What is DeepSeek : DeepSeek, a Chinese artificial intelligence start-up competing with OpenAI’s ChatGPT, Google’s Gemini, and Meta’s AI, has triggered a staggering $US593 billion ($943 billion) loss on Wall Street.
Last week, the company introduced its AI model, R1, which offers many of the same features as its competitors but at a significantly lower cost.
As a result, global stocks plummeted, leading to Wall Street’s largest single-day decline in history.
But what is DeepSeek, the company that has tech giants on edge?
What is DeepSeek AI?
DeepSeek was founded in 2023 by CEO Liang Wenfeng in Hangzhou, China.
Wenfeng previously co-founded the hedge fund and AI company High-Flyer, which acquired 10,000 of Nvidia’s A100 graphics processor chips—crucial for developing and running AI systems—in 2022.
In just over a year since its launch, DeepSeek has introduced a range of competitive AI models.
Its latest model, R1, was announced late last year and officially launched as an AI assistant last week.
Following its debut, DeepSeek’s mobile app quickly soared to the top of app charts, surpassing ChatGPT as the most downloaded free app, with nearly two million downloads so far.
Renowned tech investor Marc Andreessen—who has advised former US President Donald Trump on policy and warned that excessive AI regulation in the US could give China a competitive edge—described DeepSeek R1 as “AI’s Sputnik moment.”
“DeepSeek R1 is one of the most astonishing and groundbreaking advancements I’ve ever witnessed—and as open-source, a remarkable gift to the world,” he shared in a post on X.
What makes DeepSeek different from ChatGPT and other AI?
One of the key highlights is the significantly lower cost of running the model.
According to DeepSeek, the R1 API is 20 to 50 times more affordable than OpenAI’s O1 model.
DeepSeek-R1’s API is priced at just $0.88 per million input tokens and $3.49 per million output tokens, while OpenAI’s API costs $23.92 and $95.70, respectively.
The company also claims to spend just $8.9 million on powering its AI, a stark contrast to the hundreds of millions, or even billions, spent by competing US firms on their AI technologies.
Additionally, DeepSeek has positioned its models as “open source,” meaning key components are freely available for anyone to access and modify.
However, the company has not disclosed the data it used for training the model.
What are DeepSeek’s main features?
DeepSeek R1 features an interface similar to ChatGPT and other AI models, allowing users to input questions and receive AI-generated responses.
What sets it apart from other AI models is its ability to articulate its reasoning and demonstrate its thought process when answering questions. This self-explanatory reasoning is then used for further training, eliminating the need for the company to feed the model new information—enabling it to learn independently.
Read More :- Apple News Sends Crowds Wild As New Products Launched
Additionally, the model continues to learn and improve through the feedback it receives.
Its multi-stage learning process and cold-start data have enhanced its performance, bringing it “on par” with OpenAI’s O1 model.
Leave a Reply