One of the hardest edges to build in AI is the data behind it.
Revenue at the top labs is climbing steeply, but the cost of running the models continues to fall.
As those costs fall, the real bottleneck becomes verification. The value moves to the data and training environments where truth is hard and expensive to establish.
This favors fields where reality is hard to label like biology and robotics. The data and environments there are scarce and expensive to build. That scarcity could compound into an advantage over time.
(Original content provided by @PonderingDurian at Delphi Ventures)
Read the full report for free here.
dữ liệu sinh học và robot vẫn khan hiếm nên chi phí xác minh cao hơn hẳn.
