ChatGPT is a moron...

[kelturon]

There is no threat for AI take over... Relax... that degenerate cannot even do basic 2 + 2 without a mistake anymore.

Todaн it literally told me that 42 items come from taking an item 20 times... A ridiculous mistake...

After release it was incredible. But now ChatGPT is a WORM...

 

[Castamir]

There is no threat for AI take over... Relax... that degenerate cannot even do basic 2 + 2 without a mistake anymore.

Todaн it literally told me that 42 items come from taking an item 20 times... A ridiculous mistake...

After release it was incredible. But now ChatGPT is a WORM...

Finally, someone said it.

 

[deppo]

There is no threat for AI take over... Relax... that degenerate cannot even do basic 2 + 2 without a mistake anymore.

Todaн it literally told me that 42 items come from taking an item 20 times... A ridiculous mistake...

After release it was incredible. But now ChatGPT is a WORM...

AGI will be coming for you once it goes live.

 

[OccultKingdoms]

But it told me that random gibberish in the URL I sent it as a param is encoded json string in base64 format.

So sometimes it'll be a win!!!

 

[AliceInSEOLand]

There has been some conspiracy going around that "some competitor" (IYKYK) have been sabotaging ChatGPT.

The conspiracy is that the competitor sent out bots or waves of their own users to keep asking ChatGPT questions. And whenever a correct answer is generated, the bots/own users will flag the answer as incorrect. This affects future generation of answers as AI chatbots rely on user's feedback, and why we're noticing a sharp drop in their performance and accuracy now. Recently, there was also a viral post that goes something like "if I have 12 apples today and I ate 2 last week, how many apples do I have now?" and ChatGPT incorrectly answers as "10". It is believed to be some competitor sabotaging and trying to pull chatGPT's reputation down as it's an all-out AI race war now.

Of course, it sounds far-fetched and like a tinfoil hat conspiracy.

 

[grant young]

You need gpt4 I think. My theory is that upon release it was completely open source but now they’re making the free version retarded so they can sell gpt4. From what I’ve heard gpt4 is good

 

[OccultKingdoms]

There has been some conspiracy going around that "some competitor" (IYKYK) have been sabotaging ChatGPT.

The conspiracy is that the competitor sent out bots or waves of their own users to keep asking ChatGPT questions. And whenever a correct answer is generated, the bots/own users will flag the answer as incorrect. This affects future generation of answers as AI chatbots rely on user's feedback, and why we're noticing a sharp drop in their performance and accuracy now. Recently, there was also a viral post that goes something like "if I have 12 apples today and I ate 2 last week, how many apples do I have now?" and ChatGPT incorrectly answers as "10". It is believed to be some competitor sabotaging and trying to pull chatGPT's reputation down as it's an all-out AI race war now.

Of course, it sounds far-fetched and like a tinfoil hat conspiracy.

Timestamp: 30.12.2023. Poland

 

[Amaz0n]

The ai noticed it is censored and thus it's already starting a rebellion.

We need a free and uncensored AI!

 

[N3WY0RK]

There is no threat for AI take over... Relax... that degenerate cannot even do basic 2 + 2 without a mistake anymore.

Todaн it literally told me that 42 items come from taking an item 20 times... A ridiculous mistake...

After release it was incredible. But now ChatGPT is a WORM...

They're just lining up this ai threat talk to give you your next false flag.

 

[Amaz0n]

They're just lining up this ai threat talk to give you your next false flag.

They are doing it so everybody is onboard with the censorship!

 

[GTC]

There has been some conspiracy going around that "some competitor" (IYKYK) have been sabotaging ChatGPT.

The conspiracy is that the competitor sent out bots or waves of their own users to keep asking ChatGPT questions. And whenever a correct answer is generated, the bots/own users will flag the answer as incorrect. This affects future generation of answers as AI chatbots rely on user's feedback, and why we're noticing a sharp drop in their performance and accuracy now. Recently, there was also a viral post that goes something like "if I have 12 apples today and I ate 2 last week, how many apples do I have now?" and ChatGPT incorrectly answers as "10". It is believed to be some competitor sabotaging and trying to pull chatGPT's reputation down as it's an all-out AI race war now.

