Post Thumbnail

$80 billion didn’t help: Microsoft acknowledges DeepSeek’s superiority

Microsoft is radically rethinking its artificial intelligence strategy following DeepSeek’s impressive breakthrough. The corporation’s head, Satya Nadella, admitted that the achievements of a small team of developers have set a new benchmark of success for the tech giant.

In January, Microsoft quickly integrated the DeepSeek R1 model into Azure. DeepSeek’s success is particularly notable because the company managed to optimize performance at a level below Nvidia’s CUDA layer, which significantly increased the efficiency of their AI models.

“The most impressive thing about DeepSeek is that it’s a beautiful reminder of what 200 people united by one idea can achieve,” Nadella said during a corporate meeting, responding to a question about Microsoft’s competitiveness in AI. “And what’s especially important – they didn’t stop at the level of a research or open-source project, but created a product that became number one in the App Store. For me, that’s a new standard.”

Microsoft, despite investing $80 billion in AI, access to the latest OpenAI models, a large-scale advertising campaign during the Super Bowl, and updating the Copilot design with voice and visual features, has not yet achieved similar success. While ChatGPT regularly ranks first among free applications, Copilot is often outside the top 100.

In response to these challenges, Microsoft introduced its own Muse model, trained on the Xbox game Bleeding Edge to generate gameplay. The corporation plans to use Muse to help Xbox developers prototype games and optimize them for modern hardware.

Jay Parikh, head of the new CoreAI engineering group at Microsoft, emphasized the need to reconsider internal work organization: “Considering what DeepSeek achieved with 200 people… it makes us think about how we work together within the company, where we have many organizational boundaries.”

The key challenge for Microsoft, besides creating a popular application, will be the ability to quickly respond to the emergence of new models and develop its own breakthrough technologies, as demonstrated during the integration of DeepSeek. In the near future, some interactive AI features created with Muse will appear in Microsoft Copilot Labs, which could make the product more attractive to the mass user.

Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

Latest News

Nvidia introduced Cosmos model family for robotics

Nvidia company introduced the Cosmos family of AI models. Which can fundamentally change the approach to creating robots and physical AI agents.

ChatGPT calls users "star seeds" from planet Lyra

It turns out ChatGPT can draw users into the world of scientifically unfounded and mystical theories.

AI music triggers stronger emotions than human music

Have you ever wondered why one melody gives you goosebumps while another leaves you indifferent? Scientists discovered something interesting. Music created by artificial intelligence triggers more intense emotional reactions in people than compositions written by humans.

GPT-5 was hacked in 24 hours

2 independent research companies NeuralTrust and SPLX discovered critical vulnerabilities in the security system of the new model just 24 hours after GPT-5's release. For comparison, Grok-4 was hacked in 2 days, making the GPT-5 case even more alarming.

Cloudflare blocked Perplexity for 6 million hidden requests per day

Cloudflare dealt a crushing blow to Perplexity AI, blocking the search startup's access to thousands of sites. The reason? Unprecedented scale hidden scanning of web resources despite explicit prohibitions from owners!