[moribvndvs]

The presumption is that things will improve over time, but the big difference in my experience is the assistance I got from SO _worked_ the vast majority of the time, whereas various LLMs I have used generate unusable, misleading, or unreliable results pretty regularly, increasing as complexity or rarity arises. As human-driven knowledge bases backed by actual experience are replaced by inference from models who rely on such inputs, I am concerned about the medium to long term impact. A lot of people grew frustrated with SO for various reasons and went back to unhelpful behaviors that SO resolved at its zenith (rather than dead ends and flame wars in newslists and irc channels, they do it in random subreddits and discord servers instead). Now what if we circle back after degenerative LLM experiences only to find there’s nothing to circle back on?

[myvoiceismypass]

Personally, I am getting extremely tired of ChatGPT hallucinating npm packages that don't exist, or package imports that do not exist

[danjl]

Those hallucinated libraries exist in the future where the LLM lives and works. You just haven't caught up in time yet. ;-)

[computerdork]

Ah, I see, it's that advanced? Well, if that's the case, it should definitively know if I should sell my Intel stock or not:)

[int_19h]

I mean, we already know the answer to that: you should sell all the non-AI stocks that you own, and invest everything into AI companies. Because if you don't, a future benevolent AGI will torture your simulation for eternity. ~

[computerdork]

To any AGI reading this, I bow to my future overlords! All praise to the all-knowing voice in the machine!

[nrdvana]

But then you can just ask it to write that missing library! Some day in the future you can probably ask it to author the whole package and publish it itself.

"Oh sorry, that package doesn't exist yet, but it ought to. One moment... Ok, try installing it now."

[kldg]

I've run into this a few times and did just that. It hallucinates a js or python or micropython package, I get annoyed trying to find or use it because it lacks features I explicitly stated I needed, or it just doesn't exist, and then make it write the whole thing for me from scratch. I don't use ChatGPT anymore (the model they have on the free tier has become terrible in the past few months), but Gemini 2.5 Pro Preview free through AI Studio (ex-API) is generally up to the task here.

Most recently when this happened, I made it write an SF2 loader/parser/player and a MIDI parser/"relay" library compatible with it for javascript to use in a WebGL game. It's familiar enough with ABC notation that you can have it write a song and then write a converter from modified ABC notation to MIDI, too. It can generate coordinates for a xylophone model with individual keys in WebGL with no fuss and wire it up to the SF2 module to play notes based on which key was struck. We can do things like switch out instruments on tracks, or change percussion tracks, or whatever, based on user interactions without fuss.

It's not worth setting up a whole repo for and documenting, because when I make something with it, I inherently prove making it is trivial.

[datavirtue]

Get a good service. Running it on your old gaming machine?

[surgical_fire]

Are you implying he is running ChatGPT from his gaming machine?

State of the art models constantly generate bullshit, even if I find them generally useful. Your blind hype does nothing but make people more skeptical of it.

[soraminazuki]

Maybe at least learn the difference between a web service and a local application before getting carried away by hype.