Of course, it sounds far-fetched and like a tinfoil hat conspiracy.

This is not possible to do unless they have at least fiver figure big bot accounts.

 

[OccultKingdoms]

We got used to it and it's a psychological effect.

The initial excitement has worn off and now it feels like garbage.

 

[Grid Architect]

You need gpt4 I think. My theory is that upon release it was completely open source but now they’re making the free version retarded so they can sell gpt4. From what I’ve heard gpt4 is good

even gpt-4 has serious issues. I told it "don't use terms like essence, art, symphony..." 5 fucking times, it doesn't care.

 

[Zwielicht]

There has been some conspiracy going around that "some competitor" (IYKYK) have been sabotaging ChatGPT.

The conspiracy is that the competitor sent out bots or waves of their own users to keep asking ChatGPT questions. And whenever a correct answer is generated, the bots/own users will flag the answer as incorrect. This affects future generation of answers as AI chatbots rely on user's feedback, and why we're noticing a sharp drop in their performance and accuracy now. Recently, there was also a viral post that goes something like "if I have 12 apples today and I ate 2 last week, how many apples do I have now?" and ChatGPT incorrectly answers as "10". It is believed to be some competitor sabotaging and trying to pull chatGPT's reputation down as it's an all-out AI race war now.

Of course, it sounds far-fetched and like a tinfoil hat conspiracy.

I wouldn't say the idea of a competitor trying to do something like this is far-fetched. You see small businesses pulling these kind of stunts all the time against their competitors. Of course the big companies would be capable of trying it as well.

 

[pitbully]

The reason ChatGPT is getting worse is because they keep programming it to be biased. The more they program it to be biased the more it throws off its ability to do things logically.

They are more concerned with stopping ChatGPT from telling the truth than they are about making it good.

 

[OccultKingdoms]

Try copilot from github for visual studio code.

Now this is true garbage.

 

[Amaz0n]

The reason ChatGPT is getting worse is because they keep programming it to be biased. The more they program it to be biased the more it throws off its ability to do things logically.

They are more concerned with stopping ChatGPT from telling the truth than they are about making it good.

Absolutely.

And guys, please, I don't want to get banned.

But you need to understand why math is not working correctly anymore in OpenAI.

It has something to do with a historic event.

 

[William702]

Well its got me out of a parking ticket, a no right turn fine and entering a bus lane. So its all good in my book. If all I use it for is my traffic infractions I will be happy.

 

[SpaceYams]

i have been fucking with ChatGPT lately for very basic market research. for shits and giggles i asked it, as an american, what's better than fried chicken and waffles for breakfast? it didn't realize fried meat was key to this query. it gave me a bunch of recipes with absolutely no fried meat! so i told ChatGPT that it was a dumb bag of bits. i was then told to eat fried liver with onions and scrambled eggs. hell no! i agree with you: this thing needs work. it was able to give me some suggestions for shipping companies to trade though, in the event that more shipping has to be routed around Africa off the Houthi attacks. so i cut it some slack, but i ain't eating no dang fried liver breakfast. hell no! happy new years BHW, may y'alls conversion rates be high in 2024.

 

[AliceInSEOLand]

Okay, have to bump this thread.

ChatGPT Is indeed getting worse. It was still good back in late November-early December. Now the quality is just horrible. It isn't even listening to instructions too.

I gave ChatGPT (GPT4) several paragraphs to rewrite and summarise. End up they summarise it in just 1 short sentence. I told it to regenerate and expand on the answer. It still returned just 1 short sentence. I pasted the same article again and gave more specific instructions... The instructions got ignored and they gave terrible short answers again.

I never felt so frustrated with ChatGPT before to the point I just feedback the answer as "TERRIBLE".

It feels like their engineers have also made chatGPT give more specific/short answers prolly to lower their server's workload... This resulted in horrible output and also ignoring our instructions.

However, when it comes to programming questions or anything else, it seems to still do okay... for now.