{"id":27914,"date":"2025-01-29T14:54:03","date_gmt":"2025-01-29T06:54:03","guid":{"rendered":"https:\/\/www.1ai.net\/?p=27914"},"modified":"2025-01-29T14:54:03","modified_gmt":"2025-01-29T06:54:03","slug":"%e8%8b%b1%e7%89%b9%e5%b0%94%e5%89%8d-ceo-%e5%9f%ba%e8%be%9b%e6%a0%bc%e5%8a%9b%e6%8c%ba-deepseek%ef%bc%9a%e5%b0%86%e6%8e%a8%e5%8a%a8-ai-%e6%99%ae%e5%8f%8a","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/27914.html","title":{"rendered":"Former Intel CEO Kissinger endorses DeepSeek: will drive AI ubiquity"},"content":{"rendered":"

in the near future DeepSeek<\/a> Launched an open-source AI<\/a> inference modeling, an event that sparked a furor in the tech industry. On one hand, NVIDIA's stock plummeted, while on the other, DeepSeek's consumer app quickly climbed to the top of the app store charts.<\/p>\n

\"a6d7dc79j00squ8fl001ad000pm00hkp\"<\/p>\n

Last month, DeepSeek announced that they had completed training the model in only about two months at a cost of about $5.5 million using a data center containing about 2,000 NVIDIA H800 GPUs. Last week, DeepSeek released a paper showing that the performance of its latest model is on par with the world's most advanced inference models. And most of those world-leading models are trained in data centers that spend billions of dollars on NVIDIA's faster, more expensive AI chips.<\/p>\n

The introduction of this high-performance, low-cost model by DeepSeek has sparked a fierce reaction in the tech industry.Intel<\/a>Pat Gelsinger, former CEO, hardware engineer, and current chairman of soon-to-be-launched startup Gloo, a communications and engagement platform for churches, exclaimed at X, \"Thank you DeepSeek team.\" During his tenure at Intel, he unsuccessfully tried to catch up with Nvidia by launching Intel's alternative AI GPU, the Gaudi 3 AI. Kissinger left Intel last December.<\/p>\n

Kissinger argues thatThe emergence of DeepSeek should remind the tech industry of three overriding lessons<\/strong>: lower costs mean wider adoption; creativity flourishes within constraints; \"openness wins.\" He noted that DeepSeek will help break down the increasingly closed landscape of basic AI modeling efforts. Currently, companies like OpenAI and Anthropic use a closed-source model.<\/p>\n

Kissinger told TechCrunch that the R1's performance has been impressiveGloo has decided not to use and pay for OpenAI's services.<\/strong>Gloo is developing an AI service called Kallm that will offer chatbots and other services. He said.\"My Gloo engineers are using R1 today.<\/strong>. They could have used o1, but they could only access o1 through the API,\" he revealed.Gloo plans to rebuild Kallm from scratch in the next two weeks, using their own open source base model<\/strong>. To which he said, \"It's exciting.\"<\/p>\n

Kissinger argues thatDeepSeek will make AI more affordable, and not only will AI be everywhere, but quality AI will be everywhere too!<\/strong>. As an example, he says, \"I want better AI in my Oura smart ring, I want better AI in my hearing aids, I want more AI in my cell phone, and I want better AI in my embedded devices, such as the voice recognition system in my electric car.\"<\/p>\n

Kissinger emphasized thatDeepSeek Proves AI Advancements Can Be Made Through Engineering Creativity<\/strong>, rather than simply putting in more hardware and computing resources. He is very excited about this.<\/p>\n

Kissinger added: \"Having Chinese developers remind us of the power of open ecosystems can be a bit awkward for us in the Western tech community.\"<\/p>","protected":false},"excerpt":{"rendered":"

DeepSeek recently launched the Open Source AI Logic Model called R1, an event that triggered a wave in the technology industry. On the one hand, the share price of Ingweida fell sharply, and on the other hand, the consumer application of DeepSeek rose rapidly to the top of the app store. Last month, DeepSeek announced that it took them only about two months to complete training on the model at approximately $5.5 million, using a data centre containing around 2000 Tai Ying Weida H800 GPU. Last week, DeepSeek published a paper showing that its latest model is comparable to the world's most advanced reasoning model. Most of the world's leading models cost billions of dollars to buy Britain faster and more expensive<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[411,3606,240],"collection":[],"class_list":["post-27914","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-deepseek","tag-240"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/27914","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=27914"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/27914\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=27914"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=27914"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=27914"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=27914"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}