design is perhaps my favourite of any 8-bit machine.
The Commodore 16 with a reduced 16 KiB of memory and no user port. Bizarrely, Commodore released this in a breadbin case like the VIC-20 and C64, which likely contributed to sales confusion.
The entry-level Commodore 116, which paired the C16’s features in a smaller Plus/4 case, and a rubber keyboard like the Spectrum.
While the Plus/4 and C16 saw better success in Europe than they did in the US, the C116’s rubbery keyboard and limited memory contributed to it being a sales flop. I’ve read reports of them selling around 51,000 units total, which was probably fewer than Commodore sold of the C64 at lunch during its peak.
I always thought the C116 was the same size as the Plus/4, but as you can see in the family reunion photo below, the C116 is clearly smaller! No wonder they thought this would target the Sinclair. It’s absolutely, gobsmackingly adorable, and I can’t stop staring at it.
My history with the 264 line goes back to the early 2000s, when my parents saw I was showing an interest in computer history. They scoured eBay and picked up three machines for me: an NTSC Plus/4 and C16, and a breadbin PAL C64. The combined $50 they spent would barely cover a third-party power brick now, which is incredible to think about.
I’ve mentioned on my Retro Corner that I ended up using the C16 more than any of the others. I didn’t know about the history of these machines or how their capabilities differed; I just assumed the C16 and Plus/4 were better because they were “newer”. Perhaps there’s something to the theory that Commodore’s sales and marketing teams had no idea what they were doing with these machines, when even a budding but uninformed enthusiast like me was equally confused.
I loved the design of the Plus/4, and I ended up using the C16’s fun breadbin keyboard to write all my earliest 8-bit BASIC programs. It frustrated me at the time that the C64 had issues with the same commands; something I now understand because the C64’s BASIC is so limited.
The 264 line made 8-bit computing accessible to me, and I credit them with introducing me to this whole other world of computers I didn’t know existed when I grew up on the 486 and DOS. The C116 was the last missing piece in this beloved family of machines, and also represents the first PAL machine I have in the lineup. It’ll be so much fun trying demos on this!
10 PRINT "Does this coffee break area have a RetroTINK!?"
20 GOTO 10
I’ve tested it and it works! Stay tuned for more testing and exploration :).
By Ruben Schade in Sydney, 2025-03-31.
2025-03-31 05:50:10
I’ve talked again recently about my frustrations with certain Linux desktop fans directing their ire not at commercial software companies that are making the lives of people who don’t have a choice that much worse, but instead towards… the people who’s lives they’re making that much worse. Even if you think you’re in the right, it’s not how one wins friends and influences people, as a certain Dale Carnegie would have said.
But I have taken pains to point out in such posts that my reservations about the abrasive attitude of some amateur Linux users doesn’t extend to their underlying OS of choice. This I think is worth underscoring from time to time, just as it’s important not to blame electronic victims.
Linux on the desktop is awesome for me for a bunch of reasons, many of which are shared with FreeBSD and NetBSD, my favourite OSs. But the one reason I want to discuss here specifically is permissions. And for once I’m not talking about chmod(1), or chown(1), or chattr(1).
Rebooting my FreeBSD machine into Fedora to play a game with Clara last week made me realise something that seems so obvious in retrospect, but I only just recognised. I don’t need permission to use this computer. I turn it on, and use whatever I want, however I want.
When I install certain software for people, or OSs at work, there’s infrastructure in place to ensure you have permission to use it. There are product keys, serial numbers, activation servers, per seat or per core licensing, price lists full of complicated SKUs, and more. In some cases entire additional layers of hosting are required to install, activate, report, and track this stuff; hosting that could be put towards something technically productive instead.
It’s not even just regarding a proof of purchase, this administrative overhead often extends over the lifetime of your install. Your machine needs to be phoning home to somewhere, it needs permission to download and use certain features, it’s regularly polled to ensure it’s operating as the vendor likes, and so on. Your machine has to be in the system. Or in other words, it needs permission.
Linux isn’t entirely free of this concept either. My home Fedora partitions need permission to access an update server to remain secure, for example. But that’s a process I ultimately have agency over; I could potentially swap that repo out for something else if I wanted. Point is, I don’t need to seek permission from the Fedora Project, or the FreeBSD Foundation, and the NetBSD Foundation, to use this computer. I install them, I use them, and I update them.
That perceived lack of control, or the need to request permission, is what has made macOS feel increasingly alien to me of late, and what has made Windows feel downright hostile in a way that’s hard to put into words. I’ve talked before about feeling more like a guest on Macs thesedays, and how cleaning up a Windows machine has gone from removing OEM bloatware to disabling core features of the OS.
These OSs have always presented a Faustian bargain to an extent, but whether it’s Apple forcefully and quietly re-enabling toxic features, or Windows predictably requiring new installs and accounts to be registered with a Microsoft profile, it’s been amped up to 11 of late. Yes it’s always been bad, but now it’s worse.
On my FreeBSD desktop, my NetBSD laptops, my Fedora game machine, or any of my servers, this is an absolute non issue. I can turn my PC on, and be answerable to nobody. I don’t need to report in. I don’t need permission to unlock the door, so to speak. I just start them, and go about my day.
I can’t tell you how relieved this makes me.
By Ruben Schade in Sydney, 2025-03-31.
