Chinese AI startup DeepSeek hit the tech world by storm after achieving the seemingly impossible. At a fraction of the development cost, its R1 V3-powered model surpassed OpenAI’s proprietary o1 reasoning model across several benchmarks, including coding, math, and science.
The model’s emergence caused a drastic shift in market stocks. NVIDIA shed up to $600 billion in market valuation, debatably the largest drop in history to take place in a single day. While DeepSeek’s success seemingly remains a mystery, Meta Chief Technology Officer (CTO) Andrew Bosworth claims he’d predicted the whole thing six months ago before it unfolded (via Business Insider).
While responding to a question about DeepSeek’s emergence from one of his Instagram followers, Bosworth claimed it was “a funny one to watch unfold.” Further indicating:
“I actually had an email of me predicting it, that it would come from somewhere, didn’t know it’d be DeepSeek, like six months ago. So I think for those of us in the space, it was not as surprising as those out of it.”
Meta reportedly attempted to pick up the slack in the AI race using crude methods, including training its AI models with copyrighted content, further covering its tracks to avoid copyright infringement issues.
We were tracking DeepSeek when it launched like a month before it then became this major news item. I think it is both a big deal and also not as big as it’s made out to be. DeepSeek’s launch is a big deal because it’s a great open-source innovation. They’ve done some really truly great and novel work in memory architectures for model building. They’ve certainly advanced the state of the art for reasoning models. They’ve probably done a lot of distilling against existing models.
Meta CTO, Andrew Bosworth
DeepSeek is a “great thing, but it’s not a world-changing thing,” concluded Bosworth.
Admittedly, DeepSeek’s emergence has raised some eyebrows in the industry. An email shared by Microsoft CFO Amy Hood urged employees to remain focused on company goals amid recent AI developments in the industry, potentially including OpenAI’s $500 billion Stargate commitment to develop sophisticated infrastructure for AI advances. However, Microsoft CEO Satya Nadella touted DeepSeek as super impressive and that it’s good for business.
Related: DeepSeek is reportedly sending user data to Chinese telecom despite US ban
Elsewhere, a report suggests DeepSeek spent $1.6 billion and up to 50,000 AI NVIDIA GPUs to develop its open-source model, contradicting claims about its cost-effectiveness.
