Breaking News

Here’s what you should know


On Monday, Chinese artificial intelligence Startup Deepseek has taken over the wounded place of Openi rivals as the most interesting free apps in the US on on AppleApp Store, Dethroning Chatgpt for AI Assistant Deepseek. Global technological shares are sold out and they have been in a step Wipe billions in a market cap.

Later on Monday Deepsek Concrete user registrations temporarily “Due to the great malicious attacks” of his services, although existing users will be able to apply as usual.

Technology leaders, analysts, investors and developers say Hype-I who follow the fear of lagging in a constantly changing AI hype cycle-can be justified. Especially during the time of the AI ​​weapons generative race, where technological giants and startups are equally racing to ensure that they are not lagging behind the market foreseen to be made of $ 1 trillion in revenue within a decade.

What is Deepseek?

Deepseek was founded by Liang Weenfeng in 2023, co-founder of high-flyer, a quantitative hedge fund focused on AI. Startup AI allegedly grew from the Hedge Fund’s research unit in April 2023 to focus on large linguistic models and reach artificial general intelligence, or AGI – a branch of AI that is equal to or surpasses the human intellect on a wide range of tasks, which is Openi, Openai, What openai and his rivals say that he is quickly haunted. Deepseek is still fully owned and funded by a high flyer, according to Jefferies’ analysts.

The sound around Deepseek started picking up Steam earlier this month, when the startup published R1, his reasoning model that rival Openai’s O1. It is an open boiler, which means that any AI programmer can use and rocket it to the top of the app stores and boards of the leading industrysince users praise his ability to perform and reason.

Like other Chinese Chatboti, it has its limitations to a question about certain topics: when asked about some of the Chinese leader XI Jinping, for example, Deepsek saves user From similar test lines.

Another key part of the discussion: Deepseek’s R1 was built despite the US Publishing chip exports to China three times in three years. The estimates differ in how much Deepseek’s costs of R1 or how much the GPU in it has entered it. Jefferies’ analysts estimated that the recent version had “training cost of only $ 5.6 million (assuming that 2/800 hours of lease costs). This is less than 10% cost TargetLlam. “” But regardless of specific numbers, reports agree that the model has been developed in a fraction of the costs of the Openi rival models, anthropically, Google and others.

As a result, the AI ​​sector is full of questions, including whether an increasing number of astronomical financing rounds and a billion dollars are needed and whether the bubble will crack.

Read more CNBC reports on AI

Stocks Nvidia It fell almost 16%on Monday, with a chip manufacturer Asml down almost 7%. Nasdaq fell more than 3%. Four technological giants – Target,, Microsoft,, Apple And ASML will be ready to report to the earnings this week.

Analysts in Raymond James described in detail some questions that plague the AI ​​industry this month, writing: “What are the implications of investment? What is the open sources compared to the ownership models? Is the throwing of money on the GPU truly Panacea? The limit of American export work? [DeepSeek]?? Well, they might be terrible, or not guessed, but be sure, the industry is awful with disbelief and guessing. “

Bernstein’s analysts on Monday wrote in a note that “according to many (occasionally hysterical) heat we have seen [over the weekend,] Implications move anywhere from “It’s really interesting” to “This is the death of the knee and the infrastructure complex we know.”

The way US companies respond

Some American executive directors advocate to answer before clients turn to potentially cheaper Deepseek offers, and the target allegedly started with four deep ties. “War rooms“Within its generative department AI.

Microsoft CEO of Satya Nadella wrote On X that the phenomenon of Deepsek was just an example of Jevons’s paradox, writing: “As Ai becomes more effective and more affordable, we will see that its use is jumping, turning it into goods that we simply cannot get enough.” The Openi CEO of Sam Altman tweed a quote attributing to Napoleon, writing: “The revolution cannot be made or stopped. The only thing that can be done is that one of a few children give it to a victory.”

Yann Lecun, Meta Chief Ai Scientist, wrote On LinkedIn, this Deepseek’s success indicates a change of tide in the AI ​​sector to favored the technology of open code.

Lecun wrote that Deepseek profited from some of the targets of his own technology, ie Llam models, and that the startup “came up with new ideas and built them on top of other people. Since their work was published and an open source, everyone can profit from that .

Alexandr Wang, Executive Director of Scale AI, said CNBC Last week, that last Ai model Deepseek was “breaking the earth” and that its edition of R1 was even more powerful.

“What we have discovered is that Deepsek is … top -notch performance or about the same with the best American models,” Wang said, adding that the race AI is between the US -Ai China “AI War”. Wang -s company provides a key to key AI players, including Openi, Google and Target.

Earlier this week, president Donald Trump announced a Common investment with Open,, Prophet and Softbank To invest billion dollars in US AI infrastructure. The project, Stargate, was presented at the White House by Trump, the CEO of SoftBank Masayoshi Son, co-founder of Oracle Larry Ellison and Openi CEO of Sam Altman. Will include key initial technological partners Microsoft,, Nvidia and Oracle as well as a semiconductor company Hand. They said they would invest $ 100 billion to start and up to $ 500 billion in the next four years.

Ai evolutionary

The news of Deepseek’s skill also come in the middle of a growing hype around AI agents – models that go beyond chatbots to complete complex tasks with multiple steps for users – which chase both technological giants and startups. Meta, Google, Amazon, Microsoft, Openi and Anthropic expressed their goal of building an agent AI.

Anthropically, Amazon-Bocked AI startup found by former research executives have increased their technology development over the last year, and in October startup said That his AI agents could use computers like people to complete complex tasks. The ability to use the ANTHROPICA computer allows its technology to interpret what is on the computer screen, select buttons, enter the text, move on to websites and perform tasks through any software and internet review in real time, is stated.

The tool can “use computers basically the same way as we do,” said Jared Kaplan, Chief Scientific Director of Anthropic, for CNBC in the then interview. He said he could do this with “dozens or even hundreds of steps”.

Openai posted a a similar tool Last week, the introduction of a feature called an operator that will automate tasks such as resting rest, filling in forms, reservations of restaurants and food ordering.

AND Microsoft-Babard startup describes him as “an agent who can go to a web to do tasks for you,” he added that he was dressed in interaction with “buttons, menus and textual fields that people use daily.” It can also ask questions to further personalize the tasks it fills, such as the application information for other websites. Users can take control of the screen at any time.

– CNBC – Michael Bloom contributed to reporting.

Don’t miss these insights with CNBC Pro



Source link

Leave a Reply

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

Back to top button
Social Media Auto Publish Powered By : XYZScripts.com