AI and keeping engineering skills

AI, IT

Today I was asked two common questions:

  • If developers rely more and more on AI, how will they keep their skills up to date?
  • And how will new engineers learn programming if AI writes the code?

Here is how I see .

1. AI will write most code

AI is already moving toward generating most of the code.
Humans will increasingly focus on:

  • problem definition
  • system design
  • requirements analysis
  • validating results
  • correcting and improving outputs
  • reviewing and maintaining systems

For routine software — websites, forms, apps, chatbots, standard backend services — AI can already generate most of the code. This category probably represents the vast majority of software written worldwide.

2. AI does not reduce thinking in professional work

It is true that many technologies, over time, reduce the need for certain types of thinking:

In everyday life, this trend is real.

But in professional AI usage, I observe the opposite.

Working intensively with AI means spending more time on:

  • analysis
  • structuring problems
  • writing clear specifications
  • validating outputs
  • managing context
  • coordinating multiple agents

You spend less time typing code and more time thinking about the system.
In practice, thinking often increases rather than decreases.

3. Reading code may also become easier

A real concern is that if developers stop writing code manually, they may lose the ability to read and validate it.

My expectation is that this problem will gradually disappear as well.

If AI generates code, languages can evolve toward being simpler to read and validate, rather than being optimized for human authorship. New languages may appear that are designed primarily for:

  • readability
  • validation
  • verification of AI outputs

In that world, deep expertise in writing code in a specific language may become much less important.

4. New engineers will learn differently

Future engineers will likely start their careers differently from our generation.

Instead of learning everything end-to-end — from system design down to writing every line of code — they will focus much more on:

  • understanding systems
  • architecture and design
  • task decomposition
  • specification writing
  • validation and verification
  • working with automated systems and agents

In other words, they will learn to be engineers first, not manual code writers.

Our generation of developers — who spent years writing code line by line — may simply be a transitional stage.

5. The industry will adapt

I don’t expect a long-term skill crisis.

What will change instead:

  • new tools
  • new languages
  • new development patterns
  • new ways of working

The industry will evolve, and engineers will evolve with it.

In a sense, AI may finally push software development toward something closer to true engineering.

Добавить комментарий