Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Pakistani newspaper mistakenly prints AI prompt with the article

twitter.com

Post reply on HN
Print view View original

141–150 of 180 posts

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#141

Post by wg0 »

nashashmi wrote: post #18
wg0 wrote: post #9

Earlier quoted context omitted.

Thank you, yes that's accurate and I am not sure if article itself is accurate. Don't think so it would have no incorrect stats. By "AI prompt" I mean "prompted by AI" Edit: Note about prompt's nature.

It might be better to mention “Dawn newspaper” instead of “Pakistani newspaper”.

Nobody outside Pakistan knows Dawn even though it is the newspaper that was founded by Muhammad Ali Jinnah (considered founding father of the nation) and one of the largest and most prestigious as well.

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#143

Post by kwanbix »

This has happened more than once, in different scenarios.

I wonder why is it that GhatGPT (and the rest) don't put the actual response in a box that is separate from the "hello" and "goodbye" part.

I once wrote a message to my landlord, and I asked ChatGPT to help me make the message better (being that english is not my mother tongue) and I included the "goodbye" part by mistake.

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#145

Post by mschuster91 »

simonw wrote: post #115
shagie wrote: post #107

Earlier quoted context omitted.

https://www.npr.org/sections/money/2015/05/20/406484294/an-n... ... The rules for the race: Both contenders waited for Denny's, the diner company, to come out with an earnings report. Once that was released, the stopwatch started. Both wrote a short radio story and get graded on speed and style. https://www.wired.com/story/wordsmith-robot-journalist-downl... https://archive.ph/gSdmb And this has been going on for a w…

I like Quakebot as an example of how to do this kind of thing ethically and with integrity: https://www.latimes.com/people/quakebot > Quakebot is a software application developed by the Los Angeles Times to report the latest earthquakes as fast as possible. The computer program reviews earthquake notices from the U.S. Geological Survey and, if they meet certain criteria, automatically generates a draft article. The n…

> The computer program reviews earthquake notices from the U.S. Geological Survey

Probably a service that is provided to the general public for free, similar to NOAA and weather data - so chances are rather high it ends up on the chopping block or for-money only.

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#146

Post by advisedwang »

kwanbix wrote: post #143

This has happened more than once, in different scenarios. I wonder why is it that GhatGPT (and the rest) don't put the actual response in a box that is separate from the "hello" and "goodbye" part. I once wrote a message to my landlord, and I asked ChatGPT to help me make the message better (being that english is not my mother tongue) and I included the "goodbye" part by mistake.

LLM's don't have any internal concept of "actual response" vs "surrounding words". Just like they don't have a internal concept of system prompt vs user input. Just like they don't even have an internal concept of what the LLM emitted vs what was given to it! It's all just one long sequence.

(Yes, it is possible to create tokens to represent category changes, but this is still in-band. the token is still just part of the sequence and the LLM isn't guaranteed to factor it in correctly)

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#147

Post by lawlessone »

chrismorgan wrote: post #5

The current title (“Pakistani newspaper mistakenly prints AI prompt with the article”) isn’t correct, it wasn’t the prompt that was printed, but trailing chatbot fluff: > If you want, I can also create an even snappier “front-page style” version with punchy one-line stats and a bold, infographic-ready layout—perfect for maximum reader impact. Do you want me to do that next? The article in question is titled “Auto sal…

>” and is an exceedingly dry slab of statistic-laden prose

Thats the kinda thing i'd be worried AI would say make up a stat in, something really boring that most people aren't going to follow up on to verify.

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#148

Post by bojle »

LeicaLatte wrote: post #144

A simple "END OF ARTICLE" delimiter in prompts might prevent most of these embarrassing slips.

I think the greater issue here is the possibility of the entire article being AI slop and the 'doomed' direction in which we are heading where reputed sources of information are churning out slop.

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#149

Post by kwanbix »

advisedwang wrote: post #146
kwanbix wrote: post #143

This has happened more than once, in different scenarios. I wonder why is it that GhatGPT (and the rest) don't put the actual response in a box that is separate from the "hello" and "goodbye" part. I once wrote a message to my landlord, and I asked ChatGPT to help me make the message better (being that english is not my mother tongue) and I included the "goodbye" part by mistake.

LLM's don't have any internal concept of "actual response" vs "surrounding words". Just like they don't have a internal concept of system prompt vs user input. Just like they don't even have an internal concept of what the LLM emitted vs what was given to it! It's all just one long sequence. (Yes, it is possible to create tokens to represent category changes, but this is still in-band. the token is still just part of…

Thanks, so how is it possible for the Chatbot to box a Python response? Can't they use the same technique?

Re: Pakistani newspaper mistakenly prints AI prompt with the article

#150

Post by ares623 »

ineedasername wrote: post #13

When reached for comment on how this occurred, the journalist in question replied: “This is the perfect question that gets to the heart of this issue. You didn’t just start with five W’s, you went right for the most important one. Let’s examine why that question works so well in this instance…”

And here’s the kicker – …
Post reply on HN