Conscious agents / the 2014 model
A world,
through you.
What happens between
the world and an experience?
Based on Objects of consciousness
Donald Hoffman & Chetan Prakash · 2014
A world becomes an experience.
Four possible world states. Two possible experiences. Change the mapping, and the same world can be experienced differently.
The probability of receiving the experience assigned by the selected mapping. At 50%, both experiences are equally likely.
Each passage is a possibility.
Ready for perception / press Step to follow one transition
02 / Beneath the surface
The probabilities
are the mechanism.
These are the actual matrices driving the field. Each row totals 100%. The highlighted cell is the last sampled transition.
Adjust a channel above to change its matrix.
perception
W → X| w ↓ / x → | x₀ | x₁ |
|---|---|---|
| w₀ | 90% | 10% |
| w₁ | 90% | 10% |
| w₂ | 10% | 90% |
| w₃ | 10% | 90% |
decision
X → G| x ↓ / g → | g₀ | g₁ |
|---|---|---|
| x₀ | 80% | 20% |
| x₁ | 20% | 80% |
action
G → W| g ↓ / w → | w₀ | w₁ | w₂ | w₃ |
|---|---|---|---|---|
| g₀ | 45% | 45% | 5% | 5% |
| g₁ | 5% | 5% | 45% | 45% |
03 / A note on the model
An invitation
to look closer.
In Objects of consciousness (2014), Donald D. Hoffman and Chetan Prakash define a conscious agent using spaces of experiences and actions, three probabilistic channels, and a counter.
This study explores a finite example of those channels. Its four world states, two experiences, two actions, and adjustable probabilities are illustrative choices. Colors, particles, and spatial forms are artistic encodings. The display separates P, D and A into teaching stages; Cycles counts completed loops, each using every channel once. It is not a physical clock. Artwork mode runs a fresh sample with the default probabilities; this study lets you change them.
The animation reveals a model’s behavior. It does not establish consciousness in the simulation or demonstrate that conscious realism is true. The world W remains abstract here; the paper goes further into networks of interacting agents.
Read the original paper Hoffman & Prakash · Frontiers in Psychology · 2014