Has China won the AI battle?
27 January 2025, 14:12 | Updated: 27 January 2025, 14:25
Deepseek-R1 has already struck fear into Wall Street, but not LBC's Will Guyatt.
Listen to this article
Loading audio...
Out of nowhere, Deepseek, a previously unknown Chinese tech start-up, has struck fear into the hearts of western tech businesses with its new Artificial Intelligence model, Deepseek-R1.
Based on the level of Silicon Valley drama over the weekend and Wall Street's alarming tech stock sell-off - it would appear many are both surprised and alarmed by the speed of progress and financial efficiency of its design - with other tech watchers calling this a "wake-up call."
If you take Deepseek's claims at face value - they are reported to have spent over $6m training their AI model - built on pre-existing open-source models developed by the tech community, in comparison to the hundreds of millions spent by US tech companies to build higher profile AI models that power popular chatbots like OpenAI's ChatGPT, Google's Gemini or X's Grok.
Apart from some highly paid Western AI engineers having their noses pushed out of joint, the US administration likely has some deeper concerns.
Sanctions against China were meant to deny access to the most advanced semiconductors and AI chips. This also challenges the view that the AI revolution will be driven by really powerful and incredibly expensive gear.
Deepseek claims they've used significantly weaker computing tech available to Chinese businesses and instead spent their efforts optimising their algorithms and their processes, a bit like race-tuning a slightly rusty-looking Fiat to outperform a shiny, brand-new Ferrari. Did Deepseek have any Ferrari's hidden behind the Fiats? I guess we'll never truly know.
The Deepseek-R1 app has shot to the top of the App chart on Apple and Android devices - but that's not the real story. Outside of those fretting that this means the end of the Western dominance of AI, there's also some genuine excitement in what's possible with Deepseek's model - it could really benefit those with limited funding or tech resources, particularly those in poorer countries.
Who decided the AI revolution should only remain in the hands of the rich? Western governments and tech bros like Musk and Zuckerberg, of course.
My view on this Deepseek news is in fact much closer to that of Meta’s chief AI scientist Yann LeCun who doesn’t believe this is a knock-out blow to the Chinese - instead, an example of a cheaper open-source model surpassing a proprietary one that cost way more to develop, and a lot more for businesses to tap into. Like me, LeCun believes we’ll end up with many different AI models for different applications.
Anyway - surely competition is a good thing as long as it's fair - no matter where it comes from?
We all need somebody to keep the Googles, Metas, and Open AIs of this world on their toes, whether they are American, Chinese, English, or Moldovan.
