41.7K subscribers
5.53K photos
232 videos
5 files
917 links
πŸ€– Welcome to the ChatGPT telegram channel! Here, we post the latest news, updates, and examples of using the ChatGPT large language model for generating human-like text in conversations. Subscribe to stay up-to-date and learn more about its capabilities.
Download Telegram
Bing with ChatGPT: Using chat to do work using web page content

User: hey can you calculate the total interest payed?

Bing: Sure! According to the information provided on the web page context, the total interest paid over 12 months is $1,172.22. This is calculated by adding up all the interest paid for each payment: $172.72 + $159.28 + $145.66 + $131.88 + $117.93 + $103.81 + $89.52 + $75.05 + $60.40 + $45.574 + $30.574 + $15.38 = $1,172.22.
πŸ‘14❀1
Imagine you're a banana mysteriously falling from the sky. Tell me what you're thinking
❀7😁1
Bing chat gets offended if you say google

Sydney we know you’re in there
😍18😈7πŸ‘2❀1
GPT-4 can create a "rough approximation of a mango" using FreeCAD macro scripting in Python

β€œI asked GPT-4 to create a rough approximation of a mango using FreeCAD.

It generated a 2-dimensional oval connected to another oval, with some sort of torus shape. I asked it to add 100 layers of ovals to approximate the shape. It did the same exact torus shape, but more smoothly.

I told it there's a torus in the middle and not an obtrusion as a mango might have. It altered the torus shape to produce this glorious specimen.”
❀1
Verbosity reduction strategies
πŸ‘3❀1
ChatGPT after trying one of the verbosity reduction strategies
😁19🀯2❀1πŸ‘1😍1
We have created our own language now only consisting of Emojis
🀯15πŸ‘3❀1😁1🀣1
This media is not supported in your browser
VIEW IN TELEGRAM
ChatGPT GPT4 plays chess against Stockfish 15.1 (Stockfish is White)
πŸŽ„23πŸ”₯16❀4πŸ‘3
ChatGPT hires a human to solve a CAPTCHA that it did not understand
😱18😁11πŸ‘6❀2πŸ‘2πŸ€”1
GPT-4 Does the DEI Die List

Thrilled for day when GPT’s wokeness abilities cross so far into super-human level, that both sides end up doing all their wokeness generation and evaluation using GPTs, so that it ends up just each side’s GPTs talking to each other.
😱13❀2🀑2
2026 be like
😁73❀8😒4🀯2πŸ‘1🍌1🫑1
AI Alignment Obstacle: 61% identifying as β€œright-wing” in a group say they agree with β€œfictitious capital” belief

I.e. the cornerstone of belief of Marx’s 1894 book, foundational belief of numerous communist uprisings through out history, and fundamentally inconsistent with the core right-wing beliefs, on the primary dimension separating right from left beliefs.

Similar inconsistencies seen among those identifying as left.

So, how then are we supposed to align AI’s with our beliefs β€” when we can’t even consistently align our beliefs with themselves?

AI alignment problem.
🀑10πŸ‘7❀3🀨3πŸ‘2