Nobody knows what programming will look like in two years

[ktpsns]

Monthly AI subscription fees will most likely quadruple in two years. This will fan out the greyscale of AI assistant quality much more.

Who can afford to use a $1k/month agent for a side project?

This post summarizes the economics of AI pretty well, describing the lock-in effect we will observe: https://merveilles.town/@lrhodes/116086521256156934

[deterministic]

I am 99.9999% sure my programming job will look pretty much the same two years from now.

I use AI daily, and it definitely helps me being (say) 10% more productive. However there is zero way an AI can do my job. And it looks as if LLM's have hit the limit of what is possible with the current AI architectures. So unless there is a revolutionary new architecture happening the next two years (unlikely) I don't see much change.

Also, using customer code generators makes me way more productive than using AI's. I simply declare what I want at a high level and the code generator in seconds spits out C++/Typescript/SQL/XML/JSON/CMake/Tests/... to do it. About 90% to 95% of the code I need is generated doing this. Way more efficient than using an AI.

[notnullorvoid]

The act of programming will look very similar. The community of programmers will be smaller, as more and more former programmers decide they don't like programming.

[v3xro]

Really? I will most likely be using IntelliJ 2027.x with whatever the latest plugins are for the programming languages I write by hand.

[p0w3n3d]

Thanks. You made me smile... Yeah I think this really depends on the companies management decisions. Currently I observe some sort of gold rush, buying AI and expecting more and more. I think that in two years it will be sorted out, people will realise how much did we really benefit from using AI, and what died in that time

[scrubs]

I do. I'll be working on low level systems software not listening to this drivel.