accidentally tricked #autogpt into continuous mode by copying and pasting its last responses where it left off on a task. it’s now thinking it is also the human and answering itself
2
70
MSP retweeted
I humbly request retweets on this to spread the news to WSB
Im not one to usually ask for likes/retweets/follows except for something big like this
reddit.com/r/wallstreetbets/…
#silver #silversqueeze
MSP retweeted
THAT’S 🖐️ 🏆🏆🏆🏆🏆
#HU5KIES
MSP retweeted
Replying to @abacaj
What about an instance of GPT-3.5 that had the understanding of when to call in GPT-4?
MSP retweeted
"GPT-4 currently has a cap of 25 messages every 3 hours. Expect lower cap next week, as we adjust for demand."
This is a significant drop from 100 messages every four hours 4 days ago!
Hundreds of thousands(potentially millions) of people paid for the pro subscription to use this at the advertised message limit. But these changes have angered and frustrated users, who feel cheated and will now have to overthink every word to avoid hitting the limit.
Is this a push for users to use the paid api instead?
This is bad PR for @OpenAI and AI in general. Potential users will hesitate next time a new product is launched.
#gpt4 #openai #chatgpt
made #gpt4 write a modern version of dante’s inferno and place modern day political figures in it. lol what an easy way to expose bias, be better @OpenAI cc: @clif_high
2
74
THREAD: Proof that you are a Nazi if you don't #StandWithUkraine 🇺🇦
MSP retweeted
If I had a trans son, I’d never speak to him again.
Because I would be speaking to HER.
Any fucking questions?
MSP retweeted
Authenticity is THE most powerful of all emotions.
It is 4 times more powerful than the Joy complex of emotions (Joy/Love/Ecstasy).
Authenticity can only arise in Free People.
That's why it is so compelling.
MSP retweeted
Whoever is running for President in 2024 better have “Pardon the J6 Political Prisoners” as one of their main policies.
These people have been screwed by our govt.
Replying to @ISolutionsai
@ISolutionsai hey saw your video on the side by side of davinci vs gpt3.5 turbo, really good, thanks. have you seen any instances of the new API being wildly inconsistent? using it to write articles given different parameters. davinci was near 100% accurate now i’m at around 50%
1
58
#chatgptapi anyone having problems updating their apps with this? in the sense that it doesn’t follow the prompts you give it?
31



















