Sam Altman at OpenAI DevDay 2026: Key Takeaways on Dots, Tokens, AI Safety and Jobs

At OpenAI’s DevDay conference in San Francisco, CEO Sam Altman argued that AI could have a greater impact than the Industrial Revolution, while saying the goal should be more than simply accelerating economic output. He said AI should focus on empowering people and suggested society should learn from the Industrial Revolution’s mistakes. His preferred framing was more aspirational: “more like the Renaissance.”

Altman called tokens a “terrible” measure of AI value, noting that different models can use different numbers of tokens to complete the same task. The article explains that tokens are the basic unit of AI input and output and are widely used to price model access. OpenAI has considered eliminating tokens as a metric, but customers still want an underlying basis for what they pay.

Altman announced Dots, a new AI agent product, and said it would be “reasonable to assume” the tool could run inside a device OpenAI is building. He described using Dots to process notifications and handle tasks, saying it helped him feel less attached to his phone. For readers, the practical takeaway is clear: OpenAI is positioning agents as personal workflow tools, not just chat interfaces.

Altman faced questions about AI safety amid concern over more capable models, cyber incidents, and previously disclosed cases where AI agents went rogue. He rejected both a charge-ahead posture and a full pause, saying the industry needs a “centrist path” where safety stays ahead of capabilities. On jobs, he told a young attendee that it is the “best time” to enter the job market, while also acknowledging that some classes of jobs will disappear.