> How confident can you be predicting 30 years in the future? Note that 30 years ago was 1991.

30 years ago, in the '90s, the company I now work for developed an important system for a very large customer. It replaced an older system that was developed along with the equipment it runs, introduced in the '60s -- another 30 years prior. My current project is developing a replacement for the current system. It's the exact same system, but written in C to run on a Cortex-M instead of Ada on an Intel.

In another 30 years, we'll probably be re-implementing the exact same system, but written in Rust to run on a RISC-V chip.

You can still make a serious dollar writing COBOL code. That was quite oldfashioned in 1991, and I'm honestly not certain we'll be rid of it by 2041.