Please enable JavaScript.
Coggle requires JavaScript to display documents.
The Brilliance And Weirdness Of ChatGPT - Coggle Diagram
The Brilliance And
Weirdness Of ChatGPT
INTRODUCTION
Was built by OpenAI, the San Francisco A.I. company
Stands for ''generative pre-trained transformer''
Over a million people signed up to test ChatGPT in five days
Different. Smarter.
Weirder. More flexible
working computer code and college-level essays
can also guess at medical diagnoses, create text-based Harry Potter games
explain scientific concepts at multiple levels of difficulty
can write jokes (some of which are actually funny)
Based on what the company calls ''GPT-3.5,'' an upgraded version of GPT-3
Available to the general public through a free, easy-to-use web interface
APPLICATIONS
Good at helping programmers spot and fix errors in their code
Can remember or learn from previous conversations
Good at answering the types of open-ended analytical questions
Possible to create personalized therapy bots
LIMITATIONS
Prone to giving
wrong answers
Making probabilistic guesses about which bits of text belong together in a sequence, based on a statistical model
Trained on billions of examples of text pulled from all over the internet
Stale answering due to its knowledge
Restricted to things it learned before 2021
An instance of the bot generating several topical jokes about former President Donald J. Trump pulling out of the Paris climate accords when prompted to write the opening monologue for a late-night show
Moderate by design
The inclusion of billions of examples of human opinion, representing every conceivable view, in its training data.
ChatGPT's tendency to provide an impartial summary of both sides' beliefs without explicit prompting on charged political debates
Users have found
out how to exploit ChatGPT
OpenAI has programmed the bot to refuse ''inappropriate requests'' -- a nebulous category that appears to include no-nos like generating instructions for illegal activities
Users' circumvention of
many of these guardrails
Rephrasing a request for illicit instructions as a hypothetical thought experiment
Asking it to write a scene from a play or instructing the bot to disable its own safety features
CONCLUSION
OpenAI released the bot to the public for testing
Assessing ChatGPT's blind spots
Figuring out how it might be misused for harmful purposes are, presumably
The unpredictability of ChatGPT's future
The beginning of the end of all white-collar knowledge work
A precursor to mass unemployment