[NatSydenham]

I find it's great for very generic work, like boilerplate or where you have forgotten how to implement a specific single, well defined algorithm. However, relying on its output is very risky if you aren't verifying everything it puts out. A good test to show this is to ask it to write a short essay on something you know very well, then behold all the incorrect information it enthusiastically tries to feed you.

Also:

* It has a horrible habit of inventing properties on objects or methods in libraries.

* It will very happily straight up lie to you about things it does

* Often when you ask it to make a specific change, it will give you back exactly the same as last time.

* For the love of God, don't put any company-owned code into it.

* Maybe I'm just bad at prompting...

[DrSiemer]

Making an LLM explain parts of new code I don't understand has taught me a lot of useful new tricks and methods.

It's own code output will generally work for things that have been done before, but quickly fall apart when you are trying to create something new. You'll need to write a lot of pseudo code to make that work, which is actually excellent training as well.

It really does help if you instruct the language model to be honest and to not do it's very best to please you.

I have not found a solution to the repeating incorrect answer loop yet, although some environments offer a "banned response" list, that can mitigate the situation a bit.

As to the OPs question, yes, definitely use it! But more importantly, learn how to use it. Your approach sounds reasonable and similar to mine. Companies that ban the use of LLMs or developers that scoff at it's benefits are delusional.

Something that is easier to use might come eventually, but in the short term future I expect we will see a clear separation between developers with strong grammatical skills who have learned how to prompt and those without the ability or the experience.