Meh not sure it’s even pro or contra. It’s just happening and it’s gonna get better at what it’s doing. I find the whole 10x debate a bit ridiculous because 10x of what exactly? How are you even measuring the 1x in order to determine what would be 10x of that?
That being said, I’ve seen some crazy things, like an entire backend + frontend + dashboards + data warehouse with lots of pipelines + devops being run by a single dev at scale at one of our clients. With AWS or Cloudflare and enough AI budget such things are actually possible. But then you’re looking at a bill of $3-5k / month spent on AI.
Personally I like it for doing stupid repetitive refactoring or feature implementations. Say, I wanna add a new API endpoint so I tell it implement this new endpoint and stand up a service layer based on what all existing endpoints are doing, or I wanna migrate from one testing framework to another. It’s mostly gonna 1 shot it and run all the checks automatically until they pass. Or if I need to dive into the database I can let an agent sift through all the tables and get me a report while I go and make a coffee or work on something else.
I think it can be really useful if you know what you’re doing and it can make your life much easier if you learn where it can help you and where not, and that’s it. It’s just a tool that may or may not help you being more productive.
I tried putting together a research plan using an LLM. Like nothing crazy I just wanted it to help me structure my thoughts and write LaTeX for me. Horrible experience.
I gave it a reference paper and said "copy that methodology exactly“ and then said exactly what steps I would like to see included.
It kept making bold claims and suggesting irrelevant methods and just plain wrong approaches. If I had no idea about the topic I might have believed it because that thing is so confident but especially if you know what you’re doing they’re bullshit machines.