Another *real example* that happened to me: Person: Hey I found this issue, also here are my notes <<insert giant pile of notes>> Me: Hmm, did you write this with AI? P: No, it's entirely by hand Me: But... all of the links don't exist? P: Oh I synthesized the notes with AI Me: ... - ThreadSky

hazelweakly.me • 6 days ago

Another *real example* that happened to me:

Person: Hey I found this issue, also here are my notes <>

Me: Hmm, did you write this with AI?

P: No, it's entirely by hand

Me: But... all of the links don't exist?

P: Oh I synthesized the notes with AI

Me: ...

Comments

hazelweakly.me•6 days ago

I am begging you, if you can't even *read* the output of an LLM, try it out, or otherwise judge its quality at ALL before handing the information out... Just tell me it's from an LLM.

I promise I won't judge, just don't waste my time. Please

bram.xyz•5 days ago

Girl, you need better friends. Who want their word to be worth something.

Those people exist but it takes time to surround yourself with them.

hazelweakly.me•5 days ago

My friends are fine, they're just fatigued by the state of the world and they're trying their best (they're also, by volume, not the problem at all).

It's the more professional contexts where people are paid to be right that's harder to deal with. It's ironically the more frequent case, too.

armurray.com•5 days ago

This is why I argue LLMs drastically increase the cognitive load / critical thinking component. We now need to guard the inputs and outputs from these things. Instead, we’re turning into a society of Brondo has what plants crave

maxdeviant.com•5 days ago

I've been using LLMs pretty heavily in a recent project of mine involving graphics programming.

It's been quite useful in helping me get a grasp on things, but I've found myself wanting better signage for which parts of this were written by the LLM.

maxdeviant.com•5 days ago

I even considered having separate modules that contained the LLM-generated bits so that I can look back later and easily see "oh yea, the LLM wrote this and it *seems* to work okay, but hasn't been heavily quality-controlled".

hazelweakly.me•5 days ago

That totally makes sense. In general, I really want better ancillary information for programming languages in general. We live in a world in which metadata has an outsized effect on everything and the fact that nothing out there seems to do a good job at tangling metadata to code is disappointing.

hazelweakly.me•5 days ago

Ironically, a huge multiplier of AI would've been if we had spent the last 15 years on: a better focus on documentation, literate programming, "the human soft parts" of code, correlating ADRs to code, tying proofs to code, bi-directional linking of to code... etc.

But nooo, its "useless"

maxdeviant.com•5 days ago

This is something that we're looking to explore more at Zed: how can we build a representation of the code that includes other artifacts (docs, conversations, general metadata) along with it.

Not something we've explored super deeply yet, still focused on building the actual editor bits 😄

Comments

Posting Rules

Reply