My employer pays for Claude, and my approach is to use it as a better Google search. It's often not better.
Just today it made three glaring mistakes in one session:
1. It read a file in the wrong directory, because that file had the same name as the file in the right directory. It apologized when I challenged it, promising me that it would remember to "read import statements" in the future.
2. It miscounted the number of times a function was called in my repo. It said 20, while my built-in IDE search accurately showed 17. Again, it apologized when I corrected it.
3. It referred to a variable by name that does not exist anywhere in my code. It apologized, and said it was referring to a variable used internally by one of the third-party packages installed in my repo.
So many apologies.
It's the little things like this that remind me on a regular basis just how little I can trust artificial "intelligence."
It's a small tangent but I am constantly taken aback by just how much the discourse here has slid from nerdom to dorkeyness. All the talk here used to be pendantic and technical and overly complex nerd speak, always stuck in the process itself, feeling above being a 'user', etc.
Now everything is just like above, dorkspeak. Where it reminds me so much more of kids arguing in the playground about "who would win in a fight Darth Vader or Batman"; or console-vs-PC debates. Everything is about the genuine complexity of navigating certain products, of being first and foremost a consumer of something and putting all your energy into comparing various things you are free to choose from.
Its not even like its less techincal, or more mean now, or anything like that. It's just very different and I know its been a while but it feels like it happened overnight.
>It's a small tangent but I am constantly taken aback
I'm amused by your surprise.
>Its not even like its less techincal
I'd wager that, actually, it is less technical. We constantly see otherwise very techinical people here poking their heads up and admitting that they haven't been writing any code for 6 months or more, and the ones that brag about it are seemingly unaware that they've been reduced to being a technical PM (the ones that contest this probably havent worked with capable technical PMs). No one wants to engage with vibe-coded Show HN entries because the poster may not grasp the implementation. No one appears to be doing (or sharing) anything super novel with all this coding superpower they have suddenly attained so there's nothing technically interesting to talk about. The last time I found an AI application submission interesting was the one where the person was trying to turn their pet's random keyboard typing into a language (if I remember correctly).
If the guidelines banned any AI comments that didn't preface their comment with the model and version of AI they used, their application domain, and the programming language being used I'm sure there would be less contentious dorky debate and more technical/practical discussion.
And since many here don't see the endgame with AI resulting in anything good for their career or society, like the endless remote work debates, they comment because from a strategic point of view they do not want to cede the narrative to the other side on such an important topic. understandably. hence more dorky debates.
> No one wants to engage with vibe-coded Show HN entries because the poster may not grasp the implementation.
> No one appears to be doing (or sharing) anything super novel with all this coding superpower they have suddenly attained so there's nothing technically interesting to talk about.
stack_framer · · focus · HN ↗
Just today it made three glaring mistakes in one session:
1. It read a file in the wrong directory, because that file had the same name as the file in the right directory. It apologized when I challenged it, promising me that it would remember to "read import statements" in the future.
2. It miscounted the number of times a function was called in my repo. It said 20, while my built-in IDE search accurately showed 17. Again, it apologized when I corrected it.
3. It referred to a variable by name that does not exist anywhere in my code. It apologized, and said it was referring to a variable used internally by one of the third-party packages installed in my repo.
So many apologies.
It's the little things like this that remind me on a regular basis just how little I can trust artificial "intelligence."
garg · · focus · HN ↗
beepbooptheory · · focus · HN ↗
Now everything is just like above, dorkspeak. Where it reminds me so much more of kids arguing in the playground about "who would win in a fight Darth Vader or Batman"; or console-vs-PC debates. Everything is about the genuine complexity of navigating certain products, of being first and foremost a consumer of something and putting all your energy into comparing various things you are free to choose from.
Its not even like its less techincal, or more mean now, or anything like that. It's just very different and I know its been a while but it feels like it happened overnight.
sillyfluke · · focus · HN ↗
I'm amused by your surprise.
>Its not even like its less techincal
I'd wager that, actually, it is less technical. We constantly see otherwise very techinical people here poking their heads up and admitting that they haven't been writing any code for 6 months or more, and the ones that brag about it are seemingly unaware that they've been reduced to being a technical PM (the ones that contest this probably havent worked with capable technical PMs). No one wants to engage with vibe-coded Show HN entries because the poster may not grasp the implementation. No one appears to be doing (or sharing) anything super novel with all this coding superpower they have suddenly attained so there's nothing technically interesting to talk about. The last time I found an AI application submission interesting was the one where the person was trying to turn their pet's random keyboard typing into a language (if I remember correctly).
If the guidelines banned any AI comments that didn't preface their comment with the model and version of AI they used, their application domain, and the programming language being used I'm sure there would be less contentious dorky debate and more technical/practical discussion.
And since many here don't see the endgame with AI resulting in anything good for their career or society, like the endless remote work debates, they comment because from a strategic point of view they do not want to cede the narrative to the other side on such an important topic. understandably. hence more dorky debates.
dodkiexjwidj · · focus · HN ↗
> No one appears to be doing (or sharing) anything super novel with all this coding superpower they have suddenly attained so there's nothing technically interesting to talk about.
You don’t see a relation there?