detfalskested

Les Edwards' iconic artwork for the album "Music for the Jilted Generation" by The Prodigy.

Until further notice

So, I've decided to take my work back underground

To stop it falling into the wrong hands

The above quote is from the Intro track on The Prodigy's second album, Music for the Jilted Generation. A record which, on the inner sleeve, has the above iconic artwork by Les Edwards.

With the current "AI" hype train running on full steam, free and open source software has met a challenge that the licences we have been relying on for ages do not have a good answer to or defence against. It's very encouraging to see places like Codeberg taking a stance against "AI". But as long as the source code is still open it is unfortunately also exploitable for "AI" training, no matter how loud we say no, despite it obviously being abuse behavior by the "AI" companies.

I just finished going through my profiles and groups on GitLab and the data.coop Forgejo and made sure to make all repositories private that don't need to be public. It's a pity. I have been believing in the value of open source for decades (and still do). But I don't want my code to be used as training material for LLMs against my will.

This is of course something I should have done a long time ago. The damage has, to a large extend, already been done. But I will no longer let this happen. Until I, or someone smarter than me, come up with a solution, things will remain private.

Maybe the solution – at least temporarily – is to work together in private places, where the bots and crawlers can't scrape our free (not as in beer) work and feed it to the slop gigafactories of hypercapitalism. Because I do want to work together with others. Other people.

Server-Sent Events from behind nginx

I was having weird issues with the timing of delivery of Server-Sent Events in a production Django application. It was set up to burst out a few events at the same time. For some reason, the last one did not show up with the rest of the burst. Instead, it showed up as the first item in the next burst, which would again not include the last expected item.

I first suspected async to be the culprit, as I am only getting my feet wet in that aspect. But after a bit of debugging (and acquitting async), it turned out to be nginx doing some buffering of responses. Something I, of course, did not experience during local development, where nginx is not part of the setup.

Fortunately, nginx allow you to turn of proxy buffering. Either completely for your site, via proxy_buffering off. Or by adding a custom header to the specific response in the application behind nginx, X-Accel-Buffering: no.

A further explanation can be found in the nginx documentation.

Overskudsvarme fra datacentre

En artikel fra DR om overskudsvarme fra datacentre fik mig til at skrive en e-mail til Klima-, Energi- og Forsyningsministeriet.

Hej Samira Nawa og KEFM

Jeg læste lige DRs artikel om at kun ganske få af landets datacentre udnytter overskudsvarmen: https://www.dr.dk/nyheder/indland/datacentre-mangler-udnytte-overskud-af-varme

Der er et par citater fra Henrik Hansen fra Datacenterindustrien og ministeren selv jeg gerne lige vil kommentere på.

Henrik Hansen siger:

Overskudsvarmen har været pålagt en afgift og et prisloft, der har gjort det til en dårlig forretningsmodel

Det er måske korrekt, hvis man ser på overskudsvarmen som en selvstændig forretning. Jeg synes det er mere korrekt at påpege at det er hans datacentre som helhed der har en dårlig forretningsmodel, hvis ikke de har råd til at skille sig af med overskudsvarmen på forsvarlig vis.

Ministeren siger:

Det må ikke ende med at blive til større varmeregning for dem, der har datacentre i kommunen

Det er jeg som sådan helt enig i. Men det er også forfejlet at ville lægge regningen over på det omkringliggende samfund. Overskudsvarmen er et problem der skal håndteres for datacenterejerens regning, efter princippet om at forureneren betaler.

De skal ikke have lov at varme kloden op. Og slet ikke på vores andres regning.

Afslutningsvist vil jeg gerne påpege at det blot er endnu et brudt løfte omkring lyksalighederne ved datacentre. Hvad er der fx blevet af alle de tusindvis af jobs de har lovet hver eneste datacenter vil skabe? Eksempel: https://www.dr.dk/nyheder/viden/tech/apples-datacenter-giver-80-permanente-job-i-viborg

--

Mvh Mikkel

The Messy Middle

A long quote from the essay Why Is Everyone In Tech So Sad? by Aaron Horwath:

Employees overwhelmingly choose to remain or leave their roles because of their colleagues and/or bosses, and the quality of the experience of the messy middle a workplace offers. Those are the elements that make work fun—and valuable.

But not everyone feels that way. In my experience, there are two broad categories of Knowledge Workers. The first are outcome-first workers. Their focus is on the business, on winning, on efficiency. Human needs and faults and emotions are an obstacle to be overcome.

The second group has an experience-first perspective. These workers love the messy middle. They value the journey. They want their organizations to perform well but as a natural consequence of collaboration, debate and shared struggle with people they actually like. Many experience-first people are artists outside of work. Photographers, directors, screenwriters, sculptors, painters, poets, writers. Many are active volunteers. For these people, to be pulled away from their economically unviable passions requires something in return: a company experience with freedom for exploration, creativity, problem solving and cool colleagues.

Research suggests this group needs that environment to do their best work. Harvard Business School’s Teresa Amabile spent decades studying what produces genuinely creative, high-quality output. Her Intrinsic Motivation Principle determines that people do their most creative and innovative work when motivated by the work itself — the interest, the challenge, the enjoyment — not by outcomes or metrics. The environments that kill creativity are political, risk-averse and relentlessly outcome-focused. The environments that stimulate it are collaborative, idea-driven and free. The messy middle, in other words, isn’t inefficient. It’s the condition under which genuinely valuable work gets produced.

It’s important to note that the impact of removing the messy middle from the work experience of these two groups is asymmetrical: For outcome-first people, it is a victory. For experience-first people, it undermines the foundation of work.

I'm definitely in the group of people who thrives in the messy middle.

Another interesting quote from the same essay:

We’ve been through waves of hiring and firing a thousand times before. But this time might be different. In the past, there was always a fresh cohort of workers waiting to be brought in. But what if talent stops coming back? What if the AI transformation gone wrong makes top talent — particularly those who are experience-first by nature — lose faith in Workism en masse? What if, when companies go to replenish headcounts, talent freed from the Workism spectacle has turned to different versions of their lives? What if their side projects suddenly became profitable? The talent pool increasingly doesn’t own homes and doesn’t plan on having kids, so it has perhaps never been easier to make a career pivot toward something that is genuinely satisfying in ways corporate life isn’t.

We're building on stolen ground

A technologist who builds the Machine anyway is just building it on stolen ground.

Back in April, I attended the Cables of Resistance conference. One of the workshops that provoked the most thoughts was hosted by Tara Tarakiyee, called "Can We Build Hardware Without Capitalism?"

The introductory quote here is from a blog post they wrote about the topics covered in the workshop. It's definitely worth a read. The short story The Machine Stops mentioned in the blog post can be found online at Standard Ebooks.