System Two Thinking: longer thinking, much better results

A new AI approach from OpenAI promises significantly better results with longer computation times compared to existing models, reports Michael Nuñez for VentureBeat. OpenAI researcher Noam Brown explained at the TED AI Conference in San Francisco that 20 seconds of AI computation time delivers the same performance boost as scaling up a model 100,000 times. OpenAI’s new o1 model uses “system two thinking” – a slower but more thorough way of processing information. In mathematical Olympiad tasks, o1 already achieved 83 percent accuracy, while its predecessor GPT-4o only reached 13 percent. Brown sees great potential for applications in healthcare, energy, and finance. However, the new model works more slowly and is significantly more expensive at $15 per million input tokens than previous versions.