At Dreamforce, Business Leaders Say Last Year's AI Models Are Good Enough
TL;DR: At Salesforce's Dreamforce, business leaders reportedly said last year's models already handle most of their practical needs, CNBC writes — a reality check on the frontier-model race.
Key points
- Per CNBC, many enterprise practitioners said last year's models are "good enough" for their workloads.
- The frontier debate over model capability and safety diverged from what buyers actually deploy.
- Attention appears to be shifting from "the most powerful model" toward "the model that fits the job at the right cost."
Why it matters
Even as capability races continue, enterprise buying favors stability, cost, and ease of integration over sheer novelty. The AI safety conversation is moving from abstract risk toward a practical bar: is this safe enough to ship now?