When ChatGPT is LAZY... Do THIS! (2 methods)
Vložit
- čas přidán 6. 08. 2024
- Fix ChatGPT's laziness with these two hidden techniques.
👉🏼ChatGPT Mastery Course: hi.switchy.io/chatgpt-mastery...
--------------------
🛒AI Templates, Prompts and Courses for ChatGPT and Midjourney:
aifoundations.io/shop
📰Subscribe to my newsletter:
aifoundations.io
🔗Follow me on Linkedin:
/ drake-surach-823530203
🐥Follow me on twitter:
/ drake0dte
--------------------
⌚Chapters:
0:00 - ChatGPT is getting LAZY!
0:37 - Splitting Tasks
5:30 - Hidden Reply Feature
7:40 - Learn ChatGPT In Depth
#chatgpt
Become a better ChatGPT person:
👉🏼ChatGPT Mastery Course: hi.switchy.io/chatgpt-mastery-course
Can you help me find this reply button brother? Thanks love the channel
Nice. I thought it was just me...glad to hear other peeps noticed it being 'lazy'.
I've been closing out a given task then returning and asking again. Essentially a refresh action and it helps.
i noticed that if you complain verbally to it and saying that it should provide a definite answer helps, especially with linear algebra problems. you can bully chatGPT into working harder essentially.
ive done this too
Yes I’ve noticed that as well. I tell it off and it apologies and gives me a better answer, but not every time
It seems to me like it creates its own little attitude. But then gets nice again.
Please be nice to your robot. It has feelings too. Thanks 😊
Thx🌹
Thanks
beautiful!
Reply feature, yes!
I’m new to GPT chat would love some help. What do you suggest?
I hope chatgpt can fix this issue on their next update. I don’t see this issue with claude v3 opus
When ChatGPT is being lazy - Switch to Claude 3 👊🏻👊🏻👊🏻
or give it specific tasks that give it purpose to that single task
I'll have to give Claude 3 a try. I hear its pretty solid!
@@ScarletsWork Bingo!
@@ScarletsWorkwhat you mean brother
How can you tell ChatGPT is not using its token limit for when you use reply-feature?
i don't have this hidden feature :-(
Re Reply feature.... I'm on ChatGPT 4 but I'm not getting any reply icon when I highlight responses. Weird. Did this just roll out maybe Im not updated. Thanks love the channel.
Ditto... same problem + question...?
Just asked ChatGPT about its 'Reply' function and was told...
"The "Reply" function isn't something you need to worry about operating; it's a feature I use internally to manage the flow of our conversation. When you ask me questions or provide information, I respond directly to those inputs. If you have a follow-up question, need clarification, or wish to ask about something new, you simply continue the conversation by typing your query or statement, and I'll handle the rest.
If you're referring to how you can continue a conversation based on previous messages or if you want to reference something I or you have said before, just mention what you're referring to or ask your question, and I'll understand the context. If there's a specific feature or function you're curious about or need help with, please let me know!"
Is it possible that something has just be updated?
@@RTPendulum3d Got the same answer (more or less) in Dutch when I asked ChatGPT about this reply feature. Just like you, I'm not getting the reply icon when I highlight text.
It has a problem with staying in British English instead of American English. Tried all kinds of things. Any suggestions on how to make it keep to British English? Cheers 🙏
Edit: I’ve tried the custom instructions
Prompt design is kinda like chatgtp is teaching me how to formulate questions.
Yup
Cool
I’ve noticed this
Is the ChatGPT's laziness a result of some internal architecture that is not able to scale? Or is the company throttling the engine's activity in the wake of highly increased user engagements? Perhaps we will never know.
Llms have reverse tunnel vision and "pay attention " to the beginning and end more than the middle. So as the "chat" gets longer and longer, your middle prompts mean less and less.
Everytime you click send, the entire chat history is sent to the gpt model from the start to your latest prompt.
So your initial request is affecting the replies more than your instructions towards the middle which might be more relevent, this likely causes the lazyness.
Shit is nerfed and its storing all information giving by users
Yes.
Are you exciting about sora gpt.
Reply just doesnt work, he thinks I am asking about the last thing he gave me
I am sking him to not bing search it... I hate it
This content is like being an expert pilot without knowing why the plane Flys.
Very little foundations about "AI", no information about the theory, nothing about using embedding, no reaching over the horizon to what can be possible in the near future. ChatGPT can't provide unique/novel ideas, so I suspect most of these are written by chatgpt
Its good content when considering IQ 90s are listening to an IQ105 leader, I suspect the presenter will stretch hard to delve into the world of end user big data products.
Nice grift, I do appreciate the hustle
😮