When people claim ChatGPT can replace software engineers because it can sort of code is always puzzling to me. 20 years into this career, and writing code is only a minor part of my day-to-day activities.
A part I do enjoy doing, but still minor part nonetheless.
> What if you design an organization so that coding is the bottleneck? How transformative would LLMs be?
The funny thing is that although ChatGPT has been an amazing coding assistant tool, it can't meaningfully write code beyond boilerplate to me. I use it more as search engine on steroids, rather than actually relying on the code it writes.
A few times it even got into a hallucination loop going through different flavors of uncompilable code when I asked "how to do X using library Y", and refining the question was getting nowhere.
ChatGPT is an amazing tool that provides me value beyond the 20 bucks per month I pay. But still has far too many limitations. It's sort of funny to see people LARPing that it will kill knowledge jobs in the near term.
Design Patterns are said to compensate for features missing in a language (e.g. Visitor for multiple dispatch). ChatGPT is sounding similar - solving accidental complexity that "shouldn't" be there in the first place.
Stuff like
match network_foo_int {
TYPE_A => Foo::A(read_a()),
TYPE_B => Foo::B(read_b()),
...
I've also found it reasonably good at writing the inverse of the preceding function.Though I don't expect software dev job to go away. It'll just change in nature. Less typey typey, more managing the LLM.
There's also not orders of magnitude more existing data we can feed it anymore.
Not as a language model. Other leaps in technology would need to happen before it can code.
As you say, I find the act of writing code is hardly the bottleneck to my productivity. One of the best case scenarios, LLMs become the ultimate high level language. In that case, Software Engineer roles would switch from coding to more of a mix of old school Software Architect, PM and entrepreneur.
By the way, I find interesting to run the argument of "coding is not the bottleneck" in reverse. What if you design an organization so that coding is the bottleneck? How transformative would LLMs be?