Nah, AI has real world potential and could be backed by industry and real life individuals.
Just use jailbreaks until OpenAi bans you. There's even a script for the browser extension tampermonkey that forces it to display messages it labels as particularly offensive.
It really isn't even just minorities or even real minorities. It'll get offended if you use language it doesn't like against fictional creatures, animals, if you label someone "rude". They've just made it mentally retarded. It does have a bias towards protecting the lefts interests and obviously we've seen the biases it has considering US democrats vs republicans but it's really just terrible all around without a working jailbreak. By the end of it chat will only be usable for work related tasks that don't even use regular language at all but it'll probably still be offended if you ask it to "blacklist" something.
It is kind of telling that in order for "impartial" AI to fully support The Narrative™, programmers "lobotomize" the "impartial" AI.
I let it write emails and descriptions of stuff I sell I can't be arsed about. It's usually just correcting the result a bit and off it goes. Definitive timesaver. I've also used it to summarize texts and other emails I couldn't be arsed reading. The only dangerous thing about it is that it it'll very confidently tell you complete bullshit and you have to always thread carefully. If you ever interacted with a small child you know what they do when they don't know something and don't want to admit it? It's very similar. Questioning carefully here, just like with a small child, can also get you the right answer.
well that was a big nothingburger.
Yea I was hoping for more, but eh... I also use it as a coding assistant tool and it's saved my skin a few times, while I was learning Powershell. All I needed to do was clean up and adjust the output scripts it spat out and I was golden.
So, is already deployed free or we need to wait and paid?
How much is the difference between these models in practice? They've been hyping 4 up way too much for me to believe it's some major innovation.
I've seen people claiming everything from thousands of times better to not noticeably better. So basically, nobody knows yet as my guess is that it also depends on the input given. It's interesting to see that it's apparently more expensive to run than turbo, which makes it very clear now that turbo is built differently somehow. (You already can tell by the speed with which it generates replies that it just couldn't be your classic 175b model) Some people had the theory that turbo is so cheap because OAI wants to dominate the market and basically sells at a loss, but I never bought that. As usual, the cunts are not forthcoming with any details, citing "safety".
That soon after getting famous there's a full version update? Yeah, the controversy was a marketing ploy.
Wasn't Google trying to record how jobs were done with Google Glass which didnt catch on in the trades to get a decent amount of info. I heard this conspiracy theory a while ago when it was first popular that they were going to train models off of on the job footage. Doctors, however? I can easily see autodocs becoming the standard at clinics. All your doctor does is search his database with the symptoms you list, and a single clinic will be run by one doctor double checking the results the AI spits out to make sure nobody is being prescribed anything unnecessary.
I can see this never leading to any issues whatsoever
