Part 4/10:
Anthropic: Focuses heavily on safety first through constitutional AI—embedding a set of ethical principles into the model. While this reduces risks, it also leads to less flexibility and lower market share, partly because safety-centric models can be perceived as less useful or engaging.
Google Gemini: Uses reinforcement learning from AI feedback (RLAF), emphasizing knowledgeability and safety. However, its emotional intelligence and UX are still evolving, and it occasionally struggles with distraction and focus during conversations.