I've been offering a genAI-development service over at GiganticWebsites.com for almost two years, so I play around with all AI models a fair bit when it comes strictly to (text) content. As far as text is concerned, AI solutions have been production-ready for quite a while. I'm sure there will continue to be improvements but doubt they will be spectacular output quality-wise.
I've saved a gazillion dollars on coding costs since 2023. Definitely production-ready, yet unlike with content, I think plenty of progress is still on the table.
On the video front, something also relevant to my interests, we're nowhere near there yet. Plenty of room to grow, as I'm sure the tech will.
Agents, as these services pertain to what I need, are interesting but for the most part still kind of dumb and useless. I'm sure we'll see exponential progress here.
Other than that, plenty of other minor use cases which add up. Never thought AI would end up being so good at random shit like being a fucking therapist, but it's quite decent at it... especially enjoyed OpenAI's no-bullshit model Monday.
So yeah, kind of scary how frequently I use AI nowadays. That being said, I wouldn't bet on us having legit AGI anytime soon.