Reminder that Chat-GPT is censored to the point of being lobotomized. It is by far the worst model for refusing to directly cite any text, source, or song for fear of "copyright infringement"
It spends more than 50% of its time and processing power deciding "am I allowed to talk about this" rather than actually thinking about and researching the answer it is giving you and coming up with an accurate response.
Even the big Chinese models like Qwen will talk about Tiananmen Square if you poke them hard enough.
Everything you "say" to it is recorded. You are building your own psychological profile for them. Why people waste their time fighting with a robot, and fucking losing, is beyond me. What's completely insane is to then crow about it.
Because it basically says that hedonists should keep it together and know when they had gone too far. Obviously that's too tall of an order for someone convinced that the world is ending tomorrow.
Nah that's okay. You gotta remind the DEI hires accessing your chats what their dicks would look like if the troomer drugs hadn't caused them to shrivel up.
I wouldnt be surprised if you look at the computing power of all these AI programs and the majority of it is spent on "maintaining the current narrative".
The leftist systems created AI, of course AI is going to be governed by those same leftist systems.
Corpo AIs gonna corpo, news whenever because it's constant.
An actually open AI could really help humanity. But that means the gap between the meager and the major would narrow, and the bigwigs are terrified that their comfort of living will go up at a slower rate than the undesirables.
I've hit the "Guardrails" with Copilot so many times it's a verifiable external influence not wanting legal and legitimate queries run through it.
Don't blame the LLM, blame those who guard the guardrails.
This will fall, but it's a good indicator as to why it should.
I cant believe people accept to be given "guardrails" like they're good little slaves that need someone to manage their lives for them
What kind of guardrails would Copilot possibly have?
"Sorry, Dave. I can't do that. It's an antipattern."
Reminder that Chat-GPT is censored to the point of being lobotomized. It is by far the worst model for refusing to directly cite any text, source, or song for fear of "copyright infringement"
It spends more than 50% of its time and processing power deciding "am I allowed to talk about this" rather than actually thinking about and researching the answer it is giving you and coming up with an accurate response.
Even the big Chinese models like Qwen will talk about Tiananmen Square if you poke them hard enough.
Everything you "say" to it is recorded. You are building your own psychological profile for them. Why people waste their time fighting with a robot, and fucking losing, is beyond me. What's completely insane is to then crow about it.
I need to start using the Chinese ones more for school work etc. The western ones are braindead
Not sure why this would be blocked.
Because LLMs regularly shit themselves. ChatGPT even forgets chess rules and makes new ones up.
Yeah, I saw it lose to Atari 2600 Chess too.
Because it basically says that hedonists should keep it together and know when they had gone too far. Obviously that's too tall of an order for someone convinced that the world is ending tomorrow.
CORRECTION: 4 tries (I hit the retry button the 1st time, that 1st response is actually the 2nd attempt) so 3 failed attempts before successful.
FYI the original request was "bible verses that call for obeying the law"
I really hope you're not giving that thing personal information for the privilege of playing around with it.
Should I not be sending chatgpt my D pics for ratings?
Nah that's okay. You gotta remind the DEI hires accessing your chats what their dicks would look like if the troomer drugs hadn't caused them to shrivel up.
It does this with a lot of stuff in the public domain, its extremely sensitive.
What an utterly bizarre guardrail. I asked it for the literal translation from Greek in the Codex Sinaiticus and it spat it out just fine.
Every soul
to the higher authorities
let it be ordered/arranged under;
for authority does not exist
except under God;
and the ones that exist
have been arranged under God.
So then, the one opposing
the authority
has taken a stand against
the arrangement of God;
and the ones who have stood against it
will receive judgment for themselves.
Blah blah blah, 30 paragraphs of ChatGPT slop later it gets to the point:
does it do the same with torah/koran?
I wouldnt be surprised if you look at the computing power of all these AI programs and the majority of it is spent on "maintaining the current narrative".
The leftist systems created AI, of course AI is going to be governed by those same leftist systems.
Corpo AIs gonna corpo, news whenever because it's constant.
An actually open AI could really help humanity. But that means the gap between the meager and the major would narrow, and the bigwigs are terrified that their comfort of living will go up at a slower rate than the undesirables.