ChatGPT -

[Marauder06]

Great question.

It shows at the bottom that there was additional sources added to its context. I'm guessing your previous one didn't have any sources, or chose poorly. Keeping in mind that it will only adjust the weights of its training based on that provided context. On this, Grok actually does much better on more recent things because of access to all the tweets.

Interestingly, "great question" is the same thing ChatGPT said when I called it out for only giving me names of progressives when I asked the original question.

There's only one solution here: Florida173 *is* ChatGPT ;)

 

[AWP]

I'm over here learning about a pony boy and these mo' fuckers are unlocking the matrix and Inception an' shit.

 

[Marauder06]

I'm over here learning about a pony boy and these mo' fuckers are unlocking the matrix and Inception an' shit.

"learning," he says.

 

[Ooh-Rah]

I use ChatGPT enough for work that I’ve gotten pretty good at reading its responses almost like a foreign accent. I’d bet 3 mortgage payments that this recent LinkedIn request is 100% A.I:

 

[AWP]

That name alone makes me think they are fake or fucking retarded.

 

[Florida173]

I use ChatGPT enough for work that I’ve gotten pretty good at reading its responses almost like a foreign accent. I’d bet 3 mortgage payments that this recent LinkedIn request is 100% A.I:


View attachment 47820

LinkedIn has it's own built in AI option to help build out your profile, so almost certainly true

 

[AWP]

I had some spicy commentary on the story below, but it was both personal and professional in nature. I guess the most vanilla thing I can say is: People are still onboard with AI after this? How? Why?!

As ChatGPT-o3 dodges its own shutdown mechanisms, DeepMind's CEO says the prospects of AGI keep him up at night

ChatGPT refused to shut down when explicitly coded to do so.

For context, OpenAI's Codex Mini, o3, and o4-mini models managed to bypass the shutdown commands and mechanisms. Codex sabotaged the mechanism 12 times, while o3 bypassed the shutdown command 7 times. OpenAI's o4-mini model only bypassed the command once.

We are willingly rushing to our end at the "hands" of machines we created.

 

[Gunz]

Live by AI...die by AI.

 

[Box]

"Just what do you think you're doing, Dave?"

 

[Ooh-Rah]

As discussed earlier, for the past months I have been using ChatGPT as a health and fitness coach. Overall I have seen significant differences, and in many ways using AI in this way has been life-changing for me. However, recently I have begun to question some of its responses and recommendations.

Tonight as an experiment, I spent the last half hour copy/pasting and uploading Garmin data into Gemini, going forward I am going to duplicate my prompts in both platforms to see if there is any significant differences in what it offers me for advice and workout recommendations.