DeepSeek suffered a massive cyberattack and said it was restricting registrations


DeepSeek said on Monday it would temporarily restrict user sign-ups due to a “large-scale malicious attack” on its service, but existing users will be able to log in as normal.

This Chinese artificial intelligence startup has caused a great sensation In recent weeks, it has emerged as a fast competitor to OpenAI’s ChatGPT, Google’s Gemini and other leading artificial intelligence tools.

Earlier on Monday, DeepSeek overtook rival OpenAI as the most downloaded free app in the United States. appleApp Store, replacing ChatGPT with DeepSeek’s own AI assistant. it helped inspire sharp sell off in global technology stocks.

The buzz about the company, which was founded in 2023 and released its R1 model last week, has spread to tech analysts, investors and developers, who say the hype, and the ensuing concerns, are constantly changing. Concerns about falling behind in the AI ​​hype cycle could lead to the company’s failure. It’s guaranteed. Especially in the age of the generative AI arms race, where tech giants and startups alike are racing to ensure they don’t fall behind in the market. Expected to exceed $1 trillion Income over ten years.

Read more CNBC coverage of artificial intelligence

DeepSeek was reportedly developed in April 2023 from the artificial intelligence research arm of a Chinese hedge fund to focus on large language models and achieve artificial general intelligence (AGI) – a branch of artificial intelligence that is equal to AGI on a wide range of tasks. or surpass human intelligence, OpenAI and its rivals say they are quickly catching up.

Conversation around DeepSeek in particular began to spread last week, when the startup released R1, an inference model comparable to OpenAI’s o1. It’s open source, meaning any AI developer can use it, and it’s already soared to the top of app stores and app stores. Industry rankingsusers praise its performance and reasoning capabilities.

Although the United States Curbing chip exports Came to China three times in three years. Estimates vary on how much DeepSeek R1 will cost or how many GPUs it will use. Jefferies analysts estimate that the latest version’s “training cost is only $5.6 million (assuming a rental cost of $2/H800 hours). That’s less than 10% of the cost.” Yuanof flame. ”

But regardless of the specific number, the report agrees that the development cost of the model was only OpenAI, Anthropic, Google and others.

As a result, the AI ​​industry is rife with questions, including whether the industry’s ever-increasing astronomical funding rounds and billion-dollar valuations are necessary, and whether the bubble is about to burst.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *