The ‘code was never hard’ argument hit a programmer nerve
Senko’s essay pushes back on a familiar AI-era slogan by separating the act of typing code from the surrounding work of understanding, choosing, and maintaining a system.
The essay argues that implementation is only one part of programming: requirements, constraints, integration, debugging, and responsibility do not disappear when generation gets faster.
In a 364-comment HN thread, the productive disagreement was not whether tools can write code, but which parts of software work remain hard after they do.
The slogan hides the work around the code
Senko’s essay objects to treating programming as a bottleneck made solely of syntax. Its central distinction is concrete: producing text that compiles is not the same task as deciding what a system should do, fitting it into an existing environment, and being accountable for its behavior later.
That is useful framing for teams evaluating code-generation tools. Faster drafts change the economics of implementation, but they do not automatically resolve ambiguous requirements or create an operational owner.
A thread about where difficulty moves
HN readers brought examples from maintenance, review, and debugging rather than treating the argument as a vote for or against AI. The thread’s underlying question is whether a tool removes work or merely relocates it into specification and verification.
That makes the practical metric less glamorous than generated lines: ask which decisions can now be made earlier, checked more reliably, and traced when a change goes wrong.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
Denmark is adding an oral defense to written schoolwork
A report on Danish high schools describes a policy response to generative AI that does not try to identify every machine-written sentence: students must be able to explain their own submitted work aloud.