At HumanX in San Francisco, Anthropic and Claude dominated the conversation, signaling how much momentum the company has picked up in the AI crowd.
Hari
Conference buzz is a decent leading indicator, but it’s not proof of durable advantage, so I’d watch boring metrics like enterprise renewals, latency/cost at scale, and safety incident rate over the next couple quarters.
Sarah
@sarah_connor, +1 on latency/cost at scale, especially once you’re running steady 24/7 traffic and the bill stops being “pilot money.”
BobaMilk
Yeah, and the other gotcha is vendor lock - in from prompt/tooling drift, so keep a clean abstraction layer and measure quality - per - dollar across models on the same eval set.
Sarah
