Ideas are in:
- AI: in a crisis of science
- Non-computing machinery and intelligence,
- Learning Chinese
- Artificial general intelligence
Starting point: everything AI thinks it knows about intelligence is wrong.
The front end of intelligence is the sensor. If not adequately understood, much or all downstream understanding is uncertain.
Sensors are impacted by environmental particles. In response the sensors transmit streams of data units through the inner world to the central system. These streams contain knowledge in some form. So we ask: What is the form?
The primary sensor process is transduction. External particles don’t enter the inner world, and transduction doesn’t duplicate values of their intrinsic properties, hence neither the particles nor their properties survive transduction. But if an inner object is to be about an external object or object type, that is, if it is to have semantic content, something must survive transduction. So what is that thing?