I wonder what the issue is? Is it that is can provide wrong information?
Is it that you don’t see any legitimate use cases for them? As I work as a software developer and CoPilot in Visual Studio has saved us countless hours in writing boiler plate stuff and just solving problems in general.
Furthermore, the owner has created some pretty amazing tooling utilising LLMs and again, it saves us countless hours in the boring things. Like now if I make a new model in C#. We have something watching the code and that will use LLMs and some custom methods to scaffold changes to about 10-15 files and also create UI components in the client ready for us. All we need to do is update the schema doc for graphQL, run the migrations and update the DB. Then we are good to go. Saves about an hour every time a model is added or changed.
Yes, everyone with some common sense hate LLM generated texts.
Care to elaborate?
I wonder what the issue is? Is it that is can provide wrong information?
Is it that you don’t see any legitimate use cases for them? As I work as a software developer and CoPilot in Visual Studio has saved us countless hours in writing boiler plate stuff and just solving problems in general.
Furthermore, the owner has created some pretty amazing tooling utilising LLMs and again, it saves us countless hours in the boring things. Like now if I make a new model in C#. We have something watching the code and that will use LLMs and some custom methods to scaffold changes to about 10-15 files and also create UI components in the client ready for us. All we need to do is update the schema doc for graphQL, run the migrations and update the DB. Then we are good to go. Saves about an hour every time a model is added or changed.