My own experience is that it _might_ speed up juniors (and even that, I'm not convinced about). But juniors make up a tiny part of the overall throughput, I am easily 10x faster to do anything than they are. So a 10-20% productivity boost for juniors is pretty much negligible for the company.
For myself, everytime I try to use chatgpt it completely fails to be helpful, for the same reason that stackoverflow also fails: the problems I have as a senior are too specific or too complex. Every single time, querying an LLM ends up being a waste of time.
The boost comes from easy problems. Write a cli command without looking at the docu, convert from this json format to the other, look ath this method and do the same in this method. This can also be easily verified. Mid-fast tasks become instantly done.
Also when you wear multiple hats. Recently I had to setup an OpenResty including writing some fairly complex Lua scripts, both which I had never touched before, and it was much faster with the help of an llm.
This. I think LLMs have been most helpful to me when I'm doing something that's not really in my mainstream domain. Yesterday wanted to know the structure that SAP stores invoices in, GPT-4 had a great detailed answer and even made me sample data instantly. It would take me hours of research to do it myself.
There's still a lot of closed off knowledge that's hard to access even though it should be open. ChatGPT is great for that.
For myself, everytime I try to use chatgpt it completely fails to be helpful, for the same reason that stackoverflow also fails: the problems I have as a senior are too specific or too complex. Every single time, querying an LLM ends up being a waste of time.