- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
cross-posted from: https://infosec.pub/post/47200357
One critic called the move “petulance beyond measure.”
cross-posted from: https://infosec.pub/post/47200357
One critic called the move “petulance beyond measure.”
I don’t use mine for coding, but it can be useful for editing stuff, since a lot of agent-based systems can edit parts of a file instead.
The thing I have it do sometimes is parsing a bunch of markdown files, and parse data to put into the middle of a CSV, so it’s not out of order. Since making a script read the markdown is non-trivial, and itls not something that needs to be done very often, it’s easier to run a local model on the same machine and have it do that. Past a point, re-generating the entire file isn’t feasible, since it either consumes so many tokens doing the output that it hits the output limit that’s usually in place to prevent looping, or it takes an incredibly long time.
Is AI good for anything other than boilerplate code though? I worry when people talk about databases.
First thing is to separate out the term AI from LLM’s.
AI as a term encompasses many different technologies, some going back decades, a lot of which is used all over the place.
What we’re hearing a lot about right now are LLM’s and the surrounding ecosystem.
To answer the question though, yes, they can be used to produce output that fits a use case.
Whether or not it’s the best tool for the job is subjective, even in the cases where it’s technically viable.
There is a lot of bias and a lot of arguments for both sides.
You’d probably be best served by reading around a bit and figuring out how you feel about it.
You’re unlikely to get an unbiased discussion from a single source, especially here.
I’m not excluding myself , I’m bias AF.
The technology is interesting, the industrial implementation is an environmental and societal catastrophe.
I can only speak to personal experience, which is admittedly limited
Anyone who says yes is clearly not knowledgeable. Its like asking a spaghetti developer if they think their code is good. AI Physcosis is unlocking a new incompetence fear in me.
It’s absolutely fine at some stuff, provided you know enough to spot any mistakes it might make.
Just because you can do it with an LLM, doesn’t mean it’s the best tool for the job.
I have some experience with vibecoding, and while it does help, it only cuts down on development time a little bit. I’m talking about markup language. I have a hard time imagining that someone can build an entire app with this.
It depends. At least, I find that it has a habit of falling on its metaphorical face if the task is anything more complex than the simplest things, so the idea that people can use it to make viable programs is baffling to me.
“Put these values into the CSV” works okay enough, but if you task it with more than that, like see if a column of values in the CSV is entered correctly from the markdown, it breaks.
Or it gets stuck in a loop, and there’s a very short point where it is faster to enter it by hand. Slightly ironic, though, that a language model doesn’t do too well with natural language processing.
I’d certainly not trust it for anything important like a production database, but the csv/markdown thing isn’t, and it’s no big deal if it gets destroyed by the model/agent, so it’s interesting to poke around with, and feel out the limitations, so you know its strengths and weaknesses.