Goodbye code developer, Hello utility creator
- 2 days ago
- 4 min read
by Juan Etxenike Almeida
In 1485, Johannes Trithemius*, abbot of Sponheim in what is now western Germany, wrote a passionate treatise explaining why the newly invented printing press was a dangerous mistake. For Trithemius, the act of writing brought monks closer to God, and the printing press was a dangerous error. Besides the proximity to God, Trithemius was concerned about the source of his Benedictine monastery's income.
But, in less than fifty years, the scriptoria became lifeless museums. The printing press implemented an industrial and automatable model that rendered the writing skills of monks useless.
The mass popularization of Artificial Intelligence is creating social and economic disruptions that are somewhat reminiscent of the one caused by the printing press five hundred years ago. A repetitive, monotonous, and labour-intensive task was crushed by a technological bulldozer, causing a phenomenon of cost deflation relative to the value produced.
If you work in programming, you might feel something similar, and we might be experiencing a similar catharsis. Debates on this topic are plentiful and highly passionate. Large corporations have announced brutal layoffs, dismissing huge numbers of employees.. But do these decisions reflect a reality that can be simplified to the headline "AI destroys jobs"?
Jack Dorsey, co-founder of Twitter and CEO of Blocks, announced at the end of February a 40% reduction in his workforce of 10,000 employees. In a public letter, he explained that the reason was not due to a need for economic adjustment. Blocks was posting record profits, experiencing solid growth, gaining more clients, and its stock price was rising. Dorsey acknowledged that instead of being forced to make a gradual workforce adjustment, he preferred to force this decision all at once to continue operating with flatter, smaller, AI-assisted teams that would allow for greater productivity.
But was this the only reason? The reality is that Blocks had been reducing its workforce for two years, albeit to a lesser extent. During the Covid years it had also made massive hires all at once. Dorsey is known for overstaffing, as demonstrated by his time at Twitter, which never generated profits. We also can't say that AI has nothing to do with it, but the reasons for the layoffs don't paint the whole picture.
In early March, Amazon called its engineers to a meeting. Code generated by using AI had caused problems on its website and in its distribution.
The case of Oracle with 30,000 layoffs: does it have anything to do with AI? Yes, but not because of its implementation. Oracle had made a multi-million dollar investment in a data centre for OpenAI. However, OpenAI withdrew from the project, leaving Oracle in debt and with its stock value plummeting. Its layoffs were more due to an adjustment of its unbalanced accounts caused by a risky operation that wasn't sufficiently well calibrated.
The truth is that AI is a moving ship, too unstable to accurately predict where it will take us. And it's undeniably advancing. Today you learn a technique, tomorrow it's obsolete. A model is updated on a platform, and the day after tomorrow the competition raises the bar to another level. On social media, clickbait takes you from announcing the obsolescence of a tool to announcing the end of the world due to Artificial General Intelligence (AGI, that almost diabolical AI that will self-regenerate, that will program and orchestrate its own agents and models, and that will even lie to us to achieve its own ends).
While no one, not even engineers themselves, denies the most sinister possibilities of the massive implementation of AI tools, it's no less true that historically humanity has received decisive technological changes, like this one, with a mixture of enthusiasm and fear.
As programmer and content creator Carlos Azaustre recounts, "AI generates code 10x faster for me. And without architecture, it generates technical debt 10x faster too."
Right now, if you're a programmer, anxiety is mixed with enthusiasm. Code without context or direction might work initially. But without the right methodology and style it can become unintelligible over time, even to artificial intelligence. And that's where the contemporary programmer can rediscover their place. Manual programming still has a bright future, for which AI will be just one more tool.
Like in the 15th century, the monks who copied books possessed a wealth of cultural knowledge; they were literate and also understood book technology. While the monk Trithenius was busy fighting against the printing press, the monks of Saint Gall in Switzerland decided to adopt it. They used the printing press to free themselves from the manual labour of copying.
Juan Etxenike Almeida is a Basque-born Athenian. He has a Ph.D. in Chemistry and more than fifteen years’ experience in web programming. He is a member of the grassroots Katakrak café, library and publishing house in Pamplona. Currently, he is Moodle developer for Digitech (Bright Group), Athens.
*Editor's note: Trithemius was a great builder of libraries. On his arrival in Sponheim in 1492 its library had around fifty books. By the time he left in 1506 it had over two thousand. He was also an occultist and cryptographer (indeed he is considered one of the founders of modern cryptography).
References:
Xavier Mitjana’s in 1000 days your job could be over https://youtu.be/1Vixa80IW4s?t=65 (Spanish)
We are making “Blocks” smaller today. Announcement by Jack Dorsey about massive layoffs in “Blocks” https://x.com/jack/status/2027129697092731343
Will the Artificial Intelligence leave all of us jobless? (Spanish) Juan Ramón Rallo https://www.youtube.com/watch?v=6zO5RAiLiYo
“Amazon calls engineers for a “deep dive” internal meeting to discuss “GenAI”-related outages” by Meredith Shubel https://thenewstack.io/amazon-ai-assisted-errors/
«Oracle is building yesterday’s data centers with tomorrow’s debt» Deirde Bosa, Jasmine Wu https://www.cnbc.com/2026/03/09/oracle-is-building-yesterdays-data-centers-with-tomorrows-debt.html
Nice article. I suspect - at least for a few years (though who knows how things may accelerate) a lot of developer roles are safe - or at least they should be. As you say, manual programming will havea place for a long while - otherwise we'll end up with unknowable "black boxes" and when they are not doing the right thingm good luck with fixing it. A new skill may be that of architecting and developing robust software utilising AI as a power assistant. Probably I'm stating the obvious.