Positive users highlight AI's rapid code-quality gains and future potential in the debate shift, while negative users warn that skipping reviews still risks errors, legacy issues, and technical debt.
Based on 16 visible X reactions from 40 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Judging by my most recent grok experiences you most certainly need to be looking at the code. LLMs were trained on a shitheap of open source code. It’s not even the wild west. They do not *understand* anything. They just guess. And it’s a mess. But sometimes it *might* produce the desired outcome.
If you read the code, you will see the even the best models make "dumb decisions" quite often. Typically, they make it more complex than it needs to be. But not all code is the same. Leaf node code, that you can slice off and wont affect rest of the system, is fine to not care as much.
@Austen funny how fast the conversation changed six months ago people were counting syntax errors, now they're arguing about responsibility and trust that's real progress
@Austen at this point AI is writing better code than 98% of engineers
If you read the code, you will see the even the best models make "dumb decisions" quite often. Typically, they make it more complex than it needs to be. But not all code is the same. Leaf node code, that you can slice off and wont affect rest of the system, is fine to not care as much.
@Austen funny how fast the conversation changed six months ago people were counting syntax errors, now they're arguing about responsibility and trust that's real progress
@Austen if you stop reading the code, you're just speedrunning technical debt.
In less than one year the AI for engineering debate went from, “Can AI write good code?” to, “Do you still need to look at the code that AI has written?
@Austen If you were still debating “Ai writes good code” last year, or if you still debating either of these this year… that puts you easily 2 years behind
@NickADobos Eh I think should you look at code right now depends on use case
Positive users highlight AI's rapid code-quality gains and future potential in the debate shift, while negative users warn that skipping reviews still risks errors, legacy issues, and technical debt.
Based on 16 visible X reactions from 40 accounts; directional sample.
Ask a question below.
Published answers will appear here.
In less than one year the AI for engineering debate went from, “Can AI write good code?” to, “Do you still need to look at the code that AI has written?
@Austen If you were still debating “Ai writes good code” last year, or if you still debating either of these this year… that puts you easily 2 years behind