AI is the empowerment of "I can do X better then people who's entire careers are built around X" attitudes
-
AI is the empowerment of "I can do X better then people who's entire careers are built around X" attitudes
-
AI is the empowerment of "I can do X better then people who's entire careers are built around X" attitudes
@nullpo Or, hear me out, it's "I can do 80% of the boring work quickly and then focus on the fun and complicated parts".
I'm way better than the LLM at coding. It's just a thousand times faster at parts of it so why wouldn't I use it.
(I've had this argument before. At that time it was whether to be lazy and compile high level languages instead of writing everything in assembler)
-
@nullpo Or, hear me out, it's "I can do 80% of the boring work quickly and then focus on the fun and complicated parts".
I'm way better than the LLM at coding. It's just a thousand times faster at parts of it so why wouldn't I use it.
(I've had this argument before. At that time it was whether to be lazy and compile high level languages instead of writing everything in assembler)
An important difference is that abstract programming languages are deterministic, whereas LLMs are not.
So you either spend a lot of time proofreading code nobody can explain to you, with potentially gaslighting comments, or your software gets 0wned a year later or doesn't even run in the first place.
If that's an improvement to your previous workflow, so be it.
-
An important difference is that abstract programming languages are deterministic, whereas LLMs are not.
So you either spend a lot of time proofreading code nobody can explain to you, with potentially gaslighting comments, or your software gets 0wned a year later or doesn't even run in the first place.
If that's an improvement to your previous workflow, so be it.
-
-
No, you're trying to splain me - who just wrote what my experiences (actual ones, not gut feelings) that LLMs are quite useful for ~80% of the work - that the code is not possible to understand.
I'm one of the old greybeards in both software dev and cybersec. When I state that they're useful they ... are.
-
No, you're trying to splain me - who just wrote what my experiences (actual ones, not gut feelings) that LLMs are quite useful for ~80% of the work - that the code is not possible to understand.
I'm one of the old greybeards in both software dev and cybersec. When I state that they're useful they ... are.
-
-
R relay@relay.infosec.exchange shared this topic