From Hackaday.com, this is a simulation of a neural network featuring a pet squid.

Self-Sufficient Actions:

• The squid operates independently, making choices based on its current condition (hunger, fatigue, etc.).
• Utilizes a vision cone for identifying food, mimicking realistic hunting behavior.
• The neural network is capable of decision-making and forming connections.
• Weights are examined, adjusted, and trained using the Hebbian learning method.
• Experiences from both short-term and long-term memory can impact choices.
• The squid can generate new neurons in reaction to its surroundings (Neurogenesis).

As always, you may use this squid article to discuss the security topics in the headlines that I have not mentioned.