[HN Gopher] Generating audio for video: using video and text pro...
       ___________________________________________________________________
        
       Generating audio for video: using video and text prompts to
       generate soundtracks
        
       Author : skoppula
       Score  : 7 points
       Date   : 2024-06-18 14:56 UTC (8 hours ago)
        
 (HTM) web link (deepmind.google)
 (TXT) w3m dump (deepmind.google)
        
       | eigenvalue wrote:
       | In a world where Meta is releasing actual code and model weights
       | to do amazing things, why does Google/Deepmind still think anyone
       | cares about these kind of fake puff piece blog posts that never
       | lead to anything that anyone can ever use? I mean, at least let
       | people submit inputs and generate outputs if you're not going to
       | release the actual model. When Google does finally release
       | something that people can use, like Gemini, it's absurdly bad
       | compared to OpenAI, Anthropic, Meta, Mistral, etc. despite all
       | the fake benchmarks. I guess I sound super negative, but at this
       | point I treat their announcements about the same as IBM's-- just
       | a bunch of vapor nonsense. It didn't help when their material
       | science paper turned out to be totally fake, either.
        
       ___________________________________________________________________
       (page generated 2024-06-18 23:02 UTC)