You're viewing part of a thread.
For more conversational things that require more up-to-date knowledge and things that I need to verify the facts/sources for, I'll use bing ai. It's actually quite good for that.
I like how it references it's sources.
Though I've pretty much replaced everything but image searches with it, it parses the data really well and gets the most relevant stuff.
Plus they just hooked it up to DALLE2 so you can feed it images and ask for it to take it as inspiration for other works.
I fed a couple of Simon stalenhag's works and asked it to generate a new cover for No Man's Sky (Stalenhag made the OG one) based on his style but with oil painting. This was one of the outputs.
It's really cool how fast it's come around to be integrated with the rest of the environment. Can't wait to see what they do next with it. I have a feeling Microsoft and Google will make commercial offerings for personal AI models that will be trained on company data and you'll be able to use it without worrying if you're exfiltrating data to OpenAI by using it.
Oh man that's such a cool image! And yeah, its super cool what they are doing. I'm excited for DALLE2 to improve to the level of midjourney haha, but even now its pretty great!
And yeah I agree with you on the direction its going. Specialized models are definitely where its going to go.
I haven't seen much of mid-journey, is it pretty photorealistic?
Oh dude, its really freakin good. Its like, 2 generations past DALLE2 at this point. If you are interested in AI art stuff, I recommend checking it out!