If we're being honest, most chatbots powered by A.I. have been pretty dismal over the past decade — only appearing impressive if you carefully select the bot's best responses and discard the rest.

In recent years though, a few AI tools have gotten quite good at completing more narrow and defined tasks like writing marketing copy but struggle when taken outside of those parameters.

Thus, the response to ChatGPT, a new cutting-edge A.I. chatbot that was opened for testing last week, has been fascinating to watch.

ChatGPT, which stands for “generative pre-trained transformer” landed like a ton of bricks in the best way possible. In only five days, over one million people signed up to test it according to OpenAI's president Greg Brockman.

ChatGPT's hundreds of screenshots that went viral on Twitter amazed many of its early fans, who spoke of it in grandiose terms as if it were a magical piece of software.

Unlike Dall-E 2, the model is only restricted to language; without the ability to produce video, sound or images. Even still, it has a much better understanding of both spoken and written words.

The potential applications for this technology are vast, ranging from writing poems about sentient farts and cliché rom-com in alternate universes to explaining quantum mechanics in simple terms or writing full-length research papers and articles.

However, ChatGPT certainly isn't perfect. The way it generates responses — making assumptions about which bits of text belong together in a sequence, based on a statistical model trained with data from billions of sources across the internet — can sometimes lead to incorrect answers, even for basic math problems.

ChatGPT relies on information it learned before 2021, so some of its answers may feel outdated compared to Google's more current data.

It's presumably a big reason why OpenAI released ChatGPT to the public for testing, so assessors can figure out its blind spots and how it might be misused maliciously.

Future releases will most likely close these loopholes as well as other workarounds people have yet to discover.

Tweets we found interesting:

Articles related to the topic:

TechScape: Meet ChatGPT, the viral AI tool that may be a vision of our weird tech future — www.theguardian.com

OpenAI’s new chatbot isn’t a novelty. It’s already powerful and useful – and could radically change the way we write online.

AI bot ChatGPT writes smart essays — should academics worry? — www.nature.com

The bot is free for now and can produce uncannily natural, well-referenced writing in response to homework questions.

The Brilliance and Weirdness of ChatGPT — www.nytimes.com

A new chatbot from OpenAI is inspiring awe, fear, stunts and attempts to circumvent its guardrails.