“I will not risk open war.”
“Open war is upon you, whether you would risk it or not.”
Aragorn in his response to King Theoden in the Lord of Rings.
This week a bombshell dropped in the Artificial Intelligence (AI) market which is going to have huge repercussions across the industry and globally. If this news went over your head or you switched off because it sounded too techy, I will break it down for you in a manner which highlights the most significant points.
A Chinese start-up called DeepSeek has developed an AI platform which effectively operates at similar rates of efficiency to American-designed competitor platforms, at a fraction of the cost. For comparison, DeepSeek’s V3 model cost approximately $5.6 million to train whereas OpenAI’s GPT-4 model cost over $100 million. That is 18 times cheaper. What had previously been total American market dominance was just turned on its head.
DeepSeek has gone a step further, it has released DeepSeek-V3 online for free, which you can register for and use with your Internet browser. It has also open-sourced its models and shared its research publicly, allowing others to build on it, and effectively opening up how its AI models are able to do high-end work with less memory and computing power.
This is important to know, because the future of the modern economy is going to be shaped by Artificial Intelligence. In fact, it is most likely that you will either be someone who has the skills to tell computers what to do, or you will be someone who is told what to do by computers. Perhaps you will get lucky and find some middle ground where you actively consult computers but still make or input into final decisions. But for how long?
Everyday work
That is how to use Artificial Intelligence platforms, which I have now integrated into my everyday work, primarily to assist in research but also to work through large documents, draw out information, map out data, or search sources for information. It saves a significant amount of time, particularly if you work in a space where large quantities of information and data are generated, and you need to reflect on information from the past.
But why then such a dramatic headline? Surely AI is going to be good for the world?
Well, that all depends on the norms and standards that will shape its regulation, its deployment and uses. In this there is a challenge that faces governments across the world. Many politicians and government officials in the state do not have the skills to truly understand AI platforms, their uses and public impact. Thus, they rely heavily on the industry to assist in designing the regulations.
Take DeepSeek for example, a Chinese platform which appears to have built into its training certain biases that exist within the Chinese government. For example, the platform avoids providing a true reflection on sensitive topics such as the Tiananmen Square protests, Taiwan, and other matters deemed sensitive by the Chinese government. It is effectively a great platform for reasoning and mathematics, but not quite as “free” as was proposed.
The European Union and the United States of America are creating regulatory reforms to drive the safe and ethical use of AI. President Donald Trump has recently signed a new Executive Order: “Removing Barriers to American Leadership in Artificial Intelligence”, which seeks to remove restrictions on AI innovation and drive more “unbiased and agenda-free” development.
OpenAI, Google Cloud AI, Microsoft Azure AI, IBM Watson, Amazon Web Services, Meta Platforms, Apple, Salesforce Einstein, Adobe Sensei – all of these are American companies which are integrating AI into their service offering to customers with specific products. Total American market dominance.
Binds them
But of course, there is one company I did not mention here. The one company that brings much of this technology together. That rules them all, that finds them, that brings them all and, in their servers, binds them. Nvidia.
Nvidia took the technology industry by storm in the late 1990s with its Graphic Processing Units (GPUs). When we grew up, there was no greater luxury item than having your computer powered by one of the Nvidia G-force graphics cards which could play the latest games. These GPUs and their application expanded across the technology sector, and now they form the backbone of AI and machine-learning systems.
In the last year, as AI took the world by storm, Nvidia’s share price skyrocketed and it soon become one of the largest companies in the world.
Nvidia’s power and capability is of such a strategic nature that the USA restricted the sale of their A100 and H100 GPUs to China, due to their advanced capabilities which have extensive military and surveillance uses. So, what did DeepSeek use to create their AI platform? Nvidia’s lower-performing H800 chips which had not been banned in China.
They effectively used other optimisation techniques to make up for the decreased computing power and it is this that has turned the industry on its head. Nvidia’s share price dropped by 17% and fell below $3 trillion. Suddenly their high-end GPU chips are not as desirable as was once thought.
The DeepSeek creators have admitted that despite producing this platform, China continues to trail the USA, always the imitator and never the initiator. Nonetheless this salvo into the sector is likely to set off more disruption, and force the USA government to evaluate how much of this critical hardware and intellectual property it wishes to share with the world.
The views of the writer are not necessarily the views of the Daily Friend or the IRR.
If you like what you have just read, support the Daily Friend
Image by Chathura Anuradha Subasinghe from Pixabay