DeepSeek hit with large-scale cyberattack, says it’s limiting registrations

Deepseek said Monday that it would temporarily limit user records “due to large -scale malicious attacks” in their services, although existing users may log in as usual.

The Chinese artificial intelligence startup has generated a lot of stir in recent weeks as a rapid growth rival for OpenAi Chatgpt, Google Gemini and other leading artificial intelligence tools.

Earlier on Monday, Depseek took the coveted place of his OpenAi rival as the most unloaded free application in the USA. In Apple’s App Store, dethlting Chatgpt for Deepseek’s own assistant. It helped inspire an important liquidation of world technological actions.

Rumors about the company, which was founded in 2023 and launched its R1 model last week, have spread to analysts, investors and technology developers, who say that exaggeration, and the consequent fear of being left behind in the always changing cycle Of the exaggeration of AI, it can be justified. Especially in the era of the AI ​​generative arms race, where both technological giants and new companies are competing to make sure they are not behind in a market that is expected to exceed billion dollars in income within a decade.

As reported, Depseek emerged from the Artificial Intelligence Research Unit of a Chinese coverage fund in April 2023 to focus on large language models and achieve general artificial intelligence, or AGI, a branch of artificial intelligence that matches or It surpasses the human intellect in a wide range of tasks, that Openai and his rivals say they are chasing him quickly.

Rumors around Deepseek began to spread especially last week, when the startup launched R1, its reasoning model that rivals OPENAI O1. It is open source, which means that any IA developer can use it, and has shot at the top of the applications and industry classification tables, and users praise their performance and reasoning capabilities.

Startup models were significantly built despite the fact that the United States stopped chips exports to China three times in three years. Estimates differ on how much exactly costs Deepseek or how many GPUs are included in it. Jefferies analysts estimated that a recent version had a “training cost of only 5.6 million dollars (assuming a rental cost of 2 dollars per hour and 800 hours). That is less than 10% of the cost of the goal flame. ”

But regardless of the specific figures, the reports agree that Openai, Anthropic, Google and others developed the model at a fraction of the cost of rival models.

As a result, the AI ​​sector is flooded with questions, among them if the growing number of astronomical financing rounds and valuations of billions of dollars in the industry is necessary, and if a bubble is about to explode.



Source link

Leave a Reply

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