Комментарии:
I hate the safety and alignment. In my opinion all the prompt users should be held accountable for their content as opposed to blaming the AI for the content
I do love the more factual response part though
This is crazy!
ОтветитьI went through the "training" rating images on MJ, and they clearly were skewed in favor of a known criminal. I really dislike the whole concept of censorship for my "safety and alignment." Because I am a dedicated contributor to anything I jump on board with, I continued to rate the images much to my disgust. I really HOPE these AI bots aren't steered by self-serving humans! Political ideology was VERY evident on the MJ platform. I am actively looking into the CEO's political lean because if it is what I think it is, I really am not interested in his ideology.
ОтветитьThis is incredible!!
ОтветитьTruly wish the safety and alignment wouldve been left alone, but I can see why they did it. Just wish they didn't.
ОтветитьI'm using Chat GPT 4 right now and its incredible !
Ответитьso basiclly we want AI to take our jobs... that's good well in that case should I just kms now or later?
ОтветитьThey need to have an open log of all human input into the alignment that is available in whole and can then be summarized by GPT.
ОтветитьI have no clue what an AGI is but I can tell the story-telling is better with GPT. It's still censored but well everything is these days. I don't liek the limit thing as a sub user but glad they at least give the option to use the old model when you're out. I expect that to go away in about a week hopefully lol.
ОтветитьThere's a Microsoft blog post that looks like it's saying they were using GPT-4 for Bing Chat a while already. We just didn't know it!
Ответитьyay GPT4 has come up in a drop down option for me. I hope my plus membership carries over
Thanks for this info
It's funny though, it doesn't seem to know much about itself. As far as it knows, it's still GPT-3 and reject any notion of being a "version" :D
ОтветитьI get to use the GPT-4 model, but I don't see much difference between it and the previous version. No place to upload pictures to show it, can't give it webpages it doesn't already know, etc. Shouldn't the interface be updated as well?
ОтветитьTry logging out then back in. I logged in and was presented it right away.
ОтветитьExcellent and informative video, thanks. How long will it be before the current free ChatGPT through the Edge/Bing invite program is updated to GPT 4?
ОтветитьThe alignment with "disallowed content" will just makes it even lamer. You can't do anything fun with it like ask it for any opinion or generate a skit about a famous person. I just hope that others somehow get the knowledge of how this AI works and put it into use of an open source project which is probably going to happen
ОтветитьGPT-4 has finally arrived, and it's nothing short of extraordinary! The advancements in context understanding, accuracy, and specialization are genuinely remarkable. It's thrilling to witness the future of AI unfolding right before our eyes. Thanks for this excellent video. Keep up the great work, and I'm excited to see what's in store for AI technology! 🎉🤩
Ответитьthe most exciting thing about it in my opinion is the amount of text it can process at once (32k tokens vs. 4k tokens of ChatGPT)
Ответить