It's very difficult to see what's the thin layer on top of what IBM is doing, Google is doing, AWS is doing, a thin layer on something that's pretty commoditized, on these large language models-and what's truly valuable.
When anyone can prompt, then what's proprietary to you? And how do you go about collecting, amassing, and getting results from that data? A lot of the investing we see are just thin layers on top.
Instead, he suggested that investors value a startup's ability to continually engage customers. Continued engagement means more data is collected, which in turns builds more value for an AI application.
Collection
[
|
...
]