Tag: OpenAI
ChatGPT Finally Works While You Sleep & More AI News You Can Use
The Signal
ChatGPT has undergone a significant functional expansion, most notably by enabling its Gmail connector to send outbound emails and adding a native, recurring task scheduler. While the creator emphasizes these as breakthrough automation tools, their reliability remains a point of contention alongside unsettled claims regarding why certain high-powered model tiers like Fable were suddenly deactivated. This update marks a shift from reactive text generation toward proactive, scheduled agentic workflows.
The Case
- Users can now configure recurring tasks directly within the ChatGPT sidebar, allowing for automated daily briefings or periodic inbox scans that run on set cadences.
- The Gmail connector has moved beyond drafting to support outbound email sending, though the creator notes these connectors remain operationally inconsistent—sometimes failing as often as 20% of the time.
- OpenAI has overhauled its model picker to emphasize 'thinking levels'—ranging from instant to pro extended—to give users more granular control over how much compute and reasoning logic is applied to a query.
- Dynamic document handling has replaced static modes, with Canvas now appearing automatically for long-form content and providing a library feature for saving and later retrieving essays.
- A Chinese open-source model, GLM 5.2, is being highlighted for its 1 million token window and reported agentic coding performance, though the creator's comparison to industry leaders like Opus lacks independent, third-party validation.
- Access to the 'Fable' model tier was abruptly pulled shortly after public demonstration; the creator attributes this to government limitations on model deployment, though this causality is not officially verified.
The 1 Minute Signal Take
These updates collectively signal that the core utility of consumer AI is shifting from a 'chat' interface to an 'automation' layer, which is a welcome increase in practical value. While the new scheduling and outbound email capabilities are genuinely disruptive, the creator's unverified claims regarding model performance and the reasons for specific model takedowns require a skeptical eye. Watch this if you want a clear walkthrough of the new UI features and the specific workflows for scheduling tasks, but skip it if you are primarily seeking reliable benchmarks or verified industry news.
Time saved:
Tags
Tag: OpenAI
