Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

No, DeepSeek Is Not A ‘Sputnik Moment’

Jan 30, 2025 - forbes.com
DeepSeek, a Chinese AI company, has stirred the AI community by developing cost-efficient AI models, claiming to have trained a large language model with just $5.6 million in compute costs. However, this figure is misleading as it only accounts for one training run and excludes prior research costs. Critics argue that the market's reaction is exaggerated and politically driven, noting that cost-efficient AI model development is not new. DeepSeek's approach involved building on open-source models like Meta’s Llama, and while it has sparked discussions on cost efficiency in AI, it lacks a novel research breakthrough.

Despite the skepticism, DeepSeek's use of reinforcement learning and making its technology freely available is significant. The company has prompted a broader conversation about resource allocation in AI development, challenging the notion that vast resources are necessary for building advanced models. This has led to tensions with OpenAI, which accused DeepSeek of using its proprietary model outputs without permission. The debate highlights ongoing issues around data use and intellectual property in AI, with some industry leaders dismissing the situation as overhyped and not a groundbreaking moment for AI.

Key takeaways:

  • DeepSeek's AI model, which is cheaper to produce, has caused a stir in the AI community, but its significance is considered overblown by some industry leaders.
  • The reported $5.6 million training cost for DeepSeek's model is misleading as it does not account for the costs of previous research and multiple training runs.
  • DeepSeek's approach has sparked a debate on cost efficiency in AI model training, challenging the notion that vast resources are necessary for developing advanced AI models.
  • OpenAI has accused DeepSeek of using its proprietary model outputs to create AI systems, highlighting tensions in the AI industry over data usage and model development practices.
View Full Article

Comments (0)

Be the first to comment!