Just 30%? Astonishing. I'd have imagine it was well over 90%.
I suppose that it doesn't have the kind of natural-monopoly thing that Amazon and Google have. There's room for more than one player.
Still, is there any domain in which 70% of proof-of-concepts keep going? There are so many reasons that even valid projects fail. We're still trying to decide if there's any "there" there in generative AI.
Clearly that's something. But it's also clear that there's an awful lot of "ChatGPT for [X]" projects that are destined to just suck, many of which are redundant with each other.
I suppose that it doesn't have the kind of natural-monopoly thing that Amazon and Google have. There's room for more than one player.
Still, is there any domain in which 70% of proof-of-concepts keep going? There are so many reasons that even valid projects fail. We're still trying to decide if there's any "there" there in generative AI.
Clearly that's something. But it's also clear that there's an awful lot of "ChatGPT for [X]" projects that are destined to just suck, many of which are redundant with each other.