Hi
@SERA2g,
with all due respect, I think you're totally missing the point here.
@Bravo claims her vetted ChatGPT posts are a trustworthy source of information, as she would only ask her “friend Chatty” about topics she knows well, and would then “review, verify and refine” the replies before sharing them with us. Something I personally believe to be nice-sounding in theory, but unrealistic in practice.
When I dared to challenge this claim with supporting evidence of a hallucination I had spotted in one of the ChatGPT replies she had previously posted - combined with a dash of sarcasm (given my history with her) - I got immediately attacked by others for doing so, in a much nastier way than my post was, with those bullying me predictably deflecting from the ACTUAL issue at hand here, which is that
@Bravo's claim, while sounding noble, unfortunately doesn’t stand up to reality.
And no, it didn't take a lot of effort for me to write that post at all, in contrast to other posts of mine into which I’ve invested a LOT of time both researching and writing, such as the recent ones about my observations with regard to Spartan Group and the "The Mississippi Connection".
It was prescisely from that research that I knew Cindy Hyde-Smith does not sit on the Defense Subcommittee. So it was easy for me to spot said hallucination just by glancing over the ChatGPT replies
@Bravo had previously shared on that topic. And it didn't take more than a quick Google search to copy and paste the senate.gov links as supporting evidence and to re-quote our CTO on his experience that GPT-5 is still massively hallucinating, an opinion that to me personally has more credibility than the opposing opinion of an anonymous forum poster.
Fact is, and I trust you will agree with me on that, it distorts our interpretation of Jonathan Tapson's recent visit to Capitol Hill when posters falsely suggest he had been invited to meetings at the White House - like
@manny100 recently did - or post unverified ChatGPT replies that claim the meeting with Senator Cindy Hyde-Smith "to promote our Federal contracting agenda" (to quote Jonathan Tapson himself) matters because she sits on the Defense Subcommittee of the Senate Committee of Appropriations (when this is simply not true), or misleading ChatGPT statements such as that our CDO "met with powerful committees (Armed Services, Appropriations)".
For all we know, he merely met with 1 out of 29 senators sitting on the Senate Appropriations Committee and another 1 out of 27 members sitting on the Armed Services Committee, not with both committees in their entirety, as the following ChatGPT reply posted by
@7für7 implies:
View attachment 90900
Similarly here, where ChatGPT misleadingly claimed our CDO got “invited to meet with sitting U.S. Senators (
including members of the Senate Appropriations and Armed Services Committees)”, as that equally means he would have met with more than just the two US Senators Jonathan Tapson had named in his LinkedIn post, for which there is no evidence whatsoever:
View attachment 90899
False or misleading information of that kind results in shareholders reading things into what our company actually says that are exaggerated or totally unrealistic and occasionally even breed conspiracy theories.
While Roger Wicker happens to be the Chairman of the Armed Services Committee, I personally believe that the
main reason why our CDO got to meet with Cindy Hyde-Smith and Roger Wicker instead of any of the other 98 Senators was not because they sit on this or that committee (others do, too), but because they are the two Senators representing Mississippi in Congress, the state where the Spartan Group team that Jonathan Tapson thanked in his LinkedIn post is located.
Speaking of Spartan Group: ChatGPT also claimed the following:
https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-471608
View attachment 90903
What is ChatGPT’s source here? This does not align at all with how they present themselves online. As I mentioned in one of my posts, they only seem to have recently started to offer federal government contracting consulting services. What they do specialise in to date is in tactical training for individuals, security personnel, law enforcement officers and military members, proactive security measures, risk assessment etc., which makes sense given that the company was founded by members of the US Special Operations Community.
Again, this information by ChatGPT did not get verified by the poster before sharing and hence paints the impression we get of this company in a certain way which does not appear to reflect the reality.
Forum users are free to post their ChatGPT replies as long as readers are clearly made aware that the content is AI-generated.
At the same time, however, other forum users - including me - have every right to criticise what we see as a dangerous trend to embrace the idea of LLM applications being considered the “ultimate search engine” and to take their generated content at face value, despite the notorious percentage of hallucinations and inaccuracies those models are known to produce, which makes this tool unsuitable to use it the way most people here do.
So when
@Bravo makes bold claims about only consulting "Chattie" on topics she is familiar with and about diligently fact-checking the main points before posting “to ensure that ChatGPT is not hallucinating”, she needs to live with others challenging these claims when they see evidence to the contrary.
I see it as a service to keep this forum factual when posters claim factualness, yet, don’t live up to it, and this will ultimately benefit us all.
️ Our super-intelligent operating system, MB.OS.
. Honestly, do you really think that I am Frangipani? How can I prove that I am not?
engrish please
