[swatcoder]

Hah. Of course the authors are not aware.

OpenAI handed this tool to people with effectively no guidance or instructions so that they can trade on the glow of misconceptions and harvest user data. The whole point of giving the thing away for free these last few months is to let people make fools of themselves like this while pumping up OpenAI’s value as a technology firm. It’s no accident that they don’t engage to clarify all these persistent misunderstandings.

[goatlover]

The concern is the increasing number of anecdotes where people are taking ChatGPT's output at face value, even though everyone knows by now that it does sometimes get stuff wrong and does hallucinate. I thought it would take longer.

[HideousKojima]

A very annoying trend I've already started noticing on some forums (including here unfortunately) is someone asking a question and one of the replies is someone saying "I put your question into ChatGPT and it said..."

[Lewton]

I downvote and flag every chatgpt generated comment, even if they’re up front about it

If I want to know what chatgpt thinks about something, I can ask it myself

[breakingcups]

Case in point: https://news.ycombinator.com/item?id=35031737

[rideontime]

Apparently everyone doesn't!

[dcow]

The problem is that people also sometimes get stuff wrong and hallucinate. So ChatGPT is just mimicking how we behave and it seems completely normal.

[ehnto]

While it's true humans are fallible, it is more clear what someome should and should not get wrong, and we can make safer assumptions about the knowledge we get from certain individuals. ChatGPT doesn't "know" anything, you can get it to be right and wrong about the same detail in the same conversation. It's not clear what it should be correct about, and so it's a pretty poor source of information.

[dcow]

And then there are trolls.

[renewiltord]

> OpenAI handed this tool to people with effectively no guidance or instructions so that they can trade on the glow of misconceptions and harvest user data.

When one logs on to Chat GPT, they get the following notices:

> This is a free research preview.

> Our goal is to get external feedback in order to improve our systems and make them safer.

> While we have safeguards in place, the system may occasionally generate incorrect or misleading information and produce offensive or biased content. It is not intended to give advice.

> Conversations may be reviewed by our AI trainers to improve our systems.

> Please don't share any sensitive information in your conversations.

Effectively no guidance or instructions indeed. Perhaps ChatGPT is sometimes confidently incorrect because it learned from online commenters.

[swatcoder]

Did you read those statements?

None of them say what the product is, what it does, or how to use it — which are the kind of thing one might call guidance and instructions.

Those are just disclaimers to make sure you don’t hold OpenAI responsible for anything you do with this thing, or blame them for anything they do with your attempts.

[reacweb]

They say what it is: "This is a free research preview. Our goal is to get external feedback in order to improve our systems and make them safer."

It is not a product. They open it so that public can contribute to help them.

They do not say the aim of this research.

[creatonez]

This is baseless speculation