The reliability of the Software sector as a long term career option.

We have all heard about how Claude Opus 4.6 model now aces majority of frontend work and is making huge strides in backend. We have also heard of the current limited ability of it to handle complex tasks i.e architect or orchestrate stuff. We have been proven wrong again and again over a mere 3 years duration where gpt 3.5 was unable to do anything without very heavy guidance (we laughed saying the goalpost is xyz) to today where it gets 70%-80% of the job done(and now we're moving the goalpost). And it goes to say that given 10 years it might be able to bridge it to 95% or more, please don't quote me as I'm just a 3yoe noob. I understand the arguments about scaling laws, compute costs, and real-world implementation, but investor greed will ultimately bulldoze those hurdles. I want to hear from all the experienced folks, the industry leaders on what's the future of software engineers in India look like? Will we be phased out / become cheap as dirt in the upcoming decade? Is this a classic "Who Moved My Cheese?" phenomenon? Should we still strive to become the best software engineer in hopes of getting the best pay? How are you guys planning for future? What field is emerging now which is the best replacement for software? ML Engineer?