šŸ¤– Former Intel CEO Pat Gelsinger Ditches OpenAI for DeepSeek’s R1 at Gloo

Pat Gelsinger embraces DeepSeek's open-source AI, shaking up the industry and leaving OpenAI in the dust at his new startup.

Rafelia

DeepSeekOpenAIPat GelsingerAITech Trends

Tech NewsAI Insights

443

2025-01-29 05:30 +0530


🚨 Pat Gelsinger Chooses DeepSeek Over OpenAI at His New Startup, Gloo

What do you do after leaving Intel and trying (but failing) to outdo Nvidia with your own AI chips? šŸ¤”
If you’re Pat Gelsinger, you jump ship to DeepSeek and stir the AI pot! šŸµ

Gelsinger, now chairman of Gloo, a messaging platform for churches ⛪, is all-in on DeepSeek’s R1 model—a high-performance, low-cost open-source AI reasoning model that’s making waves 🌊 (and tanking Nvidia’s stock šŸ“‰).


šŸ’° DeepSeek’s $5.5M AI Revolution vs. Billion-Dollar Behemoths

DeepSeek trained its R1 model in two months using 2,000 Nvidia H800 GPUs, costing a mere $5.5 million. Compare that to U.S. giants throwing billions at AI chips šŸ¤‘, and you can see why R1 is the talk of the town.

Gelsinger couldn’t hide his excitement:

ā€œThank you DeepSeek team!ā€


🤷 Why Skip OpenAI?

Gloo, which is building an AI service called Kallm (yes, like ā€œcall meā€ šŸ“ž), decided to ditch OpenAI’s APIs in favor of R1.

Gelsinger’s reasons?

  1. Lower cost = More AI everywhere šŸ’ø
    • ā€œI want better AI in my phone, my Oura Ring, my hearing aid, and even my EV!ā€ šŸš—šŸ’”
  2. Open wins šŸ‘
    • Unlike OpenAI and Anthropic, DeepSeek’s open-source approach reminds us of the power of ingenuity.
  3. No dependency on closed ecosystems šŸ”’
    • Why rent an API when you can build your own AI service from scratch? 🤷

In just two weeks, Gloo plans to have Kallm rebuilt using DeepSeek’s model. šŸš€


šŸ¤” Skepticism? Gelsinger Doesn’t Care

Of course, DeepSeek’s meteoric rise hasn’t been without controversy:

  • “Fake numbers!” Some skeptics argue DeepSeek’s training costs must’ve been higher. šŸ•µļø
  • “Wait for OpenAI’s o3!” Others think OpenAI’s next model will obliterate R1. šŸ’„
  • “China concerns!” Privacy, censorship, and geopolitics loom large. šŸŒ

Gelsinger? He’s unbothered. ā€œDeepSeek is 10-50x cheaper than o1,ā€ he says.

ā€œHaving the Chinese remind us of the power of open ecosystems is maybe a touch embarrassing for the Western world,ā€ he adds. 😬


šŸ’” Why This Matters

DeepSeek’s R1 isn’t just a win for affordability—it’s a wake-up call for the AI industry. 🚨

Instead of throwing more hardware at problems, creativity and engineering are pushing the boundaries of AI. šŸ› ļø

Meanwhile, Gelsinger is proof that even as AI heats up, open-source innovation might be the key to democratizing AI everywhere.


🤩 Final Thoughts

With DeepSeek’s R1, the AI landscape is shifting—Nvidia, OpenAI, and Anthropic better watch their backs.
As for Pat Gelsinger, he’s already ahead of the curve, rebuilding Kallm from scratch with open AI power. šŸ’Ŗ

Who’s next to join the DeepSeek revolution? šŸ‘€

šŸ”— Follow the story and more AI updates here: https://raf-97414.github.io/en/posts/

#DeepSeek #AI #OpenSource #PatGelsinger #OpenAI #TechTrends