Some people would say that for truly unbaised news reporting, they should go to 
Fox Entertainment Network. Others would say NBC/MSNBC. Others would have their 
own opinion.

Just remember that ChatGPT’s knowledge base is a composite of everything it was 
trained with. Just like most people. 

The increased use of paywalls in front of so many old-school news organizations 
means that there’s an overabundance of “free” material that’s mostly from 
sources that aren’t so picky about what they publish. 

LLMs are a mirror of what they find plus some logic that may or may not alter 
things to resolve ambiguities. One of the services that makes graphics can’t 
seem to spell anything you give it correctly, although the artwork it creates 
can be quite interesting. But artwork filled with typos isn’t very appealing.

Has anybody noticed how many things come up in “reels” that have the wrong 
photos showing up for what the announcer is describing? If it’s talking about 
“Old Joe” then the image could be anybody named “Joe” who’s old. Is that 
“accurate” and “unbiased”? I guess you could say that, although it’s flat-out 
wrong for the given context. If you don’t know that, then you’re being 
mis-informed.

If you can’t SEE the images, then you’re certainly not going to know about any 
of this.

But as others have said, and I’ve frequently experienced myself, ChatGPT is 
great at showing code that refers to libraries and functions that simply don’t 
exist. They make sense, but if you aren’t experienced with that particular 
platform, you won’t know until you try to use the code, and then it will just 
keep giving you errors.

And when claiming ChatGPT is  giving you unbiased information about politics, 
especially anything that involves Trump, you’re fooling yourself. Trump says 
whatever he wants and doesn’t care because he conned the SCOTUS into ruling 
that he has “absolute immunity”. 

But when he announces that Windows is really run by the Chinese and he issues 
an EO mandating that all Federal agencies replace all Windows installations 
with Linux, then you can probably believe him, because he’d never promote 
something he can’t profit from.

Old Joe is probably turning in his grave (whichever Old Joe you care to 
imagine).

My personal stand is that I won’t trust anything I get from ChatGPT without 
first reviewing it, and I’m not going to stand behind anything I put out if 
it’s on some topic I can’t verify its accuracy one way or the other. As a 
programmer, that’s pretty easy. But when it comes to stuff from the news and 
especially politics, that’s a minefield that we should all avoid. 

LLMs are subject to the same GIGO axiom that all other computerized databases 
are. 

-David Schwartz




> On Jan 24, 2025, at 5:42 PM, Michael via PLUG-discuss 
> <plug-discuss@lists.phxlinux.org> wrote:
> 
> I found a very good means of getting unbiased reports about anything. Reports 
> the way they used to be. Just the news. Make your own decisions.  Ask 
> chatgpt. And then you can go go further into the information. For instance I 
> asked for the latest news on trump and it told me he was in California and 
> that he might withhold fema funds. I then asked why and it told me and then 
> it said  that most of the land was under federal responsibility I then asked 
> why the government didn't clean the land.... etc.
> ---------------------------------------------------
> PLUG-discuss mailing list: PLUG-discuss@lists.phxlinux.org
> To subscribe, unsubscribe, or to change your mail settings:
> https://lists.phxlinux.org/mailman/listinfo/plug-discuss

---------------------------------------------------
PLUG-discuss mailing list: PLUG-discuss@lists.phxlinux.org
To subscribe, unsubscribe, or to change your mail settings:
https://lists.phxlinux.org/mailman/listinfo/plug-discuss

Reply via email to