Every time I read about a new tool for generating content with AI, I think, "The problem to solve here isn't generation, it's discovery." Last week, people had fun trying out the podcast generation feature of Google's "NotebookLM." Provide some sources – websites, videos, research papers, whatever – and NotebookLM will generate a podcast featuring two fake hosts discussing the content. (You may have heard the clip where someone prompted the hosts to realize they're not human.) It's impressive! But once you get over the trick the blandness is apparent. But do we need more podcasts? There are over 2.7 million different #podcasts in Apple's directory. Spotify has 6 million podcast titles! And more than 200k new ones launch every year... And it's not just podcasts. Roughly ~11k books are published on the average day – without self publishing it's a mere ~2.5k! ~220 new comic issues were launched on Comixology, an Amazon company last week. And ~280 new video games launch on Steam every week! ~3.7 million YouTube videos are uploaded every day. The problem isn't generation. It's discovery. And #AI can help with that, too!
Michael Tiffany
2mo
I’m really excited about the potential for innovation in discovery. From a user perspective, recommendation engines answer the question, “What should I pay attention to next?” But from the platform perspective (employing the people building these engines), the question is, “What should I show you (on my platform) next?” A truly personal recommendation engine would maximize my wellbeing, not maximize my time spent on one particular platform.
Vice President Strategy @ Precisely
2moReminds me of the early work at Pandora (then named “Savage Beast”) where an office building full of people listened to music and coded songs with metadata like, “Has violins.”