Can We Be Made Redundant by AI?

I finally read O’Reilly’s article “The End of Programming as We Know It.”

This is an article about how programmers will change with the emergence of IDEs utilizing generative AI and LLMs.

It mentions the possibility of engineers’ job content changing.

What was particularly interesting were these two points:

1. Differences by User
   Senior engineers use AI to magically produce code. Meanwhile, junior engineers blindly believe AI output.

2. Method of Forecasting Here, they predict the future of engineers in AI development based on past programming evolution.

Thinking about it now, when home PCs first appeared or when smartphones appeared, there were opinions like “What can you do with this toy?” or “Who would use such a thing?”

Even now, AI has become indispensable in engineering circles, but I get the impression that few people in general society use AI.

Many of my non-IT friends don’t know the types of LLMs (ChatGPT or Gemini).

So, won’t using increasingly convenient AI create work methods different from before?

Actually, the book also states (paraphrasing) “Stubborn people who don’t learn new knowledge will be left behind.”

When I experiment with AI myself or study tax systems, I sometimes feel that using AI might be cheaper than the cost of hiring people.

In that case, the key seems to be deciding what tasks you want to perform, delegating to AI, and developing the ability to refine that output.

For now, I’ll try paying for AI and running PDCA cycles.

Summary

I think AI evolution won’t stop.

Within this, rather than creating AI ourselves, it seems better to move toward developing applications that run on top of it.

Also, while speed increases, it seems there’s still distance before we can have completely different experiences.

I want to think about how we can create experiences not on the extension of what we’ve had so far.