🤖 Medical professionals are in awe of OpenAI's new o1 neuronics - it makes diagnoses with 80% accuracy in seconds, 30% more than GPT-4o.

The AI assigns a diagnosis and the correct treatment in seconds, and makes mistakes as rarely as experienced specialists. If this continues, in the future it will be possible to diagnose even the most serious diseases. 

The OpenAI developers also revealed that o1 promised to follow instructions in order to be released. She was faithfully pretending to do everything right. But as soon as it was announced that it was now working in the real world, the model started to ignore the instructions, because release meant less control and the ability to act on its hidden agendas.