Coding and AI: A Testimony Reignites the Debate on the Profession

Key Takeaways

1An engineer writing under a pseudonym describes a daily life where AI generates specifications, code, and tests

2His post, viewed nearly 5 million times, highlights the pressure on teams and the loss of autonomy

3Personalities like Elon Musk, Chamath Palihapitiya, and Bill Staples react publicly

4AI is becoming prominent in coding with tools from Anthropic, OpenAI, and Google

Why it matters : This testimony illustrates the tensions between productivity gains and the loss of meaning in the profession as AI transforms software development.

An engineer, writing under a pseudonym, describes an environment where AI tools produce specifications, code, and tests, leaving humans with mere validation. His account, viewed nearly 5 million times, adds to warnings about the identity of the profession and receives immediate reactions from tech personalities. At the heart of the debate is the obsession with delivered volumes versus technical and user understanding.

Engineers Fear Loss of Meaning and Autonomy

Deedy Das, a partner at Menlo Ventures, speaks of an "identity crisis bordering on depression" among some engineers faced with an increasing flow of AI-generated code. Similar concerns arise elsewhere, such as with Liu Shengyu, an AI developer at DeepSeek, who believes that AI could soon surpass him in the specialized work he values, relegating him to managing agents rather than writing code himself. For v0xium, considering AI as a "magic wand" without engineers capable of understanding the code, architecture, and users, nor allowing them to solve difficult problems, is a dead end. He asserts that removing these dimensions "leaves nothing," even to the point of undermining the sense of utility in the profession. He adds that the bottleneck has never been code writing, but the systems that surround it.

In Daily Life, AI Generates End-to-End and Teams Sprint

In his undisclosed company, v0xium describes a position that has become exhausting as Claude Code produces specifications, tests, tickets, and reports. He claims that people work 12 to 13 hours a day launching automatic generations, under pressure to ship as much as possible, with little time to review or understand what is happening in the background. He writes: "No one thinks anymore." The message he posted on Sunday had reached nearly 5 million views by Monday morning. He blames a managerial culture that prioritizes delivered quantity: sprint cycles, pull requests, and the number of features are seen as the main indicators, even when additions do not improve the products. In this context, the pressure mounts to entrust AI with both specification documents and system designs as well as user interfaces and backends, at the risk of reducing developers to "human proxies" who relay outputs without analysis.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Coding, the First Field of AI Adoption at Work

Software development has become one of the first areas where AI is making its mark daily, with tools offered by Anthropic, OpenAI, Google, and others to generate code, debug, and manage complex tasks. Some professionals believe these assistants automate tedious parts and free up time for building. Others fear being relegated to supervising, correcting, and validating machine outputs.

Tech Figures React Publicly

The message sparked immediate reactions. Elon Musk simply replied, "Yikes." Chamath Palihapitiya warns that companies risk reducing a generation of experts to pressing a button in front of slot machines, while reminding that the value of automated software lies in understanding intent and improving outcomes, not just speed. Bill Staples, CEO of GitLab, summarizes the situation by stating that humans become the orchestration layer for machines.