At Dreamforce, business leaders say older AI models suffice

At Salesforce's Dreamforce conference in San Francisco, the gap between enterprise reality and industry rhetoric came into focus. According to CNBC, many attendees said they get sufficient value from older AI models, suggesting that the breakneck release cycle of frontier labs may not match what businesses actually deploy or need. The sentiment was consistent across conversations, with no indication that companies are clamouring for the latest versions.

Meanwhile, WIRED reported that the same conference became an unlikely battleground for the CEOs of OpenAI, Anthropic, and Nvidia, who debated whether AI development should slow down. That debate — about safety, capability, and pace — unfolded on stage while rank-and-file business users expressed contentment with the status quo. The two sources agree on this tension: the loudest arguments about AI's future are happening far above the quieter, practical satisfaction of its current users.

Where the sources differ is in emphasis. CNBC frames the business view as a matter of diminishing returns on model upgrades, while WIRED highlights the philosophical split among industry leaders about how fast to push. Neither source suggests the slowdown debate is directly influencing buying decisions, but both point to a growing disconnect between what AI vendors sell and what enterprises say they need.