让 AI 接住杂事，把注意力还给人：Sam Altman 的 Dots 使用体验与平台下注

Altman argues that greater automation can reveal latent demand and raise expectations, while acknowledging that some less creative tasks may still be displaced by AI.

OpenAI's platform ambition is to enable a wider ecosystem of small companies, but Altman explicitly says that the company has not yet found the equivalent of a successful App Store.

A persistent agent can protect focused time by triaging incoming information and escalating urgent issues, although detecting a problem is not the same as being able to resolve it.

Connecting context across messages, calendars, and images can help an agent recover information that a single search misses, but the examples remain personal accounts rather than a reliability benchmark.

Continuous feedback and low-friction note capture can turn an uncertain idea into several prototypes, allowing a busy person to iterate in small moments instead of first writing a complete specification.

Live documents can combine contributions from people and their agents while updating with changing context, but personalized summaries still need a shared basis that others can inspect.

The Decisions API reflects demand for faster and cheaper common tasks, while Altman leaves room for new architectures that OpenAI cannot cover across the entire cost-and-speed frontier.

Portable subscriptions may reduce the inference-cost burden on application developers and make experimentation easier for users, but access to a platform does not itself establish distribution or sustainable demand.

Making something people want remains a startup invariant, but the amount of work one or two founders can accomplish with agents calls older staffing assumptions into question.

Fast responses can tighten the creative feedback loop, yet Altman expects premium intelligence and speed to remain at the frontier even as the value available per dollar improves.