Ask HN: How to prepare the continuation of my career as as software engineer?

You can continue to be a software engineer all the way to a normal-age retirement. (I'm in the US, though - I don't know about France.) But you can't be the same software engineer for the whole time.

Your code should be better designed and architected than it was 10 years ago. You shouldn't write the same bugs you did 10 years ago. You should document your code better than you did 10 years ago.

Every five years or so, you should ask yourself what you need to learn now for the next few years of your career - and then learn it. (Hat tip to my wife.) One of the coolest things about software is that you can often get paid for learning that - you can find a project at your current company that will grow you in the direction you want.

What you cannot do is continue to be a better paid version of a junior programmer. You need to be more valuable than the new people. (This is why you don't need to worry about the "growing pool of software engineers" - it's growing largely by adding new people with no experience. They can't replace you, because they don't have your experience.)

At the moment I have a good contract.

Ageism definitely becomes a factor. One thing that is important is not to contribute to it. For example, part of your post implies that as you get older you will struggle to keep up with relevant technologies.

The real concern for every software engineer and actually every job is the powerful AI that we have to anticipate within the next 0-10 years.

OpenAIs newest model, although amazing, still has some weird brittleness to it, so it's not going to literally replace an actual software engineer easily. But it is a significant step closer than the day before yesterday. Probably at least 15-20% better, or a lot more depending on how you measure it.

I keep seeing people reacting to the current state of the art AI and saying that they are not concerned. This is incredibly short-sighted. The history of technology is hundreds of major innovations breaking through performance wall after performance wall.

There are not only upgrades but also new paradigms on the way for increasing AI performance, in the areas of hardware, ML model architecture, and software. And the reality is that with this latest OpenAI model and the right feedback loop and constrained platform/domain, we can actually automate quite a lot of software engineering work.

As AI is integrated better with systems, we will definitely need fewer people doing what currently is considered software engineering. And again, don't think about what the current model can do. Look at the trajectory of history over the last 60+ years of computing. Engineers always break through barriers and consistently increase performance by orders of magnitude.

This is not just a concern for software engineering but for every job. But it can be a huge improvement if society is able to adapt and take advantage of it in an equitable way.

I think the best option is to start leveraging AI and robotics to build out goods and services. Individuals without capital or very little capital can take advantage of AI and robotics to become entrepreneurs.

You're barely half way through your "commoditized workforce" lifecycle. Start doing things for yourself at any time.

Say a sentence to yourself that you want to be true. "I have made a vision detection solution for classifying cats." "I have created infographic heat maps of political manuer." "I have created a web game." And do that on the side until it is true. If you liked it as much as you thought you would, angle for that. It will at least give you something peppy for the interviews.

Otherwise, keep turning the lathe and hope abstract economic stabilities continue to grease your wheels of life.