What is DeepSeek? 3 Facts About The ‘Chinese’ AI
It seems there is no stopping artificial intelligence (AI) from making its way across different industries (even in the entertainment industry, which has caused quite a stir). Now, a Chinese-made AI model is taking over and becoming Apple’s most downloaded app.
What is DeepSeek?
DeepSeek functions in a similar way as ChatGPT. You ask this AI-powered chatbot a question, and it provides you with an answer. However, unlike ChatGPT, DeepSeek is “trained to avoid politically sensitive questions,” per BBC.
Since it’s Chinese-made, DeepSeek’s models are bilingual (Chinese and English). It can understand questions asked in Chinese and answer in the same language.
Who is Behind DeepSeek?
The AI firm is a startup based in Hangzhou. Its controlling shareholder is Liang Wenfeng, co-founder of High-Flyer, a Chinese hedge fund company incorporated in February 2016. The company is the backer and founder, according to Reuters.
Why Is It Getting Popular?
DeepSeek is making waves notably because it’s significantly cheaper to develop. According to Wedbush Securities analyst Dan Ives (as reported by CBS News), developing DeepSeek only costs $6 million, in contrast to developing OpenAI which roughly costs $1 trillion.
With this, tech stocks dropped significantly by the time it surpassed ChatGPT as the most downloaded free app on the iOS App Store in the U.S. Nvidia, the company that produces chips required for AI training, also saw a sharp drop in their stock price. Tech companies that have invested billions of dollars in AI technology became immediately concerned at the prospect of a competitor “outpacing” them, according to NBC News.
In a statement, Nvidia called it an “excellent AI advancement,” a sentiment not shared by OpenAI, the developer of ChatGPT. According to Bloomberg and Financial Times (via CNN), OpenAI and Microsoft claim DeepSeek “used OpenAI’s intellectual property to build its competitor, violating its terms of service.” A spokesperson for OpenAI said that the company is “aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more.”
404 Media published a post pointing out the irony of what OpenAI is accusing its competitor of. OpenAI has been collecting data in an “unauthorized manner” to provide answers to the questions its users have been asking. A group of media publishers, including The New York Times, is battling OpenAI in court, over copyright infringement. Their argument is that “the data that powers ChatGPT has included millions of copyrighted works from the news organizations, articles that the publications argue were used without consent or payment,” according to NPR. OpenAI is built on the premise that the more data it’s able to collect, the better it can provide answers. DeepSeek, on the other hand, is using a “reinforcement learning strategy,” meaning there “is the potential of LLMs to develop reasoning capabilities without any supervised data, focusing on their self-evolution through a pure reinforcement learning process.”
It’s ironic to see OpenAI complaining about someone else doing the same thing—only doing it better.
