Reaction
AI data vendors' staying power through the lens of early ad networks
A post argues that fast-growing AI data suppliers, like early ad networks, may struggle to become harder to replace.
TLDR
A post compares AI data vendors with early ad networks, arguing that rapid revenue growth alone may not build durable value if suppliers share experts and labs own the output. It proposes three ways to become harder to replace: own differentiated data supply, become part of labs' workflows, and build training environments that improve using model feedback. The author predicts that vertical specialists will replace large horizontal providers.
Combined views
—
2 Sources, first seen ago
