Hacker Newsnew | past | comments | ask | show | jobs | submit | Florin_Andrei's commentslogin

If you use AI, the paper trail you leave behind may now operate at a much higher level. Here's why. Also, here are some mitigation techniques.

I think we're still at the stage where model performance largely depends on:

- how many data sources it has access to

- the quality of your prompts

So, if prompting quality decreases, so does model performance.


Sure, but the study is saying something slightly different, it's not that people write bad prompts for artifacts, they actually write better ones (more specific, more examples, clearer goals,...). They just stop evaluating the result. So the input quality goes up but the quality control goes down.

Seems like it’s impossible for output to be good if the prompt is bad. Unless the AI is ignoring the literal instructions and just guessing “what you really want” which would be bad in a different way.

> On two occasions I have been asked, — "Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" In one case a member of the Upper, and in the other a member of the Lower, House put this question. I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.

- Charles Babbage, https://archive.org/details/passagesfromlife03char/page/67/m...

EDIT: This is a new iteration of an old problem. Even GIGO [1] arguably predates computers and describes a lot of systemic problems. It does seem a lot more difficult to distinguish between a "garbage" or "good" prompt though. Perhaps this problem is just going to keep getting harder.

1. https://en.wikipedia.org/wiki/Garbage_in,_garbage_out


What does prompting quality even mean, empirically? I feel like the LLM providers could/should provide prompt scoring as some kind of metric and provide hints to users on ways they can improve (possibly including ways the LLM is specifically trained to act for a given prompt).

That would be a quality metric, and right now they are focused on quantity metrics.

Some cleanup needs to happen, when the dust settles.

It's just not clear to me who, or what, will do it.


I would very much prefer that he focuses on SpaceX and Tesla, in this order, seeing as he's more or less custom-built for those jobs, and there aren't that many people like that around.

Running Twitter OTOH? There's an endless supply of "captains of industry" in the Silicon Valley, who would be willing and able to do it.


It's near a place where that amount is just a drop in a bucket.


> Those data centers come in different forms.

It's like the birds and the bees.


Crossing the street is also potentially deadly.

Not crossing it, staying home all day until you starve or go insane from isolation is more deadly.

So we choose to cross the street.


> because the other side is too stupid to make their own decisions

The unfortunate reality is - that is the actual crux of the problem. Maybe say "incapable of making" instead of "too stupid to make", since that makes the scope wider. But that is the long and short of it.


Do you routinely analyze and challenge your own biases? Do you ever wonder if someone you disagree with might be right?

It's easy to call someone stupid while ignoring what they are saying. It's hard to charitably hear someone's argument. I'm certainly at fault there.

Either way I categorically reject the worldview where a "benign" paternalism protects us from ourselves.


> Either way I categorically reject the worldview where a "benign" paternalism protects us from ourselves.

That's nice. unfortunately the world as it is is categorically providing practical counterexamples to this.


What is generally known as ethics acts as a safeguard against random individuals running rampant and inflicting tremendous damage on society for personal gain.

In a broad context, Thiel acts like someone who has no such safeguards. The future he's pushing the world towards is the caricatural dystopia from Back To The Future (which was originally meant to lampoon Trump-like characters). Fortunately, he will probably fail.

If you don't see this at all, it's probably time to ask yourself just how similar to Thiel you actually are.


what kind of future do you think Thiel is pushing the world towards?

a lot of his writing involving Girard and other political philosophers reads as studying mimetic desire as a means of avoiding absolute total war in society and maintaining the hegemony of what he believes to be "enlightenment values" — destruction and unnecessary suffering is generally the antithesis of those beliefs

again, I don't even personally have to agree with all his ideas and writings to want to have a discussion about it


"avoiding absolute total war in society" sounds like the rhetoric of the Holy Roman Empire ..


Dark Enlightenment thinking isn't far off that Holy Roman Empire. It's essentially an embracing of monarchy and caste, with rule belonging to the Peter Thiels of the world, based on various ways they can demonstrate their ability to claim that throne.


I'd greatly appreciate an example of someone arguing that we should "[embrace] of monarchy and caste".


Tucker Carlson interviewed Curtis Yarvin recently and they did touch on monarchism. It's a fairly accessible conversation.


> Tucker Carlson

That's the moment when reality-based folks stop listening to what you're saying.


No argument. Just an ad hominem. What's new?


Very similar to my play style.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: