@scottmichaud But does it? It's not a display of skill and I don't feel like they get the kind of recognition that "bragging" would point at. Write about how you use "AI" you get a lot of shit (for very good reasons)
Discussion
Loading...
@tante The code generated by LLMs is generally a jumbled mess without rhyme or reason, so it's quite impossible to "check it". You can only run it and report all the errors you are getting back to the LLM, which will say it's sorry and generate another jumbled mess.
@tante It's not a tool. When wielding a tool I'm in control of the output. It's an entity you can delegate tasks to, giving you the impression of power after said task has been seemingly executed successfully.
"Look! I've made the machine do a thing for me! I'm the Machine Whisperer!"
@tante
It's a cult.
Again.
@tante The description sounds like they have to pretend that they *could* check the generated code, if they wanted to. Gives them an excuse when things go wrong without ever actually having to do it.
@tante I just take it as bragging.
@scottmichaud But does it? It's not a display of skill and I don't feel like they get the kind of recognition that "bragging" would point at. Write about how you use "AI" you get a lot of shit (for very good reasons)
@tante Personally, my brain goes to "They're not confessing. They're bragging."
It's dumb but... it's dumb.
It is specifically weird because - outside of a few specific circles - "I just had an LLM generate some code I didn't check" does not give you praise but ridicule or even attacks. So it must be very relevant to say that instead of just using the magic machine and not saying anything.
So either is is a form of prayer or ritual, a spiritual thing to enforce one's own belief? Or is it a shibboleth to mark your affiliation: Kind of like mafia members used to commit crimes in front of one another to show one's authentic affiliation?
@tante Just today I was experimenting with claude, using it to build a simple web-based game. At one point, when it implemented tests for a new set of messages, it deleted the tests for all of the rest of the message handlers. If I hadn’t been inspecting the code as it went along, I would not have noticed.
I haven’t written any code so far, but I have been inspecting the code and then providing a set of “best practices” that guide the agent to produce the code in the way that I want it. Today I added the idea that message tests should be write-once, and only modified when explicitly asked.
I suspect that people who are vibe-coding without even looking at the code can have some success, but that it leaves a lot of holes in the code that will eventually make things break in ways that they won’t be able to easily (or cheaply?) fix.
