The bottleneck in engineering is shifting.
It is less about writing software and more about processing engineering information.
Modern engineering requires understanding code, telemetry, cloud state, incidents, dependencies, security controls, customer requirements, compliance obligations, and prior decisions.
That is an information problem.
AI is excellent at generating implementation. But safe implementation depends on retrieving and synthesizing the right context.
The quality of AI-assisted engineering will depend heavily on the quality of engineering information architecture.
Engineering organizations will need better systems for:
Software engineering is not becoming less technical.
It is becoming more information-dense.
The organizations that manage engineering information best will make better decisions faster.