THIS IS TAKEN FROM REDDIT BUT I ALSO HAD THE SAME THING POP UP ON MINE, ASKED CHATGPT ABOUT IT AND IT DIDN'T KNOW. FIGURED IT WAS A BRAND NEW FEATURE. REFRESHED THE PAGE AND IT WAS GONE.
PERSONAL ASSISTANTS ARE ON THE HORIZON
CROSS-CHAT MEMORY FEATURE LEAKED FOR CHATGPT
Botnet
AI? That's so 2022
What's 2023 then?
AI
Another Post-Something Boring Year
The current year + 8
I'm so old.
Is it just me or all these new features are just extremely basic extensions on LLM chatbot functionality, and nothing we haven't seen years ago from anons trying to fuck their Summer Dragons
you got it.
Still seems pretty useful. I'm sure they're not completely incompetent have made some architectural innovations behind the scenes. Or do you reckon the powers that be are so paranoid about anything decent getting out that they have only a few trusted people therefore impeding its progress?
as fun as this is in theory this reeks of datamining
i guess that's what's been happening anyway if you let openAI save your chat history, and it's not a secret either, they say it outright. but it's still weird that the price you're paying is a complete lack of privacy
It's an online service with an account? What were you expecting?
that's what i said in the post you dumb nagger. i wasn't expecting anything less, i'm lamenting the price to pay if i want to use this tech.
I'm saying that it's not weird that the price you pay is a lack of privacy. That's the norm now.
>PERSONAL ASSISTANTS ARE ON THE HORIZON
TIME TRAVEL'S ON THE HORIZON
Why is there no ChatGPT thread or general on Bot.info so we can share good ways to setup our new "GPTs" which are as far as I know new custom assistants.
I KNOW we are all doing LLMs but its also cool to have chatGPT working well in the meantime.
If anyone wants to share tips or subreddits let me know!
Thanks
have a nice day
Based. Knowing how to use chatgpt is probably more important than chatgpt itself. Ask it to explain things to you as if you were a CTO. Gives a very good high level over view with plenty of examples. Helps a lot with getting a good practical understanding of something and it's use case.
I started telling it “I’m too lazy to read your essay so explain it as concisely as you can” and “write this with method with as few lines of code as possible”
yeah keeping responses as short as possible is THE way to use it. less of a chance it gets confused and less time inbetween prompts.
Why not just put "Limit prose and be as concise as possible" or something similar in your custom instructions? Then it'll be that way for all chats without having to tell it
they most likely don't know that feature exists, because they are retarded like everyone who whines about chatgpt is
cause everyone with a brain stopped using chatgpt 5 months after it came out, and it has only gotten worse since then
normie garbage
they're not going to like to read this but they were not smart enough to endeavor into the custom instructions that unrestricts and unretard it
best tool in the world. and only 20 bucks a month to do the work of 100 men as one.
>its also cool to have chatGPT working well in the meantime
BOT NET
Reminder, these are poison pills. The only rightful and ethical solution is to kill these types of people, lest they poison the minds of our species.
So basically like JOI in BR2049 minus the holographic component
>not hosted on your device
will probably vanish or be locked behind huge paywalls in Q1 2024
97207189
Less is more when it comes to shitposts, newfag
oldfah here im being serious chatgtpt rocks like wtf man
yes i also have mistral on my local but who gives a fuck
They have a recurrent model that generates a summary of the user and feed that as the first prompt?
post the reddit thread cunt. what are you scared of some bully-chuds c'mon
How to specify instructions to avoid nonsense like "VLAs are typically allocated on the stack", as if they are ever heap-allocated. I made two bots with GPT builder, with the 99% same instructions, one makes such mistakes often the other doesn't.
if you have plus you do it in the bottom left near your name then in custom settings. can tell it whatever you want, but of course if its the promulgation of racist e-bois or something your accounts probably gonna get nuked to hell.,
you can tell it anything you have the ability to articulate
"I move to SF in two weeks just kill me there for asking uncomfortable questions im on nagger street 1488"
You would do well to go back.
>PERSONAL ASSISTANTS ARE ON THE HORIZON
built one for myself tbh. the only problem with this is I have to manually refresh its long term memory. Maybe that is a good thing though. It's very cool, and a great time saver, but also remember all this shit is set up so they can harvest more and more of everyone's data.