On January 20th a Chinese tech company released the free version of their chatbot called DeepSeek. The AI chatbot, by all accounts, is about on par with existing widely available chatbots, like ChatGPT. It does not represent any new abilities or breakthrough in quality. And yet the release shocked the industry causing the tech-heavy stock market Nasdaq to fall 3%. Let’s review why that is, and then I will give some thoughts on what this means for AI in general.

What was apparently innovative about DeepSeek is that, the company claims, it was trained for only $8 million. Meanwhile ChatGPT 4 training cost over $100. The AI tech industry is of the belief that further advances in LLMs (large language models – a type of AI) requires greater investments, with ChatGPT-5 estimated to cost over a billion dollars. Being able to accomplish similar results at a fraction of the cost is a big deal. It may also mean that existing AI companies are overvalued (which is why their stocks tumbled).

Further, the company that made DeepSeek used mainly lower power graphics chips. Apparently they did have a horde of high end chips (the export of which are banned to China) but was able to combine them with more basic graphics chips to create DeepSeek. Again, this is what is disruptive – they are able to get similar results with lower cost components and cheaper training. Finally, this innovation represents a change for the balance of AI tech between the US and China. Up until now China has mainly been following the US, copying its technology and trailing by a couple of years. But now a Chinese company has innovated something new, not just copied US technology. This is what has China hawks freaking out. (Mr. President, we cannot allow an AI gap!)

There is potentially some good and some bad to the DeepSeek phenomenon. From a purely industry and market perspective, this could ultimately be a good thing. Competition is healthy. And it is also good to flip the script a bit and show that innovation does not always mean bigger and more expensive. Low cost AI will likely have the effect of lowering the bar for entry so that not only the tech giants are playing. I would also like to see innovation that allows for the operation of AI data centers requiring less energy. Energy efficiency is going to have to be a priority.

But what are the doomsayers saying? There are basically two layers to the concerns – fear over AI in general, and fears over China. Cheaper more efficient AIs might be good for the market, but this will also likely accelerate the development and deployment of AI applications, something which is already happening so fast that many experts fear we cannot manage security risks and avoid unintended consequences.

For example, LLMs can write code, and in some cases they can even alter their own code, even unexpectedly. Recently an AI demonstrated the ability to clone itself. This has often been considered a tipping point where we potentially lose control over AI – AI that an iterate and duplicate itself without human intervention, leading to code no one fully understands. This will make it increasingly difficult to know how an AI app is working and what it is capable of. Cheaper LLMs leading to proliferation obviously makes all this more likely to happen and therefore more concerning. It’s a bit like CRISPR – cheap genetic manipulation is great for research and medical applications, but at some point we begin to get concerned about cheap and easy genetic engineering.

What about the China angle? I wrote recently about the TikTok hubbub, and concerns about an authoritarian rival country having access to large amounts of data on US citizens as well as the ability to put their thumb on the scale of our internal political discourse (not to mention deliberate dumbing down our citizenry). If China takes the lead in AI this will give them another powerful platform to do the same. At the very least it subjects people outside of China to Chinese government censorship. DeepSeek, for example, will not discuss any details of Tiananmen Square, because that topic is taboo by the Chinese government.

It is difficult to know, while we are in the middle of all of this happening, how it will ultimately play out. In 20 years or so will we look back at this time as a period of naive AI panic, with fears of AI largely coming to nothing? Or will we look back and realize we were all watching a train wreck in slow motion while doing nothing about it? There is a third possibility – the YdK pathway. Perhaps we pass some reasonable regulations that allow for the industry to develop and innovate, while protecting the public from the worst risks and preventing authoritarian governments from getting their hands on a tool of ultimate oppression (at least outside their own countries). Then we can endlessly debate what would have happened if we did not take steps to prevent disaster.