Welcome to another month of rambling status reports. Not much in terms of technology this month, my work at Codethink is still focused on proprietary corporate infrastructure, and the weather is too nice to spend more time at a computer than necessary. Somehow I keep reading things and thinking about stuff though, and so you can read some of these thoughts and links below.
Is progress going backwards?
I’ve been listening to
The Blindboy Podcast
from the very beginning. You could call this a “cult podcast” since there isn’t a clear theme, the only constant is life, narrated by an eccentric Irish celebrity. I’m up to the episode
“Julias Gulag”
from January 2019, where Blindboy mentions a Gillette advert of that era which came out against toxic masculinity, very much a progressive video in which there wasn’t a single razor blade to speak of. And he said, roughly,
“I like the message, and the production is excellent, but I always feel uneasy when this type of “woke” video is made by a huge brand because I don’t think the board of directors of Proctor & Gable actually give a shit about social justice.”
This made me think of an excellent Guardian article I read last week, by Eugene Healey entitled
“Marketing’s ‘woke’ rebrand has ultimately helped the far right”
, in which he makes largely the same point, with six years worth of extra hindsight. Here are a few quotes but the whole thing is worth reading:
Social progress once came hand-in-hand with economic progress. Now, instead, social progress has been offered as a
substitute
for economic progress.
Through the rear window it’s easy to see that the backlash was inevitable: if progressive values could so easily be commodified as a tool for selling mayonnaise, why shouldn’t those values be treated with the same fickleness as condiment preferences?
The responsibility we bear now is undoing the lesson we inadvertently taught consumers over this era. Structural reform can’t be achieved through consumption choices – unfortunately, we’re all going to have to get dirt under our fingernails.
We are living through a lot of history at the moment and it can feel like our once progressive society is now going backwards. A lot of the progress we saw was an illusion anyway. The people who really hold power in the world weren’t really about to give anything up in the name of equality, and they still aren’t. World leaders were still taking private jets to conferences to talk about the climate crisis, and so on. The 1960s USA seemed like a place of progress, and then they went to war in Vietnam.
As Eugene Healey says towards the end of his piece, one positive change is that it’s now obvious who the bad guys are again. Dinold Tromp appears on TV every time I look at a TV, and he dresses like an
actual supervillain
. Mark Zuckerburg is
trying to make his AI be more right-wing
. Gillette is back to making adverts which are short videos of people shaving, because Gillette a brand that manufactures razors and wants you to buy them. It is not a social justice movement!
The world goes in cycles, not straight lines. Each new generation of people has to ignore most of what we learn from teachers and parents, and figure everything out for ourselves the hard way, right?
For technologists, it’s been frustrating to spend the last decade telling people to be wary of Apple, Amazon, Google, Meta and Microsoft, and being roundly ignored. They are experts in making convenient, zero cost products, and they are everywhere. Unless you’re an expert in technology or economics, then it wasn’t obvious what they have been working towards, which is the same thing it always was, the same that drove everything Microsoft did through the 1990s: accumulating more and more money and power.
You don’t get very far if you tell this story to some poor soul who just needs to make slides for a presentation, especially if your suggestion is that they try LibreOffice Impress instead.
When 2025 kicked off, CEOs of all those Big Tech companies attended the inauguration of Dinald Tromp and donated him millions of dollars, live on international news media. In the long run I suspect this moment will have pushed more people towards ethical technology than 20 years of campaigning about
nonfree JavaScript
.
Art, Artificial Intelligence and Idea Bankrupcy
Writing great code can be a form of artistic expression. Not all code is art, of course, just as an art gallery is not the only place you will find paint. But if you’re wondering why some people release groundbreaking software for free online, it might help to view it as an artistic pursuit.
I took a semi retirement from volunteer open source contributions
back in October of last year
, having got to a point where it was more project management than artistic endeavour. In an ideal world I’d have some time to investigate new ideas, for example in desktop search or automated GUI testing, and publish cool stuff online. But there are two blockers. One blocker is that I don’t have the time. And the other, is that the open web is now completely overrun with data scrapers, which somehow ruins the artistic side of publishing interesting new software for free.
We know that reckless data scraping by Amazon, Anthropic, Meta and Microsoft/OpenAI (those US tech billionaires again), plus their various equivalents in China, is causing huge problems for open source projects and other non-profits. It has led The Wikimedia Foundation to declare this month that “
Our content is free, our infrastructure is not
“. And Ars Technica also published a good
summary of the situation
.
