An opinion piece argues that the claim 'coding was never the bottleneck' is actually bearish for employment, as AI acceleration of coding will lead companies to optimize organizational bloat and coordination overhead, resulting in leaner teams and potential job losses.
It seems like with the acceleration of software coding through AI, many programmers claim that while coding itself has become faster, the overall productivity gain has not been as dramatic because, in their companies, coding was never the main bottleneck. Instead, they point to other factors such as meetings, coordination with other teams, bureaucracy, organizational friction, etc. However, I remember that even in the pre-LLM days, a lot of developers treated these “other” parts of their jobs as inefficient bullshit that often got in the way of real progress. Of course, some coordination is genuinely necessary, especially in large systems, regulated industries, or products with many stakeholders. But a lot of it also seems to come from organizational bloat: too many teams, too many handoffs, too many layers of management, and too much process. So if you take the 'coding was never the bottleneck' argument to its logical conclusion, it does not necessarily make the employment outlook better. In fact, it may make it worse. If AI accelerates coding, but productivity is still limited by coordination and bureaucracy, then the next target for optimization is not coding itself but the organizational structure around coding. This creates a path toward much leaner teams. Newer companies can be built from the ground up with fewer people, fewer layers, fewer meetings, and more AI-assisted execution. They can learn from the inefficient work processes of older, bloated companies and potentially outcompete them with smaller teams that move faster. And if that happens, older companies will eventually have to respond. To remain competitive, they may need to reduce coordination overhead, flatten management structures, automate more internal processes, and eliminate jobs that mainly exist because the organization is large and inefficient. So it seems like this argument that “coding was never the bottleneck” is brought up a lot when saying that AI doesn't help that much and to somehow help the argument for jobs to the developers when it seems like overall, the conclusion is actually more bearish and reveals dents in the bloated companies that can be rooted out from ground-up. Thoughts?
David Sacks argues that AI's automation of coding has led to a surge in code creation and increased demand for software engineers, challenging the narrative that AI will cause mass job loss.
Developers increasingly refuse to work without AI coding tools, but studies and reports suggest this reliance may not boost productivity and could increase maintenance costs, raising concerns about long-term impact.
The article critiques the shift from outcome-based productivity claims (e.g., 55% faster task completion) to volume-based claims (e.g., 75% of code AI-generated) by AI coding tool vendors, arguing the latter are less meaningful and harder to falsify.
The article discusses how AI coding tools may lead to a paradox where they require expertise to use effectively but prevent the development of that expertise in novice developers, potentially causing a collapse in coding proficiency.
Cal Newport reflects on the initial enthusiasm for AI coding tools like Claude Code and the growing disillusionment among developers due to hidden bugs, quality issues, and unsustainable workflows, arguing that outsourcing all code production to AI is not currently viable.