In my predictions for 2025 I suggested AI would be the Technology story for 2025. This has proved the case even before we are out of January.
The Associated Press reports today that “A frenzy over an artificial intelligence chatbot made by Chinese tech startup DeepSeek upended stock markets Monday, fuelling debates over the economic and geopolitical competition between the U.S. and China in developing AI technology”.
Market favourites NVIDIA, Oracle, Alphabet and others fell when it became known that DeepSeek outperformed ChatGPT on Apples App store. The S&P 500 tech sector dropped 5.6% with NVIDIA one of the worst losing almost US$600 billion in value over the day -although it is still valued at almost US$3 trillion.
DeepSeek was founded just a couple of years ago and has around 200 employees. Its total investment appears to have been less than NZ10 million.
Compare that to OpenAI which is 10 years old, has 4500 emp0loyees and has invested US$6.6 billion.
DeepSeek has all the hallmarks of a disrupter.
Is it any good? I am not a tech expert. I’m an economist. But I listened to Hosking this morning and his tech correspondent waxing lyrical about how well it came up with the rules of a complicated game the app seemed to work well.
DeepSee3k is apparently more efficient in that it uses 75% less memory. At the same time it is twice as fast. It apparently runs well on older technology and doesn’t need the extra special GPUs developed by NVIDIA and Google. All of which means it uses fewer energy resources. And that all means it is cheaper to run.
NVIDIA makes GPUs which have been the key innovation to allow AI to become accessible to the masses. Alphabet has Tensor Processing Units which are also focussed on AI. The market’s first reponse is that these products might become obsolete really fast.
What does this mean?
AI is possibly now not a rich person’s game. Some have likened this development to when the PC replaced the mainframe. Is this such a breakthrough?
This also got me thinking about how most people are not quick at jumping on newer, especially disruptive, technology. Often it takes a generation to make something commonplace.
What is also true is that there is a culture in parts of society that says only the old things are good. We hear it in marketing “Like grandma used to make”, “2000 year old herb recipe”, “our traditional ways” etc. I like old stuff. I’ve just finished re-reading a classic novel written in 1948. It is called 1984 by a Socialist called George Orwell and should be compulsory reading in Year 13.
And how will NVIDIA and Alphabet respond to the new chatbot in the playground?