Besides the bandwidth costs, there’s something uncomfortable about everything we publish online being immediately slurped into the next generation of large language model. If permissive software licenses
lead to extractive behaviour
, then AI crawlers are that on steroids. LLMs are incredibly effective for certain use cases, and one such use case is “copyright laundering machines”.
Software licensing was a key part of the discussion around ethical technology when I first discovered Linux in the late 1990s. There was a sense that if you wrote innovative code and published it under the GNU GPL, you were helping to fight the evils of Big Tech, as the big software firms wouldn’t legally be able to incorporate your innovation into their products without releasing their source code under the same license. That story is spelled out word-for-word in Richard Stallman’s article
“Copyleft: Pragmatic Idealism”
. I was never exactly a disciple of Richard Stallman, but I did like to release cool stuff under the GPL in the past, hoping that in a small way it’d work towards some sort of brighter future.
I was never blind to the limitations of the GPL. It requires an actual threat of enforcement to be effective, and historically only a few groups like the
Software Freedom Conservancy
actually do that difficult legal work. Another weakness in the overall story was this: if you have a big pile of cash, you can simply
rewrite
any innovative GPL code. (This is how we got Apple to pay for LLVM).
Long ago I read the book
“Free as in Freedom”
. It’s a surprisingly solid book which narrates Richard Stallman’s efforts to form a rebel alliance and fight what we know today as Big Tech, during which he founds the GNU Project and invents the GPL. It is only improved in version 2.0 where Stallman himself inserting pedantic corrections into Sam Williams’s original text such as
“This cannot be a direct quote because I do not use
fucking
as an adverb”
. (The book and the corrections predate him famously
being cancelled
in 2019). He later becomes frustrated at having spent a decade developing an innovative, freely available operating system, only for the media and the general public to give credit to Linus Torvalds.
Right now the AI industry is trying to destroy copyright law as we know it. This will have some interesting effects. The GPL
depends on copyright law
to be effective, so I can only see this as the end of the story for software licensing as a way to defend and ensure that the inventors of cool things get some credit and earn money. But let’s face it, the game was already up on that front.
Sustainable open source projects — meaning those where people actually get paid do all the work that is needed for the project to succeed — can exist and do exist. We need independent, open computing platforms like
GNOME
and
KDE
more than ever. I’m particularly inspired by KDE’s growing base of “supporting members” and
successful fundraisers
. So while this post might seem negative, I don’t see this as a moment of failure, only a moment of inflection and of change.
This rant probably needs a deeper message so I’m going to paraphrase Eugene Healey: “Structural reform can’t be achieved just by publishing code online”. The hard work and meaningful work is not writing the code but building a community who support what you’re doing.
My feeling about the new AI-infested web, more to the point, is that it spoils the artistic aspect of publishing your new project right away as open source. There’s something completely artless about training an AI on other people’s ideas and regenerating it in infinite variations. Perhaps this is why
most AI companies all have logos that look like buttholes
.
Visual artists and animators have seen DALL-E and Stable Diffusion tale their work and regurgitate it, devoid of meaning. Most recently it was the legendary Studio Ghibli who
had their work shat on by Sam Altman
. “
I strongly feel that this is an insult to life itself”,
say the artists. At least Studio Ghibli is well-known enough to get some credit, unlike many artists whose work was coopted by OpenAI without permission.
Do you think the next generation of talented visual artists will publish their best work online, within reach of Microsoft/OpenAI’s crawlers?
And when the next Fabrice Bellard comes up with something revolutionary, like FFMPEG or QEMU were when they came out, will they decide to publish the source code for free?
Actually,
Fabrice Bellard himself
has done plenty of research around large language models, and you will notice that his recent projects do
not
come with source code…
With that in mind, I’m declaring bankruptcy on my collection of unfinished ideas and neat projects. My next blog post will be a dump of the things I never got time to implement and probably never will. Throw enough LLMs at the problem and we should have everything finished in no time. If you make the thing I want, and you’re not a complete bastard, then
I will happily pay a subscription fee
to use it.
I’m interested what you, one of the dozen readers of my blog, think about the future of “coding as art”. Is it still fun when there’s a machine learning from your code instead of a fellow programmer?
And if you don’t believe me that the world goes in cycles and not straight lines: take some time to go back to the origin story of Richard Stallman and the GPL itself. The story begins at the Massachusets Institute of Technology, in a computing lab that in the 1970s and 80s was at the cutting edge of research into… Artificial Intelligence.
chevron_right
.svg)
This year I have again received a grant from the WMF to attend to the annual
