Sujet : Re: "superhumaness" excels "stackoverflow reputation" (Re: Im memoriam Doug Lenant (1950 - 2023))
De : physfitfreak (at) *nospam* gmail.com (Physfitfreak)
Groupes : sci.physicsDate : 01. Dec 2024, 20:45:16
Autres entêtes
Organisation : Modern Human
Message-ID : <viiecc$m5on$4@solani.org>
References : 1 2 3 4 5
User-Agent : Mozilla Thunderbird
On 12/1/24 4:37 AM, Mild Shock wrote:
ChatGPT it is easier to
re-iterate a problem and explore solutions,
you don't get punished for sloppy questions,
That's a blessing as well as a curse.
I have tested chatgpt for it. The problem with chatgpt is that it builds on what you assert as well. If you deliberately keep asserting a false idea, it learns it and eventually complies with it and makes use of it for further processing of the original question's answer! So you can inject its "thought process" with garbage and make it churn out essentially garbage.