Ah so surpassing Gemini 1.5 Pro and all other Models on Vision understanding by 5-10 points is "not ground breaking" all while doing it at insane latency.
Jesus if this shit doesn't make you coffee, and make 0 mistakes no ones happy anymore LOL.
the only thing you should be celebrating is that its 50% cheaper and twice as quick at generating text but virtually no real ground breaking leaps and bounds to those studying this space carefully.
basically its chat gpt3.9 at 50% of chatgpt4 prices
Cool so ... just ignore the test results and say bullshit lol It's not GPT3.9 many have already said its better than GPT4 turbo, its better than Gemini 1.5 Pro and Opus on Vision recognition. but sure... the price difference is whats new lol
> virtually no real ground breaking leaps and bounds to those studying this space carefully
What they showed is enough to replace voice acting as a profession, this is the most revolutionary thing in AI the past year. Everything else is at the "fun toy but not good enough to replace humans in the field" stage, but this is there.
Between this and Eleven Labs demoing their song model, literally doing full on rap battles with articulate words people are seriously slacking on what these models are now capable of for the voice acting/music and overall "art" areas of the market.
Jesus if this shit doesn't make you coffee, and make 0 mistakes no ones happy anymore LOL.