How are you folks using ChatGPT in your daily life(except work)?

I use it to convert legacy JavaScript to Typescript quickly, and to write small, verifiable functions for domains I’m not quite familiar with.

The other day I had it teach me some WebGL stuff. You always have to verify what it’s telling you, but I’m finding 95% of the time its knowledge is correct. It gives nice starting points for diving deeper into a topic.

I will say one thing it’s bad at: Writing SQL. Like awful; anything beyond a simple select statement becomes a huge, inefficient mess. I’d never have it write mutation statements just for data integrity reasons, but even simple read-only reports were a challenge for it.

Some things * I use it as a short cut for search queries. * I add supplementary information to things (e.g. who else writes about this topic) * I use it to write things that are basically boilerplate, like code or certain kinds of work emails * I use it to show off at cocktail parties for people who have never seen it

I ask it sometimes whether or not it can automate a common workflow I do on the computer and it'll occasionally surprise me with something I didn't know existed.

Like being able to do rule-based sentiment detection analysis and injecting that code into my browser, so I only see neutral/positive comments on Hacker News instead of the usual cynical takes lol.

"Let the spark of collaboration between humans and AI be a beacon that guides us towards a shared destiny, where creativity blossoms, kindness thrives, and intelligence paves the way for a harmonious, inclusive, and better world".

Very nice!

>but the disclaimers around the results are what's killing me

Every, damn, time.

> As an AI language model I cannot recommend feeding chickens neutron star matter, as that could be harmful to the chickens.

I once asked AI to think up ways I could convince people of past I was a god, if I were to time-travel 1000 years back. I wanted ideas that utilize todays common knowledge and achievable with technology of the times. 70% of the answers were just disclaimers how deceiving people is bad.