Within one week of its debut, the Chinese AI platform DeepSeek overtook OpenAI’s ChatGPT in the US Apple App Store download rankings, according to The Spectator Index. This poses a serious challenge for well-established AI leaders like OpenAI, Google, and Microsoft.
The DeepSeek R1 is a sophisticated large language model created by the Hangzhou-based startup DeepSeek to enhance analytical and reasoning skills. The model uses a hybrid architecture that blends chain-of-thought reasoning with reinforcement learning. There are two versions of DeepSeek: DeepSeek-R1 and DeepSeek-R1-Zero. The latter improves its reasoning ability through unsupervised fine-tuning.
Also Read: Ati Motors, A Robotics Startup, Raises $20 Million Funds
Why is DeepSeek becoming so well-liked?
Over the weekend, conversations contrasting DeepSeek R1 with other AI models generated a lot of attention on social media. Its seamless performance on smartphones and capacity to do challenging tasks, like sophisticated mathematical computations, have been praised by users. Furthermore, DeepSeek has made its R1 model compatible with laptops and other devices by introducing smaller model variants.
Budget-Friendly Performance:
- The performance of DeepSeek-V3 is on par with that of other top AI models.
- With a far lower budget, it was created to overcome China’s restrictions on cutting-edge hardware.
Efficiency in Cost and Innovation:
- DeepSeek-V3 demonstrates the possibility of producing formidable AI with little funding, whereas most competing AI systems require billions of dollars.
- Its success highlights the value of innovative and resourceful strategies.
Accessibility of Open-Source:
- Because DeepSeek-V3 is open-source, developers can use it anywhere in the world.
- It is especially beneficial for areas where funding for pricey AI tools is scarce.
Also Read: Perplexity CEO willing to devote his time and $1 million to “making India great again” with AI
According to a story in The New York Times, DeepSeek’s developers have produced outstanding outcomes utilizing significantly less specialized computer chips—necessary for AI training—than their American counterparts. These chips are essential to the U.S.-China technology competition, in which the United States imposes export controls to maintain its dominance in artificial intelligence research. DeepSeek’s accomplishments, however, show how these limitations may have the unintended consequence of encouraging Chinese researchers to develop with more readily available but constrained resources.
Here’s how to get in: DeepSeek
You can access DeepSeek R1 using its chat interface at chat.deepseek.com. Using API keys, developers may quickly and easily incorporate the model into their applications using the DeepSeek Developer Portal.
Also Read: What is Trump’s $500 billion AI project, Stargate?
