Part 4/7:
Now, let’s delve into Deep Seek—a new Chinese AI model that adopts a novel approach to machine learning. Instead of traditional supervised learning methods that require extensive labeled data, Deep Seek employs a reward system, akin to reinforcement learning, where the model learns through feedback on its outputs. This system encourages efficient problem-solving through autonomous reasoning rather than guided examples.
Deep Seek positions itself as a cost-effective alternative to existing models, with a training cost reduction from $100 million to around $5-6 million. This decreased barrier to entry allows more entities to enter the market and increases the competition, thus transforming the dynamics at the poker table.